THE

AMERICAN NATURALIST

A MONTHLY JOURNAL
DEVOTED TO THE ADVANCEMENT OF THE BIOLOGICAL SCIENCES

WITH SPECIAL REFERENCE TO THE FACTORS OF EVOLUTION

VOLUME XLIX

NEW YORK
THE SCIENCE PRESS

1915

1915

ARY,

7

, 577

NO

x

VOL

“4

fik
a

a

me

wit

The American Naturalist

MSS intended for publication and books, etc., intended for review should be
sent to the wep = THE AMERICAN NATURALIST, Garrison-on-Hudson, New York.
es containing summaries of research work bearing on the

problems of Fabien evolution are especially oleae and will be given preference

One hundrea Sy aips of contributions are supplied to authors free of charge.
Further rrun will agpi oe at cost

Subse rtisements should be sent to the publishers. The
erah price “8 Soe a dolare a year. Foreign postage is fifty cents and
Canadian 'y-five cents additional. The cha ape for sane copies is
forty cents. The E moveri A rates are Four Dollars for a

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.
_ NEW et _Sub-Station 84
Entered : , April 2 ster, Pa., under the Act of
ria sai of March 3, 1879.

FOR SALE JAPAN NATURAL HISTORY SPECIMENS

ARCTIC, ICELAND and GREENLAND __ Perfeot Condition and Lowest Prices.
ae BIRDS’ SKINS, Specialty: Bird Skins, Oology, Entomology, nen
© Well Prepared Low Prices | AMimals and others. Catalogue free. Correspond-

= Particulars of

THE
AMERICAN NATURALIST

Vou. ALIX January, 1915 No. 577

. SOME FUNDAMENTAL a aaaa
OBJECTIONS TO THE MUTATION
THEORY OF DE VRIES

PROFESSOR EDWARD C. JEFFREY

HARVARD UNIVERSITY

Tuer hypothesis of the saltatory origin of species has
received a new impetus from the investigations of De
Vries,* published in his ‘‘Mutationstheorie’’ and subse-

1To an address delivered in Brussels before the outbreak of the war and
published in Science (Vol. 40, No. 1020, July 17th. ), Professor de Vries
appends a criticism of the eitri ‘rolinliney article on mutation, likewise
published in Science (Vol 39, No. 1005, April 3d.). The gist of his objection
to the writer’s tae that Œnothera and other members of the Ona
are in a position of hybrid contamination, as evidenced by the frequent

rility or en sterility of their pollen, is the contention that polle
sterility and gametic sterility in general is not sufficient evidence of hybrid
contamination. To this statement two replies may be made. In the first
Place prominent geneticists for many years have recognized pollen sterility
as important evidence of edema Secondly investigations, which have

now become very extensive, on the iosperms as a whole, show very inter-
esting conditions in many natural Faaite While the monotypie species
and those which are isolated geographically or phenologically (that is by a
time of flowering later or earlier than that of the mass of species belongin
to the genus) have invariably good pollen, those species, which me in
their geographical range and in their times of flowering in many ¢
characterized by abortion of the reproductive cells. In other words speia
infertility is only found where the possibility of crossing is present. This
PT has been illustrated in the body of the present article by reference

the Rosaceæ. ing a further illustration from the large family
Barne Ranunculus acris and R. repens, which overlap both in range
and time of flowering have pollen, which is ithe largely imperfect, particu-
larly in the first mentioned species. R. rhomboideus on the other hand,
flowering in the very early spring has perfect pollen development. _

5

6 THE AMERICAN NATURALIST [Vor. XLIX

quent works. The chief foundation for his views, in re-
gard to the instantaneous origin of species, is furnished
by the conduct of Enothera lamarckiana in cultures. It
has been somewhat generally recognized that O. lamarck-
iana, and more recently, other species of the genus as well,
constitute crucial evidence in regard to the validity of the
mutation hypothesis on the botanical side. A great many
investigations on the genetics and cytology of O. lamarck-
iana and other species, as well as crosses between species
and ‘‘mutants’’ of Enothera, have been carried on during
the past decade by De Vries, and his followers and oppo-
nents. As a result a huge and highly technical literature
has grown up. @Œnothera is obviously regarded, on the
botanical side at any rate, as the touchstone of the muta-
tion hypothesis as formulated by De Vries. Obviously
if this genus does not stand the test of critical investiga-
tion, the mutation hypothesis, so far as its validity de-
pends upon De Vries’s chosen illustration, is discredited.

Since @nothera and by obvious implication the Ona-
grace, to which it belongs, have become authority for
the mutation hypothesis, in its latest revival, they must
like Cæsar’s wife be beyond suspicion. Like Cæsar,
(Enothera has become a name of authority and its family
affairs accordingly, should be beyond suspicion, when sub-
jected to the most searching investigation. It is appar-
ently just in this direction that the weak spot of the muta-
tion hypothesis lies. Too much attention has apparently
been given to ringing the changes on the so-called mutants
of Gnothera and not enough to the investigation of
the general morphological situation in the Onagracee, to
which this much-discussed genus belongs.

Unusual variability in plants is ordinarily regarded as
prima facie evidence of hybridism and the suggestion has
in fact frequently been made by professional geneticists
(e. g., Bateson, Davis, East, Gates and others) that
(nothera lamarckiana is a hybrid. It is perhaps of inter-
est in this connection to recall that one of the commonest
expedients adopted by the practical breeder, for breaking

No. 577] MUTATION THEORY OF DE VRIES 7

up the continuity of the germ plasm, is hybridization.
Apposite in this connection is the wholesale hybridizing
practised by Burbank, for the purpose of bringing about
the necessary genetic plasticity in his cultures and thus
obtaining by resultant mutation or variation, new and
desirable varieties of useful plants. The morphological
peculiarities of hybrids have been clearly recognized for
nearly a hundred years. They are for example clearly
set down in Gaertner’s rare and classic prize essay, en-
titled ‘‘ Versuche und Beobacthungen ueber die Bastarder-
zeugung im Pflanzenreich’’ (Stuttgart, 1849). Curiously
enough these important criteria have been largely ignored
by the adherents of the mutation hypothesis of De Vries.
A very important and generally observed difference be-
tween hybrids and genetically pure species, is the very
easily detected one of pollen sterility, partial or complete.
Of course when the hybridizing forms show a considerable
degree of compatibility, this character may be inconspicu-
ous or even absent. Further even in cases where it is
originally present, it may be subsequently largely elimi-
nated by selection. De Vries himself has noted that about
one third of the pollen of O. lamarckiana is abortive. The
English geneticist Bateson was struck with this peculiarity
of the species, so much discussed in recent years, in rela-
tion to its variable offspring in cultures and promptly and
first called attention to the obvious significance of this
feature, suggesting that O. lamarckiana was a hybrid and
that its remarkable conduct was the result of hybridiza-
tion. This objection has in reality never been met. It is
the purpose of the present article to show on grounds
commonly accepted by geneticists and morphologists, that
not only is genus Œnothera in general characterized by
genetically impure or hybrid species, but that the condition
of genetical impurity is extremely common in the nma
graceæ as a whole.

It will be convenient to begin with the examination of
our common and very variable garden Fuchsias, which
belong to the family Onagraceæ. The common Fuchsia,

8 THE AMERICAN NATURALIST [ Vou. XLIX

sometimes known to gardeners as Fuchsia speciosa, is
recognized as a hybrid derivative of Fuchsia magellanica,
a native of southern South America. Fig. 1 illustrates

bic. 1

photomicrographically, the condition of the pollen in one
of the garden varieties of Fuchsia. The sound pollen
grains appear as dark bodies with two or more germina-
tion pores projecting from their surfaces. The dark color
of the grains is due to the deeply staining character of
their protoplasmic contents. More than a third of the
pollen present in the anther cavity is abortive and is
represented in the photograph by shrivelled light-colored
objects, which are in fact empty and collapsed pollen
grains. In other varieties of the garden Fuchsia, the
grains may either be entirely abortive and empty (as is
the case for example in the so-called mutant of Ginothera
lamarckiana, known as O. lata) or they may all be more or
less well developed so far as their protoplasmic contents
are concerned, but extremely varying in size. In the pres-
ent description, perfection or imperfection of pollen is
judged only from the morphological aspect, because this
is the significant point of view from the standpoint of the

No. 577] MUTATION THEORY OF DE VRIES 9

detection of hybridization. Physiological sterility is fre-
quently due to entirely different causes than genetical lack
of harmony, as for example in the horseradish “or the
potato (Solanum). In the former it has been found pos-
sible to bring about the formation of fertile seed by simply
girdling the top of the subterranean storage region of the
plant, so as to prevent the undue descent of assimilates.
The common white lily, Lilium candidum, presents a
similar condition, for here the setting of seed takes place
only when the leafy flowering axis is severed from its
bulb and kept in water. So far as I am aware, there have
been no experiments as to the result of severing the con-
tinuity of the phloem (girdling), in relation to the restora-
tion of seed production in the potato. The common yellow
day lily (Hemerocallis) possibly presents a case similar
to that of Lilium candidum, for it does not ordinarily set
seed, although in all the examples I have examined the
pollen was morphologically perfect. I have not yet been
able to secure flowers of any pure species of Fuchsia, a
genus which flourishes mostly in the remoter parts of
South America and in the New Zealand islands. The
cultivation of Fuchsias, although once very popular, has
now gone out of vogue and it is consequently difficult to
secure specimens of the species. As has been pointed out
the commonly cultivated Fuchsias are of hybrid origin.
We may now turn our attention to a very puzzling genus
of the Onagracex, namely Epilobium. This genus has
been a great riddle.to systematists and the determination
- of species has been extremely difficult on account of their
extreme variability. In European systematic works, this
high degree of variability is recognized clearly to be
largely due to hybridization and in such a standard work
as the ‘‘Naturliche Pflanzenfamilien’’ of Engler and
Prantl, the statement is definitely made that the various
species of Epilobium frequently and commonly hybridize
with one another in nature. Let us consider in this con-
nection the northern hemisphere cosmopolitan species,
known as Epilobium angustifolium, the willow herb or

10 THE AMERICAN NATURALIST [Vou. XLIX

fire weed, which by contrast to many of the other Epilo-
biums, is so constant and distinct that it is frequently
referred to a separate genus, Chamenerion. This species
shows its most marked distinction from other species of
Epilobium (Epilobium proper) in the fact that its pollen
grains are separate and not in tetrads, as is the case in
other common species. Fig. 2 reproduces photograph-

ically a transverse section of a mature flower bud of
E. (Chamenerion) angustifolium. On the outside are
seen the floral envelopes, two in number, composed, as is
the rule in the Onagracex, of four parts each. Within lie
four stamens represented by their anther sacks and inter-
nal to these are four stigmas representing the carpellary
or ovarial portion of the flower. The photograph is on a
sufficient scale of magnification to show the pollen grains
in the loculaments or cavities of the anthers. Obviously
the pollen is very uniform and perfect in its development.
Fig. 3, likewise photographic, illustrates the organization
of the pollen as viewed with a much higher magnification
of the microscope. Although some of the grains are only

No. 577] MUTATION THEORY OF DE VRIES 11

partially included in the plane of section, it is quite clear,
that like those of Fuchsia, figured above, they have pro-
jecting germination pores, but unlike the Fuchsia of our
illustration, all the pollen grains of Epilobium (Chame-
nerion) angustifolium are perfectly developed. I have
examined the pollen of the species under discussion from
widely separate geographical regions and under different
conditions of growth and season, with the uniform result,

E n As p 7
a AR “en pa
Fie. 3

that the pollen is perfect and invariable in any important
respect. E. angustifolium is a species which apparently
is not known to hybridize with other species and indeed it
is not easy to see how it could cross with those having
their pollen grains in tetrads. The perfection of the
pollen in view of this condition appears particularly sig-
nificant. The failure of E. angustifolium to hybridize in
- nature with other species of the genus is doubtless due to
the fact that it is morphologically very distinct from these
and would in all probability produce, if artificially crossed,
only sterile hybrids.

We may now turn by way of comparison to a species
of Epilobium of the ordinary type. Fig. 4 illustrates

12 THE AMERICAN NATURALIST [ Vou. XLIX

photographically the floral organization of Epilobium
hirsutum, as seen in transverse section of the bud just
about to open. The illustration shows the floral envelopes

and the stamens, together with the pistillary portion of the
flower, the latter being somewhat displaced in the figure
and cut through the region of the style. The long hairs
] teristic of the calyx of this species have been
trimmed off, for the purpose of facilitating photo-
mechanical reproduction. As in the two illustrations
above, the anther sacks are the most significant feature.
Even with the low magnification employed for the purpose
of illustrating the whole flower, the pollen grains in the
loculaments of the anthers are easily discernible and pre-
sent a striking contrast to those of E. angustifolium, in the
respect that they are in groups of tetrads. Some of the
groups are partially or wholly made up of individual
grains without protoplasmic contents, which are smaller
in size than the normal grains. Fig. 5 shows one of the
anthers much more highly magnified. The anther walls,
cavities and the pollen grains are now clearly distinguish-

No. 577] MUTATION THEORY OF DE VRIES 13

able. Some of the grains are full size and present dark
contents. Others are considerably smaller and are devoid
of protoplasm. The latter are abortive or sterile grains.
We have in fact before us a hybrid derivative of E. hir-
sutum, commonly found near ballast in New England and
not unfrequently cultivated in gardens. Other species of
Epilobium in the stricter sense of the generic appellation,
show similarly abortive pollen development and the con-
clusion reached by old world systematists on the external

Fie. 5

characters, that hybridization is common among the spe-
cies of Epilobium proper, is entirely confirmed by the
study of the pollen. It need hardly be emphasized in this
connection, that imperfect pollen development has been
recognized for nearly a century by scientific plant breed-
ers, as a criterion of hybrids.

The genus Œnothera may now be profitably considered.
Fig. 6 presents a magnified view of a transverse section of
a mature flower bud of one of the commonest of eastern
species of (nothera, namely Œnothera biennis. The
floral envelopes are more voluminous than in the two
genera illustrated above. Within are the stamens and in
the center of the figure the style appears as a large
rounded structure. Even with the low magnification em-
ployed, it is easy to discern that the contents of the anther
sacks present a very different appearance from those of

14 THE AMERICAN NATURALIST [ Von. XLIX

Epilobium angustifolium. Many of the grains of pollen
are light colored and devoid of the protoplasm which
gives a dark appearance to the sound grains. Fig. 7 illus-

r ia PERRE 2

mx
Pere
Passio =D.

Fig. G

trates a single stamen under a high degree of magnifica-
tion. The characteristic layers of the wall of the anther
sack, described comparatively and in detail in the classic
memoir of Chatin, can readily be distinguished. Within
lie the pollen grains. Clearly only a few of these are fully
developed and possess normal protoplasmic contents.
The greater number are shrivelled and empty. Judged
from the generally accepted canon of the abnormalities of
hybrids, O. biennis is of hybrid origin. This view of its
nature is in harmony with its wide degree of inconstancy
throughout its very extended range. This feature is
doubtless responsible for the fact that the genus Œnothera
is at the present time undergoing considerable elaboration,
on the part of systematists. I have satisfied myself that
the pollen peculiarities of O. biennis are uniformly pres-
ent in specimens collected hundreds of miles apart, from
the Province of Ontario, the shores of the Gulf of St.

No.577] MUTATION THEORY OF DE VRIES 16

Lawrence and the New England States. I have further
examined a large number of species of Œnothera from
various parts of the continent and in every instance have
found a greater or smaller amount of abortive pollen as a
characteristic feature of the anther contents. De Vries
in his ‘‘Mutationstheorie’’ describes the abortive condi-
tion of about one third of the grains in O. lamarckiana:
This feature has been seized upon with insight by Bate-
son, as indicating the hybrid origin of O. lamarckiana. It
is extremely curious that its significance should have

Fig. T

escaped De Vries and his numerous disciples on this con-
tinent. Not only is O. lamarckiana itself characterized by
a large proportion of abortive pollen but its so-called
mutants are similarly characterized. In the feebler ‘‘ele-
mentary species’? the pollen is often almost entirely
abortive (O. nanella) and this is also generally the situa-
tion in O. lata. It should be further noted in this connec-
tion that if O. lamarckiana is of hybrid origin, the: same
statement must hold of the other species of @nothera,
since like this much-disputed one, they are similarly char-
acterized, so far as they have been studied, by two corre-
lated features, namely more or less abortive pollen and the
peculiarity of throwing so-called mutants or ‘‘elementary
species’’ in cultures. As a consequence of this condition,

16 THE AMERICAN NATURALIST [Vou. XLIX

it becomes more or less a superfluity to study any partic-
ular species of Œnothera from the genetical and morpho-
logical standpoint, since it is the genus as a whole which
manifests the peculiar features, which have brought it so
much into the foreground of biological controversy during
the past decade. This is on the whole a satisfactory situa-
tion as it enables us to cut the perplexing gordian knot
involving the controverted origin of O. lamarckiana.
The mutation hypothesis of De Vries accordingly turns
not upon the finding of new herbarium specimens which
may throw light upon the origin of a particular species
but upon the much larger question of the genetical status
of the genus Ginothera as a whole. This question can be
settled only by consideration of the Onagracee as a whole
and of other families of the Angiosperms, which present
similar reproductive peculiarities. _

Before proceeding however to the discussion of the
facts recorded above in their relation to the mutation
hypothesis of De Vries, based on the conduct of O. lamarck-
iana in cultures, it will be necessary to make some brief
_ reference to other studies carried on in the laboratories of
plant morphology of Harvard University, which will be
published elsewhere, either at the present time or at a
later period. Obviously of great importance in the pres-
ent connection is a comparison of the conditions of spo-
rogeny found among the lower plants, the Bryophyta, the
Pteridophyta and Gymnosperms, which are not character-
ized by enormous multiplication of species, with the
sporogenic features of the Angiosperms in which the
multiplication of species has run riot. Further compari-
son of liverworts, belonging to the Marchantiales, Antho-
sperms, manifesting similar sporogenic and specific pecu-
liarities, is both pertinent and necessary, in the present
connection.

It will be convenient to deal first summarily with the
sporogenic conditions found in the lower forms of the
Embryophyta from the Bryophyta to the Gymnosperms.
In the present connection a considerable number of spe-

No. 577] MUTATION THEORY OF DE VRIES 17

cies of liverworts, belonging to the Marchantiales, Antho-
cerotales and Jungermanniales, both acrogynous and
anacrogynous have been examined with the general result
that the only sterile cells present in the capsule cavities
were the elaters. Infertile spores and hybridism both were
conspicuous by their absence in the forms studied. The
same statement mutatis mutandis holds for the true
mosses. Some indication of spore abortion was detected
in the extremely variable genus Sphagnum. It would
seem that natural hybrids exist to some extent in this
genus. Among the Pteridophyta both the Lycopsida and
Pteropsida were studied. None of the numerous Lycopsid
forms investigated showed signs of spore abortion or
hybridism. Among the Pteropsida, the only well-known
hybrids are found among what is probably the highest
family, the Polypodiacew. There is a considerable litera-
ture upon hybrid ferns, in which references to spore abor-
tion as an accompanying feature are common. No evi-
dence of hybridism in the form of abortive spores was
found in examples of the Marattiacee, Ophioglossacex,
Osmundacee, Gleicheniacex, etc., were found, although a’
large amount of material was examined. Among the
Gymnosperms, the Cycadales, Ginkgoales, Coniferales
and Gnetales were examined. The Coniferales yielded
only a single species of Abies, which showed evidence by
the presence of abortive pollen grains of hybrid origin.
The genus Pinus is very old and its species accordingly
very distinct. Not the slightest evidence of hybridization
was found here or in other numerous and widely distrib-
uted species of conifers, other than Abies mentioned
above. This does not of course preclude the discovery of
such conditions later. The writer has had the opportunity
of examining the spores of a number of fossil forms from
the Paleozoic and Mesozoic, still contained within the
sporangia, and in no case were abortive spores recognized.
The general conclusion can be drawn from the forms just
considered that hybridism is rare among them and that

18 THE AMERICAN NATURALIST [Vor. XLIX

where it occurs it is accompanied by the phenomenon of
spore abortion.

If we turn to the Angiosperms with their nearly one
hundred and fifty thousand recognized species, we find
that hybridism is very commonly recognized. It would
take us much too far to discuss the situation here at any
length. The consideration of a single important family
must suffice. The one chosen, as being of particular sig-
nificance in the present connection, is the Rosacee. We
have had a recognition for many years past on the part of
systematic botanists in this country and in Europe that
hybridism is extremely common as a natural condition in
certain genera of the Rosaceæ. The inference in such
cases is generally based on the blended character of the
hybrids themselves, which show to a large extent a com-
bination of the characters of their parent species. Pro-
fessor Brainerd has recently made some very interesting
investigations in this direction in the case of American rep-
resentatives of the Rosaceæ. The recognized hybrid forms
in the Rosaceæ are usually characterized by a considerable
degree of pollen sterility, unless the parents happen to be
species not very remote in relationship. In addition to
the recognized hybrids of the rosaceous species, the work
carried on in the Harvard laboratories has revealed a
large number of hidden hybrids or erypthybrids, which
are quite constant in their characters and are recognized
by systematists as good species, but differ from normal
species in the fact that their reproductive cells are to a
greater or less degree abortive. Species of this kind are
extremely common among those rosaceous genera, which
have become of economic importance, such as Rubus,
Rosa, Pyrus, Malus, Sorbus, Crategus, ete. Taking Rosa
as an illustration, in addition to numerous recognized
hybrids, there are many types recognized as good species,
e. g., Rosa blanda, in which the pollen is normally largely
abortive, in still other species, frequently those which are
isolated geographically, the pollen is quite sound, ʻe. g.,
Rosa rugosa of Japan. The latter type of species must be

No. 577] MUTATION THEORY OF DE VRIES 19

regarded as a species in the strict sense, while those of
the type of Rosa blanda, in which abortive pollen similar
to that characteristic of forms clearly recognized as hy-
brids, is present, are hidden hybrids. It follows that in
Rosa (or practically any of the other rosaceous genera
cited above), there are three types of individuals, namely
good species, hidden hybrids and open hybrids. The
middle condition is extremely common among the Angio-
sperms and is of the greatest importance in connection
with clear views in regard to the origin of species. Obvi-
ously constant or relatively constant hybrids can not rank
with pure species, such as are characteristic for example
of the Gymnosperms, in discussions in regard to the
origin of species by mutation or otherwise. The conduct
of such forms is conditioned to a greater or less extent by
their mixed blood. We may appropriately designate obvi-
ous hybrids as phenhybrids and those hybrids which are
recognizable as such by their internal morphological char-
acters as crypthybrids. Crypthybrids will probably when
studied more extensively in cultures by the geneticist, give
evidence of their hybrid origin in cultures. There can
be no doubt that many of the recognized species of the
Angiosperms are in reality erypthybrids. The enormous
multiplication of species in this great group of plants is
in all probability largely related to hybrid crossing. It
is of the utmost importance however to keep clearly in
mind that such hybrid species or erypthybrids are not at
all in the position of true species from the evolutionary
standpoint and that conclusions derived from their study
can not be applied without large reserves, to the question
of the origin of species in the strict sense. The species of
Pinus, so far as we have any evidence, since the main
types are known to have existed well back into the Meso-
zoic, in all probability illustrate the origin of species some-
what along the lines of the Darwinian hypothesis. On the
other hand the species of Rosa present obviously an
entirely different problem in evolution and the necessity
of making distinctions if we are to reach any definite bio-

20 THE AMERICAN NATURALIST [Vou XLIX

logical goal is very clear. A great deal of the pessimism
which at the present time is sending too many biologists
after strange gods in other scientific shrines is doubtless
to be traced to the failure to make this distinction. It may
not be possible to make the distinction in all cases even
among the higher plants; but it certainly will be necessary
to realize its significance. Probably plants will in regard
to this possibility enjoy in this respect, as in so many
others, an advantage over animals in the studies of the
experimental evolutionist.

.We may now consider with advantage the status of the
species of the genus @Œnothera. The pollen sterility
which characterized them all to a greater or less degree
is indisputable evidence of their probable hybrid origin.
The general situation in regard to the criteria of hybrid-
ism in plants has been recognized for nearly a hundred
years. It has been made clear by Bateson in regard to
(Enothera lamarckiana. The observations chronicled
here appear to make it obvious that all the species of
(nothera are in the same boat genetically, that is that
they are all of hybrid origin. They likewise probably will
all be found to ‘‘mutate’’ just as O. lamarckiana, O. bien-
nis, etc., are already known to do. It may appear later
that there are certain species which have escaped, through
geographical isolation or other causes, the mingling of
blood, which is certainly characteristic of the Ginotheras
of the Eastern United States. So far as we know them
at present, the species of @nothera are obviously in the
same position as such species as Rosa blanda, that is they
are crypthybrids. Doubtless the peculiarities of O. la-
marckiana, O. biennis, etc., can be more clearly explained
in the present condition of our knowledge as the result of
hybrid origin than in any other way. It follows that the
doctrine of mutation so far as it depends for its support
upon the @notheras is in a discredited condition, as an
explanation, in any proper sense of the term, of the
origin of species.

No. 577] MUTATION THEORY OF DE VRIES 21

CONCLUSIONS

1. The Onagraceæ are largely characterized by hybrid
contamination in nature,

2. This statement holds with particular force for
(nothera lamarckiana and other species of the genus
(nothera, which have served as the most important basis
of the mutation hypothesis of De Vries.

3. Constant hybrids or erypthybrids are of very com-
mon occurrence among the Angiosperms and have been
illustrated in the present article by reference to the gene-
tical conditions occurring in certain Rosacee.

4. The species of Œnothera are to a large extent, if not
wholly, erypthybrids.

5. The objection raised by Bateson to the genetical
purity of @nothera lamarckiana is confirmed and is ex-
tended to the Onagracee in a general way, as well as to
other species of @nothera.

6. Hybridism is the best explanation yet put forward of
the peculiar conduct of @nothera lamarckiana, as well as
other species of the genus in cultures.

7. The mutation hypothesis of De Vries, so far as it is
supported by the case of (nothera lamarckiana, is
invalidated.

Ics. 1 AND 2 at Tor; Fics. 3 AND 4 IN THE MIDDLE; Fics. 5 AND 6 AT
1-4

~4, in terms of hia all the rabbits described in these experiments have been
classified. Fig. Rabbit 4214, Fat of the Serles I Young. Fig. 6, Rabbit
40A, Father of the Series II Youn

THE ENGLISH RABBIT AND THE QUESTION OF
MENDELIAN UNIT-CHARACTER CONSTANCY

W. E. CASTLE AND PHILIP B. HADLEY +

Wuatever the theoretical importance of Mendel’s law,
its practical utility depends largely upon the purity of
the gametes. If Mendelian unit-characters can through
hybridization be recombined in desirable ways without
essential modification during the process, Mendel’s law
is evidently a distinct acquisition to the practical breeder.
` Nevertheless, if crossing is likely to produce considerable
changes in the characters which it is desired to combine
in a new race, it is evident that Mendelian crosses must
be used judiciously and with caution by the practical
breeder.

Considerations such as these have led the senior author
for several years to concentrate his studies of genetic
problems upon the question of gametic purity. As a
crucial experiment he conceived the plan of deriving an
entire race of animals, not from a single pair of ancestors,
but from a single gamete, so far as concerns a particular
unit-character. It was thought that in a race so derived,
if the principle of gametic purity holds, there should be
no variation whatever in the particular unit-character
concerned.

Color patterns of mammals seemed especially well
adapted for such studies, since they are early differen-
tiated and clearly Mendelize in crosses. The so-called
“English” piebald rabbit presents an especially fine
example of such a color pattern. The figures give a
good idea of this striking pattern in which white and
colored areas are interspersed much as in the ‘‘coach-

1 Joint publication of the Laboratory of Genetics of the Bussey Insti-
tution, Harvard University, and of the Agricultural Experiment Station
of the Rhode Island State College (Contribution 211).

23

24 THE AMERICAN NATURALIST [Vou. XLIX

dog.” It would be a distinct gain to breeders if they
could reduce the variation in details of the English pat-
tern so that ‘‘prize-winners’’ could be bred without the
production of so many ‘‘wasters,’’ which depart in essen-
tial points from the standard pattern adopted for the |
breed. This was an additional reason for undertaking
work with the English rabbit.

The first standard-bred English rabbits which the
senior author had under observation, when mated inter se,
produced young of three sorts. About half the young
were fairly good ‘‘standard’’ English extensively marked
with colored spots (see Fig. 3). About one fourth were
much whiter than the standard demands, their spots being
fewer and smaller (see Fig. 1). And the remaining
fourth were without spots, that is, were self colored.
This last class was found to be recessive and not to pro-
duce English offspring, if mated inter se.

The whiter-than-standard English proved to be homo-
zygous for the pattern, the ‘‘standard’’ English being
heterozygous and breeding like their parents.

From these observations it was clear (1) that the Eng-
lish pattern is a Mendelian dominant and. (2) that the
breeding of English rabbits resembles that of blue
Andalusian fowls. For the standard-bred animal is a
heterozygote in the production of which there is bound
to be a constant production of ‘‘wasters’’ unless either
the standard is changed or the homozygote can be changed
to conform with the standard, producing an animal with
more color. In the latter case homozygotes could be
bred with each other and wasters eliminated. The ques-
tion whether the pattern can be changed becomes there-
fore one of practical as well as theoretical interest.

In making crosses of English with other breeds of
rabbits, there was found to be considerable variation '
among the heterozygous English produced, some being
much whiter than others, i. e., having less extensive
colored spots. Plus (dark) and minus (light) selections
were made to see to what extent the pattern was capable

No. 577] UNIT-CHARACTER CONSTANCY 25

of modification. These selection experiments are still in
progress, but will be reported upon at another time.

The single-gamete experiment, with which this report
will deal, was placed in the hands of the junior author,
who has carried it out at the Rhode Island Agricultural
Experiment Station.

As foundation stock for the experiment a single hetero-
zygous English rabbit of standard character (grade 2,
Fig. 5) was selected. To mate with him, it was desired
to obtain a distinct breed of rabbits, free from the Eng-
lish pattern, and as pure (uniform) in all respects as
possible. For this purpose the ‘‘Belgian hare’’ was.
chosen. A buck and two does obtained from Mr. G. W.
Felton, Cliftondale, Mass., were found to breed very true.
From them was bred a stock of does very uniform in
character, twelve of which, together with one of the par-
ents (24), were mated with the selected English buck
which we may henceforth call by his record number §214.
The young thus produced will be called ‘‘Series I’’ off-
spring. About half of them were self (non-English), the
remainder (187 in number) were English.? The latter,
although all undoubtedly heterozygous, varied in white-
ness from grade 1 to grade 4 (Figs. 1-4), the modal or
commonest condition being about the same as that of the
: father (grade 2). The distribution of the young in rela-
tion to our grades is shown in Table I. Statistical treat-
ment of the table gives the average grade of the young as
2.43, that is somewhat darker than the father. Inspec-
tion of the table shows that more than half of the young
are darker than the father, which supports in a general
way the statistical average grade. If we consider sepa-
rately the average grade of the young produced by each

2 The total number of young obtained from 421A, when mated with Bel-
gian hare does, has been to the time of writing 436. The English young
now number 210, the non-English (self) number 226. For Series II mat-
ings presently to be described the corresponding numbers of young are:
English, 219, non-English 196, total 415. For Series I and II combined
the numbers are: English 429, non-English 422, total 851. This is un-
mistakably a 1: 1 Mendelian ratio. -

= 26 THE AMERICAN NATURALIST [Vou. XLIX

mother, we find that it ranges from 2.15 in the case of
?18F, which had 5 English young, to 2.79 in the case of
916D, which had 14 English young. The average number
of English young to a mother is 14.4.

After this series of matings had been completed, a
second series was begun in which the same 13 females
were mated with one of the darkest bucks produced in the
Series I matings (a son of 2916H#). The selected buck was
340A (Fig. 6), grade 3.75, considerably darker than his
father (Fig. 5). This series of matings produced 189
English young, together with a like number of self (non-
English) young. The grade distribution of the English
young is shown in Table I, Series II. All of the 13
mothers except one (9916F) produced darker offspring in
the Series II than in the Series I matings. The lowest
average grade was shown by the young of 917G, viz., 2.44.
For Series I matings the lowest average was 2.15. The
highest average grade in the Series II matings was given
by the young of ?162#, viz., 3.50. For Series I matings
the highest average was 2.78. Consequently, both maxi-
mum and minimum averages were higher in the Series IT
than in the Series I matings. The grand average of all
the 189 Series II offspring was 2.92 as compared with 2.43,
the average grade of the Series I young. Their modal
grade is 3.25. The modal grade for Series I was 2.00.
Since the mothers were identical in both series, the differ-
ence in the young can be attributed only to the difference
in the fathers. The male used in the Series II matings
differed genetically as well as somatically from his father,
who sired the Series I young. Not only was he darker,
but he also produced darker English young. Yet the
father contained only a single dose (one gamete) of Eng-
lish pattern and the son derived his English pattern ex-
clusively from this same source. Hence the English
unit-character had changed quantitatively in transmission
from father to son. This seems to us conclusive evidence
against the idea of unit-character constancy, or ‘‘gametic
purity.’’ If unit-characters are not constant, selection

No. 577] UNIT-CHARACTER CONSTANCY 27

reacquires much of the importance which it was regarded
as possessing in Darwin’s scheme of evolution, an impor-
tance which many have recently denied to it.

TABLE I

SHOWING THE DISTRIBUTION OF GRADES OF OFFSPRING IN THE First AND
SECOND SERIES OF MATINGS FOR EACH INDIVIDUAL MOTHER

| Grades of Young Totals | yess

Mother Series! > | | PEA, 5 = | age
ewes R E EOR E E 5 4.00 Ser. I Ser.II

iča Di EEK Oo Ro ihe aed a 2S Lk ERR ga ear 2.30

II | Sie Ses Bee oh hae ee ee 21 | 3.03

MB rc a a Be a eT eo 1 Fi Seen 2.39

Bas ed a ie ES a hs as 12 | 2.67

16D | I epee 2] 5].../...] 1) 2) 8) 1 ita 2.79

U ikea e e a E Bno.: 12 | 3.19

IOR A Tee akat GR barle A cle i ea 2.29

of: Spgs PA ees ess TER Gi Nae Sey BR oper 5 | 3.50

16F | I |1 Caeo ads bA OE Bares WES 2.62

II | Ee Bae St 2 Se wie i gd Roe aes a 11 2.48

16651 Bed r E Be Bet ah. Wi 2.35

II et ak eA be e p 13 3.06

16H | I dio ee 81 1) Ss) Bt 2 Bt a 8 1 2.76

II rq eeeGn 2 oy Cane kes ge HOM Eini 20 | 3.06

yd ag ke Le a Otay See) 2 2 PD ot BS ig SRR 2 2.36

II 9) 87 G Sy Se a be: 27 | 2.97

17G i í POT Ot 81 St. Sik St aer AE 2.27

be pie ecm ip a ee 2 a oe Et ee ee We et ke ge Been 9 | 2.44

isp | I Pitot aseen TI Rie Iio 2.58

II ie Re ee Sie Bie 8 Sr. ee Da 16 | 2.91

sf | I ao Po e oes clos eee OP a eee OK: 2.15

II ree ie Se Ge ee ae) Oe Oe ie e 16 | 2.97

18H | I Blac are Bi ie al 10 a. 2.43

H Pete ey Bieta ice e aoe 19 | 2.87

RA ay De ee ee OT Bt 2d ors ie chs O Recess | 2.22

II UALS i gan ter es ey a Bfe 8 | 2.78

Totals] I | 1 | 5 |10|18|33/31/24/16/18/11/13| 6| 1 | Tg | 2.43

II 1 5117 13/10912% 18 | 37/27/14} 7 |..... 189 | 2.92

The question whether an imaginary ‘‘unit-factor’’ for
English pattern has or has not changed in correlation
with the visibly changed English unit-character is not
here discussed. We recognize that it has an academic
interest, which, however, scarcely affects the practical
question whether the visible Mendelizing characters of
animals are subject to change through crossing or through
selection or both.

™
CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY
OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT
HARVARD COLLEGE, NO. 256.

ON THE NUMBER OF RAYS IN ASTERIAS
TENUISPINA LAMK. AT BERMUDA!

BY W. J. CROZIER

HARVARD UNIVERSITY

I. It was suggested by Clark (1901) that the average
number of rays borne by Asterias tenuispina was different
for separate localities in Bermuda. He examined forty
specimens of this species—eleven from Coney Island and
twenty-nine from Harrington Sound; in the first set the
average number of rays was 5.91 (I find the mode to be 6),
in the second set 6.93 (with a mode of 7). If this condi-
tion really obtains, it would be an exceedingly interesting
matter to determine the factors responsible for this sort
of difference. I have therefore examined a number of
Asterias (312 in all) from several localities in the Bermu-
das, namely: Agar’s Island, Spanish Point, Hawkins
Island, Ely’s Harbor, Hungry Bay, Harrington Sound
and Coney Island; the first four are situated on the
periphery of Great Sound, the others at widely removed
` points on the north and south shores. For the identifica-
tion of these places, references may be made to the maps
published by Mark (1905).

These observations were made at the Bermuda Biolog-
ical Station, mostly during the summer of 1914.

II. The first lot of starfishes was collected in the imme-
diate vicinity of Agar’s Island in 1913. The number of
rays varied from 2 to 9; the frequency distribution is
given in Fig. 1. The modal number of rays is clearly 7.
In 1914 a collection of Asterias from this place gave the
ray frequency distribution shown in Fig. 2, where the
modal number of rays is again 7. Collections, during
1914, at the other stations named gave the following ray
frequency counts:

1 Contributions from the Bermuda Biological Station for Research, No. 35.

28

No. 577] NUMBER OF RAYS IN ASTERIAS ° 29

Station and Year Pheer, By aol See Figure
Agar’s Isl., 1913 43 7 1
Agar’s s Isl., 1914... 62 7 2
Spanish Point, 1914 33 7 3
w sl., 1914 39 7 4
y's Harbor, 1914 36 7 5
Hungry Bay, 41 7 6
Coney Island, 1914... eien oes 20 7 1
Harrington od 1914 38 7 8

The modal number of rays is in each case 7. This is
true for the same locality in two successive years, for
near-by localities and for places widely enough separated

ole queney

Island - I914.

to yield oiii data relative to the suggestion which
prompted this inquiry. For the total population ex-
amined the ray frequency distribution, which of course
gives a mode of 7 rays, is plotted in Fig. 3. It is to be
noted further that according to Ludwig (1897, p. 345) the
most common number of rays in A. tenuispina from the
Mediterranean is also 7.

III. It has been observed by every one who has studied
A, tenwispina that in most of the individuals the rays oc-
cur in two groups, those of one group being longer than
those in the other, though within each group the rays are
of about the same length. This condition is evident in
259 (83.6 per m of my specimens. There is ee a

30 THE AMERICAN NATURALIST [Vou XLIX

agreement (cf. Ludwig, 1897, and Ritter and Crocker,
1900), that in some cases, if not in all, ‘‘this disparity in
size is due to the regeneration of halves of automatically
bisected animals.” My observations fully confirm this.
I have witnessed, as did Ludwig, several cases of spon-

Fre quency

taneous self-division in the laboratory. The casting off
of one or more rays may at any time be induced by holding
or injuring one or several rays, or by the stimulation of a
ray with dilute acid applied with a pipette. The autotomy
of a single ray takes place very much as described by
King (1898) for Asterias vulgaris ; the existence of a
ss Jo in the region of the ae ae

No. 577] NUMBER OF RAYS IN ASTERIAS 31

ossicle is shown by the fact that even in preserved ma-
terial the rays part very easily in that region.

The relative abundance of cases in which there are evi-
dent two groups of rays of different length indicates that,
as in Linckia (Clark, 1913), autotomous division is a
normal method of asexual reproduction.

FIG. I
istribution ofray Ser ok a bey
H individuals with all fe

Jor nearly equał len gts BSE

Distribvtion of rey frequencies*
in the total population. i
Relatien of long to shert

IV. The numerical relations of the old rays to the re-
generating ones, and the topographical arrangement of
the latter, yield evidence relative to certain questions in
the physiology of regeneration.

(a) It is to be observed that the regeneration in ques-
tion has taken place apart from experimental control;
therefore information as to the number of rays usually
present just before Asterias undergoes self-division must

deduced from the data at hand. The modal-ray fre-
quency for specimens with rays of very nearly equal —
length is 7 (Fig. 10), but it is a question whether this ap-

pearance of equality in ray length may not be due to a — .

32 THE AMERICAN NATURALIST [Vou XLIX

variety of conditions, especially the rapid growth of re-
generated rays. The regenerating rays of Linckia (Clark,
1913; Monks, 1904) and the newly formed rays of multi-
radiate types (Ritter and Crocker, 1900) grow more rap-
idly than the old ones and soon reach the dimensions of
the latter; this is also indicated in my series. But the cor-
relation of the number of long with the number of short
rays (using only those cases in which the two groups were
clearly distinct) makes it evident (Fig. 11) that the condi-
tion in which there are 3 or 4 long rays and 4 short ones is
by far the most common; and further, that the cases in
which there are either 3 or 4 long rays are almost equally
abundant. It seems not unlikely, then, that A. tenuispina
usually has 7 rays before it divides, and that it divides
into two parts having, respectively, 3 and 4 rays, the divi-
sion-surface then giving rise, in the greater number of
cases, to 4 new rays, but sometimes to 5, 3 or even 2.

If all the individuals observed had undergone autotomy
and regeneration according to this scheme, then those
with 7 and those with 8 rays would be expected to occur
in equal abundance; 8 is next in frequency to 7, but the
latter preponderates because some starfishes have prob-
ably not autotomized at all, and because all the animals
which have divided do not adhere to this paradigm (see
Fig. 11). Yet, in the majority of cases, 4 rays are regen-
erated whether there are 2, 3 or 4 long (old) rays in
evidence.

It would seem that self-division may occur at any time
in the life history of A. tenuispina, or at least in animals
of all sizes, though it is my impression, gained from hand-
ling many live individuals, that the smaller (younger ?)
ones autotomize more readily than larger ones. Those
showing two distinct ray groups ranged in longer ray
length from 11 mm. to 65 mm. There is no evidence that
autotomous divisions follow one another rapidly, or in-
deed that they occur more than once in any given indi-
vidual.

One case was observed in which there was one long ray
only, and 6 shorter ones. This may mean that a single
ray can regenerate the whole body, as suggested by v.

No. 577] NUMBER OF RAYS IN ASTERIAS 33

Martens (1866, quoted by King, 1898) for this species.
I have not been able to substantiate this idea by laboratory
experiments, for, in my tests, single isolated rays did not
live more than a few weeks.
(b) Newly forming rays have a tendency to appear in
symmetrically disposed pairs (see Fig. 12), which gives to
Se FIG, 12.

Ase Wat

DIAGRAMATIC REPRESENTATION OF THE RAYS OF Four ASTERIAS, SHOWING TEND-
Y OF RAYS TO APPEAR IN Parrs. Measured from the mouth along
the ventral side.

X e X

FIG,.13.
RB Raia THE RELATIONS OF New Rays.

many individuals a strikingly bilateral aspect. This is
accentuated by their behavior, for, in the absence of
directive stimuli, they commonly move with the longest
rays in advance. In moving away from the light, the loco-
motor movement of the group of longer rays also tends,
in many cases, to produce a spurious ‘‘orientation.’’
When placed oral side up, the larger rays exert a de-
termining influence on the direction and manner of right-
ing. These effects are due to the greater pedicel and
muscle development of the longer, thicker, rays.

e formation of two rays at a radial cut on the disc
was found by King (1900) in Asterias vulgaris.

V. I have suggested, above, that Asterias with 7 sub-
equal rays have probably arrived at that condition by
different routes. One method of ray multiplication ap-
pears to be the spontaneous addition of new rays at any
point on the disc. Twelve starfish were found which
showed but one ray markedly shorter than the others.

34 THE AMERICAN NATURALIST [Von. XLIX

Of these, 4 had 5 long rays, 4 had 6, 2 had 7, 1 had 3 and
1 had 4 (see Fig. 11). The addition of new rays during
adult life is, so far as known, unusual among starfish, ex-
cepting in the multiradiate forms (cf. Ritter and Crocker,
1900; Clark, 1907; M’Intosh, 1907). The twelve cases
found in A. tenuispina may mean merely that a single ray
has been cast off and is being regenerated, for there is
found about the same percentage of naturally occurring
regenerating examples of A. vulgaris (King, 1898; 1900).
Yet I am inclined to interpret this condition as indicating
the way in which the modal hepta-radiate form is derived
from the fundamental penta-radiate one, or from a hexa-
radiate plan, if the young of A. tenuispina be like the post-
larve of Pycnopodia (Ritter and Crocker, 1900) previous
to self-division.* The three smallest Asterias seen had
6 rays. These were subequal and 8+ mm. long. Other
specimens, slightly larger, had either 7 or 8 rays.

Cases such as those illustrated in Fig. 13 may further
prove that addition of new rays occurs independently of
the reformation of rays subséquent to self-division. :

VI. The number of madreporites in A. tenuispina is
also variable, as noted by Ludwig (1897, p. 358) and
others. The number of madreporic bodies is certainly not
correlated with the size of the starfish. One of the small-
est ones seen had 8 rays and 5 madreporites, its mean ray
length being 10 mm.; while the largest animal collected
had 5 rays, with a mean ray length of 70 mm., and but one
madreporite. The table in Fig. 14, which includes all
cases in which the madreporites were counted, shows that,
while the distribution of these bodies is irregular, their
number is to some extent correlated with the number of
rays. Ludwig gave it as his opinion that there was no
correlation of this sort. The relation stands out more
clearly if only those individuals having equal rays (and
therefore presumably ‘‘full grown’’) are included (Fig.
15). Unfortunately, the number of animals is small.

Multiple madreporites were noted in 5 out of 101 ex-

3 According to Clark’s (1907) studies, the young Heliaster has five rays
only; his results throw considerable doubt upon the correctness of the
conclusions of Ritter and Crocker.

No. 577] NUMBER OF RAYS IN ASTERIAS 35

amples. Three of these showed a condition which might
have arisen either by the fusion of two plates or by the

division of a single one. The other two cases were simi-
lar, but of trefoil form. Dissection showed, in each in-
stance, that a single stone canal was present. Therefore
these multiple plates had probably arisen by the division
of an originally single one. (For a similar condition in
A. vulgaris, see Davenport [1901].) Only one multiple
madreporite was found in any one individual.

SUMMARY

1. The modal number of rays in Asterias tenuispina
is 7. The range in ray number is from 2 to 9.

2. The 7-ray condition is uniformly the most frequent,
even in widely separated localities.

3. The modal ray number is the same for animals with
subequal rays as for those with a group of regenerating
rays, 43 |

4. The evidence indicates that, most commonly, A.

36 THE AMERICAN NATURALIST [ Von. XLIX

tenuispina has 7 rays before it undergoes autotomy, that
it divides into 3-ray and 4-ray portions, and that each of
these parts regenerates 4 rays.

5. Regenerating rays tend to appear in bilaterally dis-
posed pairs, as regards size.

6. There is no evidence that self-division occurs often
in the life of individuals, though possibly it does.

7. New rays may be added at any point on the disc.

8. The number of madreporites varies from 1 to 5, and
is to some extent correlated with the number of rays; it is
not correlated with the size of the animal.

9. Double or triple madreporites occur in about 5 per
cent. of the individuals.

REFERENCES

Clark, H. L. 1901. Bermudan Echinoderms. A Report on Observations
and Collections Made in 1899. Proc, Boston Soc. Nat. Hist.,
Vol. 29, No. 16, pp. 339-3.
1907. gs kojen of the Genus Hetioster. Bull. Mus, Comp. Zodl.,
1, pp. 23-76, 8 pls.
1913. cate in Linckia. Zool. Anz., Bd, 42, No. 4, pp. 156-159.
Davenport, G. C. 1901. Variation in the Madreporic Body and Stone
Canal of Asterias vulgaris. Science, N. S., Vol. 13, pp. 374-
375

King, Helen D. 1898. Regeneration in Asterias vulgaris. Arch. f. Entw.-
mech., Bd. 7, Heft 2-3, pp. 351-
1900. Further Studies on Regeneration in . ateriat vulgaris. Arch.
f. Entw.-mech., Bd. 9, Heft 4, pp. 737.
Ludwig, H. 1897. Die Saciigins des Mittelmeeres. Fauna u. Flora d.
Golfes v. Neapel, Monogr. 24 x + 491 p., 12 Fig. and 12 Taf.

M’Intosh, D. C. 1907. Meristic Variation in the Common Sun-Star (So-
laster papposus). - Proc. Roy. Phys. Soc. Edinburgh, Vol. 17,
pp. 75-78.

Mark, E. L. 1905. The Bermuda Islands and the Bermuda Biological
Station for Research. Proc. Amer. Assoc. Adv. Sci., Vol. 54,
pp. 471-501, 16

Martens, E.von. : 1866. ela ostasiatische Echinodermen. Arch. f.
Naturg., Jahrg. 32, Bd. 1, pp. 57-88.

Monks, Sarah P. 1904. Variability and Autotomy of Phataria1 [Linckia].
Proc. Phila. Acad. Nat. Sci., Vol. 56, pp. 596-600.

Ritter, W. E., and Crocker, G. R: 1900. Multiplication of Rays and Bi-
Jateral Symmetry in the 20-rayed Star-fish, Pyonopodia heli-
anthoides (Stimpson). (Papers from the Harriman Alaska
Expedition, III.) Proc. Wash. Acad. Sci., ‘Vol. 2, pp. 247-274,
pl. 13, 14.

_ 1‘*Phataria’’ is an error, as pointed out by Clark (1913).

SHORTER ARTICLES AND DISCUSSION

MR. MULLER ON THE CONSTANCY OF MENDELIAN
FACTORS

In discussing the selection experiments of Phillips and myself
with hooded rats,1 Mr. Muller? accepts the explanation of ‘‘ modi-
fying factors ’’ which we offered to account for certain peculiar
results obtained, but rejects the idea which we also suggested,
that the chief genetic factor concerned may be undergoing quan-
titative variation. He rejects it on the ground that this ex-
planation is not ‘‘ in harmony with the results of Johannsen and
other investigators.’’ The work of Johannsen with seed-size in
beans and the work of others with Drosophila is cited in support
of this statement.

It is difficult to understand how the experiments of Johannsen
have any direct bearing on the case since no single Mendelizing
unit-factor was demonstrated in that connection; but in the
hooded pattern of rats a Mendelizing unit-factor is unmistakably
present and it is the quantitative variation of this which is under
discussion, not the presence of many or few additional factors,
concerning which Muller adopts our explanation. Appeal to the
work of Johannsen with bean-size to show that our conclusions
concerning color pattern in rats are incorrect is illogical because
the cases are not parallel. The citation by Muller of the work
on rabbit-size by MacDowell and myself? is equally non-germane,
because no demonstrable Mendelizing unit-factor is involved in
that case either. He might with propriety cite the bean work as
bearing on the interpretation of the inheritance of body size in
animals, or vice versa, since both involve blending inheritance.
But. neither of these cases has any direct bearing on the question
of unit-character constancy, since in neither case has a unit-
character, either constant or inconstant, been shown to exist.

The citation of work with Drosophila is more to the point, since
the ‘‘ mutations’’ of Drosophila Mendelize. But is it certain
that they do not vary? Muller admits that they do occasionally
vary, stating that ‘‘ in one case (possibly in two or three cases)

1 Castle and Phillips, ‘‘Piebald Rats and Selection,’’ Publ. No. 195, Car-
negie Institution of Washington.

2 AMER. NAT., Vol. 48, p. 567.

8 Publ. No. 196, Carnegie Institution of Washington.

37

38 THE AMERICAN NATURALIST [ Von. XLIX

a locus has mutated three times, each time in a different way.’’
He does not think that smaller changes than these have occurred,
since ‘‘ much smaller could easily have been detected.’’ From this
statement I infer that the opinion rests on casual inspection
rather than measurement, for which reason I do not attach much
importance to it. The hooded pattern of rats was not supposed
to vary quantitatively until its quantitative study was under-
taken. Two types of hooded rats were recognized, one more ex-
tensively pigmented than the other, and these were supposed to
be discontinuous like the several ‘‘ mutations of a locus’’ in
Drosophila. Quantitative study has completely dispelled this
idea as regards the hooded pattern of rats, and I have no doubt
the same would be true of Drosophila. How easy it is to be sure
of a thing which has not yet been investigated, so sure that in-
vestigation of it is considered a waste of time. Muller is con-
fident that such variation as occurs in Drosophila ‘‘ can not even
remotely be compared to fluctuating variability,’? and he gen-
eralizes thus:

“Tn no known case do the variations of a gene among, let us say,
several thousand immediate descendants of the individual possessing it,
form a probability curve.”

The use of the word ‘‘ gene ’’ in this sweeping statement safe-
guards the author, since no one, so far as I know, claims ever to
have seen a ‘‘ gene’’ or to have measured it. How could the
‘* variations of a gene’’ be expected to ‘‘ form a probability
eurve’’ if the gene is not measurable? But if the author will
allow the substitution of visible character for ‘‘ gene’’ in his
challenge, I will gladly accept it and I will add this generalization
for his consideration—No one has by actual observation and
measurement shown the existence of any visible character in any
animal which is not quantitatively variable.

As regards the mutations of Drosophila which Muller is con-
fident (apparently without having studied the matter himself)
do not vary so as to form a probability curve, I had sufficient
curiosity some months ago to suggest a quantitative study by one
of my pupils, Mr. D. H. Wenrich. Mr. Wenrich:studied the

wing-length of flies from a culture kindly supplied me by Pro-
fessor Morgan under the name “‘ vestigial.” In advance of a
more detailed publication, Mr. Wenrich kindly permits me to
state the following facts. The wing length measured in ocular
micrometer units was found to vary as follows:

No.577] SHORTER ARTICLES AND DISCUSSION 39

eee 25-29 30-34 35-39 40-44 45-49
Tensy Seay ee rs 6 34 67 43 3

Dep nhl R E E N Sawa S 50-54 55-59 60-64 65-69
Hume OS honey eae 1 1 0 1

The wing-length manifestly varies so as to form a pretty good
probability curve; what the ‘‘ gene’’ is doing, I do not under-
take to say.

It is, of course, conceivable that the variation here observed in
actual wing length might be due to variation in general body
size, larger flies having longer wings. To determine this point
measurements of tibia-length were made on the same flies, and
in the case of each individual the ratio was computed between
wing-length and tibia-length. These ratios are distributed as
follows: :

HAVIOR ssi. .70—.79 .80-.89 .90-99 1.00-1.09 eta 1.20-1.29
Frequencies 2 7 26 49 23

e SVU wks saute 1.30-1.39 1.40-1.49 1.50-1.59 1. wa 69 1.70-1.79
Frequencies ........ 7 3 0 0 2

It is evident that there is no constant relation between wing-
length and tibia-length, and so between wing-length and general
size, with which tibia-length is closely correlated. "App we ob-
tain a good probability curve. Does the ‘‘ gene’’ vary or are
we dealing also with additional modifying ‘‘ genes’’? We are
confronted here with the same problem as in the case of the rats.

But it is possible to assume that the considerable variation
shown by vestigial wings in Drosophila is purely somatic, ‘‘ phe-
notypic,’’ not due to genetic causes, and so would not show any
effects if subjected to selection. So it was thought in the case of
the plus and minus variations in the hooded pattern of rats,
before the experiment was made, but experiment has shown, even
to Mr. Muller’s satisfaction, that the variations are in part due
to genetic causes and that selection slowly and surely changes
the range of variability. Is it safe to assume the contrary for
Drosophila in the absence of all experiment?

Mr. Wenrich has also studied the wing-length of ‘‘ extracted ”’
vestigial flies obtained in the second generation from a cross
between pure vestigials and normal flies, and he finds that the
variability is regularly increased as compared with that of the
uncrossed vestigial race. This again is parallel with what occurs
when hooded rats are crossed with wild or with Irish rats, and
indicates that similar causes are at work in the two cases. Such

40 THE AMERICAN NATURALIST [Vou. XLIX

cases present to the genotype theory the following dilemma.
Hither one gene is concerned in the case or many genes. If one
only is concerned, it is variable. If many genes are concerned,
they are so numerous (whether or not constant) that they present
to the observer of the visible character affected a continuous
variation series, one capable of indefinite displacement up or
down the quantitative scale. The supposed distinction between
continuous and discontinuous variation then vanishes. Selec-
tion in that case meets with no “‘ fixed limit ” beyond which it
cannot go.

Mr. Muiler is seriously disturbed (p. 573) because we are will-
ing to consider it possible that the ‘‘ factor for hooded ”’ may be
contaminated by ‘‘ its allelomorph (the factor for self) ’’ while
associated with it in the zygote represented by the F; rats. (The
evidence of modification is unmistakable, however one at-
tempts to explain it.) He says this is ‘‘ violating one of the most
fundamental principles of genetics—the ‘non-mixing of factors—
_in order to support a violation of another fundamental prin-
ciple—the constancy of factors.” Now, when, I should like to
inquire, did these principles become ‘‘ fundamental ”?; by whom
were they established and on what evidence do they rest? I
should suppose that Bateson, president of the British Association,
might be considered fairly well posted on the ‘‘ principles of
genetics,’’ but neither in his earliest papers nor in his latest do
we find any mention of these sacred principles. In his recent
presidential address* he frankly states his belief that segregation
is often imperfect and that ‘‘ fractionation ” of factors fre-
quently occurs as a result of crossing.

We shall look in vain, I think, for those ‘‘ principles ’’ outside
of the ‘‘Exakten Erblichkeitslehre’’ (or its imitations), and when
we inquire as to the experimental basis of the principles in ques-
tion we are met with the satisfied reply, ‘‘Johannsen’s beans.’
What a slender basis and what and absurd one from which to
derive the ‘‘ fundamental principle ’’ that Mendelian factors are
constant! Yet to date this case, which admittedly involves no
clear Mendelian factor, is the only evidence worth mentioning in
favor of the constancy of Mendelian factors! Do biologists take
themselves seriously when they reason thus? Certainly no one
else will long take them seriously.

Finally, I may be permitted to correct two misapprehensions

4 Science, August 28, 1914,

No.577] SHORTER ARTICLES AND DISCUSSION 41

into which Muller in common with the Hagedoorns® has fallen,
viz., (1) that individual pedigrees were not recorded in the
course of our selection experiments and (2) that no considerable
amount of inbreeding occurred in our work.

It has been our invariable practise, upon recording the birth of
an animal and its grade, to record on the same line of the ledger
the record number of its mother and father. This enables one in
any particular case to trace back the pedigree to the very begin-
ning of our experiments. We have spent much time writing out
and studying individual pedigrees, but without discovering any
evidence of pure or prepotent lines or individuals, except in a
single case, that of our ‘‘mutant ’’ series, the origin and complete
history of which we have described in detail. The pedigrees,
however, of our rats are on record available for study at any
time; their full publication would be a quite impossible under-
takin

That extensive and intensive inbreeding has occurred in our
experiments will be obvious when I state that all our animals
were descended from a very small initial stock, less than a
dozen individuals, that from the beginning we have made the
most extreme selections possible, mating like with like, never hesi-
tating to mate brother with sister, and putting aside for strict
brother-sister matings any litter of young which seemed espe-
cially promising. I may say that in no single case (except that
of the ‘‘ mutant ”’ series) have these ‘‘ special’’ pens given us
advancement obviously greater or less than that of the general
selection series of which they formed a part. Nevertheless, we
are still continuing to follow them up and will later publish a
detailed account of them. Finally I would call attention to pp.
20 and 21 with Tables 48—49 of our full publication, in which
` are described the hooded offspring of a single selected hooded
and a single wild rat. The hooded and the wild rat produced
several young resembling the latter, that is, not hooded; these `
were mated inter se, brother with sister. Among the grand-
children (F,) occurred the usual 25 per cent. of recessives,

hooded. Two males were selected from these and mated with
females of as nearly the same grade as were available. This
process was repeated through seven generations in succession.
Seven times animals of like grade were mated together, brother

5 Zeit. f. ind. Abst. u. Vererbungslehre, 11, p. 145. See also my reply in

the same journal, 12, p.

42 THE AMERICAN NATURALIST [Vou. XLIX

with sister when possible, less often brother with half-sister,
rarely cousin with cousin. In this way were obtained 804 young
from rigidly selected, closely inbred descendants of a single pair
of rats, the series extending into generation F,. We have shown
(l. c., p. 21) that the progress of selection within this inbred
family follows a remarkably close parallel, generation by gen-
eration, to the progress of selection in our plus series as a whole.
Muller’s anticipation that a different result would follow close
inbreeding is not justified by our observations.

In discussing this experiment (p. 21) we have italicized the
statement that (so far as the hooded character is concerned) the
entire series is derived from a single hooded individual! When
the Hagedoorns made the statement that our stock had not been
sufficiently inbred, they had apparently not seen our full pub-
lication and so had no means of knowing to what extent it had
been inbred, but Muller, with our full publication before him,
apparently repeats the statement without taking the trouble to
verify it.

. W. E. CASTLE

BUSSEY INSTITUTION,

October 23, 1914

NO CROSSING OVER IN THE FEMALE OF THE
SILKWORM MOTH

IN a recent review’ of a paper by Y. Tanaka? on linkage in the
silkworm moth, I pointed out that some of his data suggested
that crossing over was occurring in only one sex. While the data
were not sufficient to establish this conclusion, there was at this
time another paper by the same author*® which I had not seen.
In this paper are presented data which clear up the matter.

Tanaka has now made back-cross tests of both sexes. That
crossing over does occur in the males was shown by the mating
sysy 2 X SYsy 3 which gave a total of 865 cross-overs among
2,907 offspring. The cross sysy 2 X SysY g gave 151/488 as the
proportion of cross-overs. But when females were tested,
SYsy 2 X sysy ¢ gave no cross-overs in 1,183 offspring. Tanaka
refers to another paper, apparently in press, in which he has
shown the same relations (i. e., crossing over in males, none in

1 AMER. NAT., XLVIII, 1914.

2 Jour. Coll. Agr. Tohoku Imp. Univ. Sapporo, V, 1913.

8 Jour. Coll. Agr. Tohoku Imp. Univ. Sapporo, VI, 1914.

No. 577] SHORTER ARTICLES AND DISCUSSION 43

females) for the combinations NynY and MYmy. As stated in
my former review, there was in the earlier paper a record of the
mating sysy 2 < SysY 4, giving no cross-overs in 128 offspring.
Tanaka now says, referring to this case: ‘‘ Whether there may
exist, in certain occasion, a complete reduplication [linkage] in
male, or whether the above result is due to any mistake by which
sex-signs have been reversed, is at present uncertain. No similar
case has as yet been found in other families.”’

The evidence seems to make it highly probable that crossing
over in the silkworm moth occurs only in the male; a surprising
result when we remember that in Drosophila it occurs only in the
female. One is immediately reminded that in Drosophila the
male is heterozygous for the sex-differentiator, while in Abrazas
and probably all moths the female is the heterozygous sex. These
facts are highly suggestive, and lead one to wonder what will be
found with regard to crossing over in the two sexes in birds and
mammals, where similar differences in sex-determination occur.
Another point worth noting in this connection is that in the
hermaphroditic sweet pea and Primula crossing over occurs in
the formation both of pollen and of ovules.

Tanaka reports two cases of aberrant results which, as he says,

-may be explained as due to mutation (‘‘dropping out’’) of S in
one case, and of both S and Y in the other. He adds that such
an assumption is premature. To the writer it seems more prob-
able that the females involved were not virgin. The results are
easily explained on the assumption that they had paired with
brothers before isolation, since brothers of the necessary composi-
tion are shown by the pedigrees to have been present in each case.

Another interesting point brought out by Tanaka’s more re-
cent paper is the relation between the larval patterns known as
striped, moricaud, normal, and plain. In my earlier review I
followed Tanaka in treating these patterns as affected by three
pairs of genes: S (striped) and s, M (moricaud) and m, and
N (normal) and n, plain being the triple recessive. The same
scheme has been followed in the early part of this paper. On
this assumption, as Tanaka points out, it is necessary to suppose
that complete linkage occurs between these three pairs of genes.
The evidence need not be gone over in detail here, but there are
over 10,000 larve recorded from various tests of this relation,
without a single cross-over among them. Although Tanaka does
not mention the point, this at once brings up the possibility that

44 THE AMERICAN NATURALIST [Vor. XLIX

we may be dealing with a system of multiple allelomorphs, No
two of the types when mated together give a third in F,; and,
unless one or both carry a recessive in heterozygous form, any
two types give a 3:1 ratio in F,, or 1:1 on back-crossing to a
recessive. The four patterns involved seem, from the descrip-
tions, to fall roughly into a series in the order striped, moricaud,
normal, and plain, That is to say, the second two are rather
intermediate in appearance between striped and plain. Al-
though I believe any arguments as to the nature of genes which
are based on the appearance of characters are open to very seri-
ous objections, it must still be admitted that the different char-
acters involved in a case of multiple allelomorphism are gener-
ally of the same sort.*

On the chromosome view, if the genes just discussed are allelo-
morphs they occupy identical loci in homologous chromosomes.
If they are not allelomorphic but closely linked, they occupy
different but closely opposed loci in homologous chromosomes.
In either case, any combination of them should give approxi-
mately the same linkage to the Y-y pair of genes, which occupy
a locus in the same chromosome, but some distance away. The
linkage of the striped-normal, striped-plain, and moricaud-plain
combinations with the Y—y locus appear from Tanaka’s data to
be in fact about the same, though the data on the first (striped-
normal) are the only ones sufficiently large to be very significant.

A. H. STURTEVANT

COLUMBIA UNIVERSITY,

October, 1914

THE INFLUENCE OF POSITION IN THE POD UPON
THE WEIGHT OF THE BEAN SEED

IN a note on the pure line problem Belling’ has emphasized the
significance of position in the pod as a factor in determining the
weight of the bean seed. Since this point in his paper seems to
have attracted some attention among those interested in genetics,
it may not be out of place to call attention to a series of quanti-
tative determinations of the intensity of the relationship? and to
illustrate the results secured.

If one numbers the successive ovules of the pod from 1 up,

4I have discussed this aspect of the matter briefly in another paper
(Amer. Nar., XLVII, 1913, p. 237)

1 Belling, J., ‘‘Selection in Pure Lines,’’ Amer. Breed. Mag., 3: 311-312,

912.
2 Harris, J. Arthur, ‘‘A Quantitative Study of the Factors Influencing the

No.577] SHORTER ARTICLES AND DISCUSSION 45
he may regard the numbers as measures (in units of intervals
between adjacent ovules) of the distance of ovules from the

Scale of Mean Weight.
LOTTIE TTT TTT TT TTT TTT Tr

oe p ces Bes eo pe
+ A it
sss
Alad
A

4 OVULES 5 OVULES 6 OVULES 7 OVULES
FIG. 1.

N

v

o

Position in Pod.
>

>

~

proximal end of the pod, and may then express in terms of corre-
lation the prenan i between the T of the seed and its
position in the p

Scale of Mean Weight.

O Ly Oh

N

>

ee S
N
Ñ
\

n

\
aN
\

X ÍI a
ANE X

\

\
\

Position in Pod.

>

Ea |

~

4 OVULES | 5 OVULES | 6 OVULES | 7 OVULES 8 OVULES

Fia. 2.

In doing this, the pods should of course be sorted into classes
according to the number of ovules which they produce and the
relationship computed for each group of pods separately, for
there is no reason for believing that the fourth in a pod with 4
ovules is comparable with the fourth seed in a pod with six. This

Weight of the Bean Seed. I. Intra-Ovarial Correlations,’’ Beih. Bot. Cen-
tralbl. Abt., I, 31: 1-12, Pl, 1-4, 1913,

46 THE AMERICAN NATURALIST — [Vou. XLIX

has now been done for twenty series of pods, drawn from five
cultures belonging to three distinct varieties (Navy, Golden Wax
and Burpee’s Stringless) and embracing altogether about 23,000
individually weighed seeds. In every one of these cases a positive
correlation has been found, i.e., the weight of the seed increases
as its distance from the base of the pod becomes greater. The
intensity of this interdependence is, however, not very great, at
least in the varieties so far studied. The correlations range from
014 + .046 to .238 + .068, with an average value of about .132,
or about 13 per cent. of perfect correlation.

The rate of change has been expressed by the slope of a straight
line for four different classes of pods studied for a culture of
Navy beans made at Sharpsburg, Ohio, in 1907 (Diagram 1°) and
for five classes from a culture of Burpee’s Stringless beans grown
at the Missouri Botanical Garden in the same year.

In the first of these, the

? Navy series, it appears that

P- ANS | | the observed mean weights at
PA ASAS ae first increase rather rapidly,

E A ei Se | then the rate of increase falls
HEDE SNS | off and finally the seeds near- -
TA _ est the tip (distal or ‘‘blos-
ef’ af Pais - som” end of the pod) become
Ai aot ‘aoe somewhat lighter than those
a fe E a little lower down. Here a
ei a Ao curve would fit the observed
R: E ee means better than a straight
AA ies line. In the Burpee’s String-
rA r Se oe less culture (Fig. 2) however,
i7 12 FA aaa the change in seed weight can
oe Sten maa k < for all practical purposes be

| ify represented by a straight line

ify on ES as well as by any curve.

ita The percentage of ovules
Jil which develop into seeds also

he Ba Tome increases from the base to-
ig es = ward the stigmatic end of the
Ke pod. In small pods the rate

Fic. 3. of increase may be fairly

regular, but in larger pods
s Ta the diagram for this series published in the original paper there is a
slip in the representation of the slope of the line for pods with 4 ovules.

No.577] SHORTER ARTICLES AND DISCUSSION 47

it falls off toward the stigmatic end, where the fecundity
may be even lower than it is a little farther down in
the pod. This is admirably shown in Diagram 3, in which GG
stands for a series of Burpee’s Stringless grown at the Missouri
Botanical Garden, NH for a series of Navy beans grown at
Sharpsburg, Ohio, and LL for a series of Golden Wax beans
grown at Lawrence, Kansas. All were grown in 1907. Here the
percentage of development of ovules at different positions in the
pod is shown for the different classes of pods by the scales to the
left of the figures. The reader may ascertain the class of pods
represented by any particular curve by noting the number of
circles representing percentage development in the various posi-
tions. These correspond to the number of ovules per pod. In the
diagrams the positions (abscisse) from left to right represent the
positions from the base to the tip of the pod.
J. ARTHUR Harris

ANOTHER GENE IN THE FOURTH CHROMOSOME OF
DROSOPHILA

UNTIL the appearance of bent wings, only three groups of
linked genes had been found.in Drosophila amelophila, although
four pairs of chromosomes had been identified in the diploid
group. Since the character bent wings, worked out by Mr. H. J.
Muller, was found to be unassociated with any of the three
groups, the gene producing this character was said to be located
in the fourth chromosome.

Recently a new character, designated as eyeless, appeared.
Flies having this character either lacked eye pigment and om-
matidia or had one or both eyes reduced in size. All of the pure
stock showed some loss of eye structures. Eyeless is recessive to
the normal eye. In order to determine the linkage, eyeless males
were crossed in turn to females of the stocks at Columbia Uni-
versity. These stocks representing the three groups were (1)
miniature wings, (2) black body and vestigial wings, and (3)
spread wings. The genes producing these characters are in the
first, second and third chromosomes, respectively. The F, from
all three crosses had normal eyes. They were inbred in each
case and gave the following..

The equation should be w==9.987 + .021 p. The line as it appears here is
_ correctly drawn.

48 THE AMERICAN NATURALIST [Vou. XLIX

Cross 1. Miniature 9 by Eyeless g
F; Normal Long Normal Miniature Eyeless Long Eyeless Miniature
1142 0 245 193

Since the eyeless flies were females as well as males, the character
eyeless is shown not to be a sex-linked character; for, if it were,
it would be inherited only by the grandsons of the eyeless male.
Since the eyeless flies are not nearly as viable as the wild stock,
the eyeless classes fall below the expectation.

Cross 2. Black Vestigial 2 by Eyeless g
F.» Normal Long Normal Vestigial Eyeless Long LEyeless Vestigial
1303 417 $ 278 97

The same count, when grouped according to the body color, was
as follows:

F, Normal Gray Normal Black Eyeless Grey Eyeless Black
1289 431 - 293 82

Cross 3. Spread ? by Eyeless ¢
F.. Normal not Spread Normal Spread Eyeless not Spread Eyeless Spread
1349 373 300 76

Allowing for the decreased viability of eyeless, both of the pre-
ceding crosses may be regarded as 9:3:3:1 ratios. Hence they
show that there is no linkage of eyeless with the characters whose
genes are in the second and third chromosomes.

Eyeless females were then crossed to bent-winged males (Cross
4). No bent eyeless flies were produced in the F,. As the count
was small, the F, bent flies were crossed to the F, eyeless, and
then the F, normal, which had the same germinal constitution as
the F,, were inbred to give F,, which should give the same re-
sults as the F..

Cross 4. Bent g by Eyeless ?
Normal not Bent Normal Bent- Eyeless not Bent Eyeless Bent
596 193 195 0

a

F. 741 172 131 0
Total 1337 365 326 0

Since an approximate 2:1:1:0 ratio, instead of a 9:3:3:1
ratio, was realized, the conclusion that eyeless and bent belong

No.577] SHORTER ARTICLES AND DISCUSSION 49

to the same group and in this sense may be said to be in the
same chromosome pair is evident. Until a bent eyeless fly—a
cross over—is obtained, the amount of crossing over between
these two characters in the fourth chromosome can not be directly
determined.
MILDRED A. HOGE

AN ABNORMAL HEN’S EGG

In a frequently quoted paper, Parker (’06) has classified
double eggs on the basis of the factors supposedly concerned in
their formation. Considering the ovarian and oviducal factors as
independent, Parker says: `

As a result of these two factors, three classes of double eggs can be
distinguished; first, those whose yolks have come from an abnormal
ovary but have passed through a normal oviduct; secondly, those whose
yolks have come from a normal ovary but have passed through an ab-
normal oviduct; and finally those produced by an ovary and oviduct
both of which have been abnormal in their action.

Fig. 1. PHOTOGRAPH OF THE SPECIMEN X 1.

Cases of ovum in ovo have been attributed by Parker and
others to antiperistalsis. Patterson (’11) mentions a case of an
inclosed double egg in which there were two distinct peristaltic
actions. Féré (’98) has called attention to the fact that hens fre-
quently lay several double eggs in succession. Féré claims that
he succeeded in producing double eggs in a hen which normally
laid single eggs, by drugging her with atropine sulphate. Glaser
(713) has described the ovary of a hen which habitually laid
double eggs and coneludes that fusion of the follicles is the expla-
nation of some double eggs.

50 THE AMERICAN NATURALIST [Vou. XLIX

The case which I wish to record is very similar to that figured
by Hargitt (’12) and termed by him a ‘‘gourd-shaped’’ egg. Un-
fortunately, the egg which Professor Hargitt studied was not
preserved carefully and on account of evaporation, the condition
was such that he could not be certain of the presence of yolk in
the smaller end. He assumed that the egg was comprised of

Fig. 2. DIAGRAMMATIC MESIAL VIEW OF THE ABNORMAL EGG, SHOWING THB
ELATION OF THE YOLK TO THE ALBUMEN,

about normal parts in the larger end, and that the smaller con-
sisted of only albumen, ‘‘its yellowish tint having resulted from
the evaporating process which had taken place.”’

The egg shown (Fig. 1) was presented to Professor Julius Nel-
son, of Rutgers College, several years ago and was carefully pre-
served in a jar of alcohol. The result was that although the ac-
tion of the alcohol had partially decolorized the yolk, it was pos-
sible to trace it throughout the entire extent with no difficulty.
As can be readily seen from the photograph, that part of the egg
which might be termed the ‘‘neck’’ presented a much roughened
appearance from the excessive accretion of lime. A nodule of
lime at the smaller end of the shell would seem to indicate that
the last deposit of the shell glands was there received.

For convenience in examination of the irregular shaped egg, it
was separated at the circling line seen in Fig. 1, and then the two
parts were halved with a sharp scalpel, after the penetration of
the shell by means of scissors.

When the first separation was made at the line indicated, one
could readily discern the presence of a constricted yolk sur-
rounded by apparently normal albumen. Examination of the
halved portions showed that the yolk extended from the larger
end through the constricted region to occupy a position approxi-
mately normal in the smaller end. It seems possible that this
particular abnormality may have been caused by a constricted

No.577] SHORTER ARTICLES AND DISCUSSION 51

oviduct rather than from the fusion of two eggs during ap-
position, induced by anti-peristalsis.?
F. E. CHIDESTER
RUTGERS COLLEGE.

LITERATURE CITED

Féré, C.
’98. Deuxième note sur le devellopment et sur la position de 1’embryon
de poulet dans les oeufs a deux jaunes. C. R. de Soc. de Biol.,
1898, p. 922.
Glaser, O.

713. a the Origin of Double Yolked Eggs. Biol. Bull., Vol. 24, pp. 175-
186.
Hargitt, C. W.
99. Some Interesting Egg Monstrosities. Zool. Bull., Vol. 2, pp.
225-229.

712. Double Eggs. AM. NAT., Vol. 46, pp. 556-560.
Parker, G. H.

706. Double Hen’s Eggs. Am. Nart., Vol. 40, pp. 13-25.
Patterson, J.

TLA Double Hen’s Egg. Am. Nart., Vol. 45, pp. 54-59.

1Since the above was written, but before the proof came to hand, an
authoritative paper has been published (Maynie R. Curtis, Studies on the
Physiology of err kaeee in the Domestic Fowl, Vi. Double- and Triple-
Yolked Eggs. Biol. Bull. Vol. 26, pp. 55-83.) in which no mention has been
made of the possibility of incomplete separation of both yolk and albumen
of a single egg. Evidence of such separation is not wanting in other verte-
brates, however scant it may be in the fowl.

SCIENTIFIC LITERATURE

GENETIC DEFINITIONS IN THE NEW STANDARD
DICTIONARY

THE widely advertised aim of the Funk & Wagnalls Company
to include in their ‘‘New Standard Dictionary of the English
Language ’’ all of the new additions to scientific terminology natu-
rally invites the specialist in each branch of science to examine
the definitions of the new words in his own field. Professor
Miller’ has called attention to the fact that the mathematical
definitions are not reliable. The same criticism must be made
regarding the definitions of many terms now familiar in the liter-
ature of genetics. For some of the errors in these definitions the
editorial staff can not be blamed, because the errors were passing
current among genetic writers themselves, at a time when further
changes in the dictionary probably became impossible; other
errors are less easily explained. While such a monumental work
as the Standard Dictionary tends to fix the usage of language,
the shortcomings of the genetic definitions may not be expected to
seriously affect the terminology actually used by the specialists
in this field; but for those who are engaged in other scientific
fields, who hive only a casual interest in genetics, and who must,
therefore, depend upon the dictionary for the meaning of any
genetic terms they may happen to meet, the erroneous definitions
are unfortunate. While very few of the genetic definitions are
free from defects, either of omission or of commission, only those
which seem most obviously defective will be considered here. In
the following list of words the definition of the New Standard
Dictionary is stated first, and then follows, in italic type, a defi-
nition which I believe will meet with the approval of most
geneticists.

Acquired. Transmitted by inheritance to subsequent generations ; as,
acquired characters.

Acquired character. A modification of bodily structure or habit which is
impressed on the organism in the course of individual life.

Both of these definitions occur in the New Standard Dictionary,
the first under ‘‘acquired,’’ the second under ‘‘character.’’ Al-
though ‘‘impressed on’’ may not be the best figure of speech to use
in this connection, the second definition represents fairly well the
correct usage of this phrase. It is difficult to understand east

1 Science, N. S., 38: 772, November 28, 1913.

52

No. 577] NOTES AND LITERATURE 53

essentially the same definition should not have been given at
both places.

Allelomorph. ‘‘In Mendelian inheritance a pair of contrasted ariaa
which become segregated in the formation of reproductive cells.’

elomorph. One of a pair o contrasted characters which are alternative
to each other in Mendelian inheritan Often used with doubtful pro-
priety as a synonym for gene, factor or r aiia

The defects in the dictionary definition in this case are two:
(a) The definition is plural, while ‘‘ allelomorph ’’ is singular;
the ‘‘allelomorph’’ is not a pair of characters, but a single char-
acter. (b) No segregation of allelomorphs takes place in the
formation of asexual reproductive cells.

elomorphism. ‘‘The presence of ae pairs of characters. ’’

Allelomorphism. A relation between two characters, suc t the de-
terminers et both do not enter the same juin. but are separated into
mer gamet

terna See inheritance. ‘‘The transmission to alternating generations
of descendants of the characteristics of either parent, as that of the father
to the odd, and of the mother to the ois generations. ’’?

Alternative inheritance. A distribution of contrasting parental or an-
cestral characters among offspring or descendants, such that the individuals
exhibit one or other of the characters in question, combinations or blends of
these characters being absent or exceptional.

Biotype. ‘‘In Mendelian inheritance a race or strain that breeds true or
almost true; a term introduced by Johannsen.’’

Biotype. A group cf individuals all of which have the same genotype.

The word ‘‘biotype’’ was introduced into English by Dr.
Johannsen? in 1906 with the definition ‘‘one single ‘sort’ of
organisms.’’ It is a term of general applicability and not limited
to Mendelian races, as stated in the New Standard Dictionary.
Although homozygous biotypes generally do breed true, this is not
an essential feature and therefore should not be included in the
definition. Ever-sporting varieties are now well known which do
not breed true, but which, so far as present evidence goes, do
constitute single homozygous biotypes. Heterozygous biotypes
genen do not breed true.

Clon. ‘‘A plant-group the members of eja have been grown from an
original stock, but which do not come true from seed.

yore _A group of individuals produced Wok a single original individual
by some process of asexual reproduction, such as division, budding, slipping,
grep, parthenogenesis (when unaccompanied by a reduction of the
hromosomes),

There are me defects in the dictionary definition of this
word, even if restricted to a plant-group in accord with the
original meaning given to it by Webber, who introduced the word.

2 Report of the third International Conference of Geneties, p. 98.

54 THE AMERICAN NATURALIST [Vou. XLIX

The defects consist, first, in the ambiguity of the word **stock,’’
because we may grow plants ‘‘from an original stock’’ of seeds,
quite as well as from cuttings, while a clone is derived from a
single individual; second, the statement that clones do not come
true from seed is incorrect, for a clone formed by cuttings, ete.,
from a homozygous individual does ‘‘breed true,’’ å. e., it pro-
duces seedling offspring of its own type. The word is now
being generally applied to animals as well as to plants.

Coupling. (‘‘ Genetic coupling’’ is not defined in the dictionary.) Such

over. (Not given a genetic definition in the dictionary.) A sepa-
ration into different gametes, of determiners that are usually coupled, and
the association of determiners in the same gamete, which are generally alleto-
morphic.

Cryptomere. ‘‘A plant a which may exist in the germ-cells with-
out making its presence visible

mere. A factor or gene whose presence can not be inferred from
an inspection of the individual, but whose existence can be demonstrated by
means of suitable crosses.

The chief defect in the dictionary definition is the restriction
of this term to plant characters. ‘‘Cryptomere’’ is a general
genetic term which may be applied as well to animals as to plants.

Determiner. ‘‘The same as determinant 3.’

Determiner. An element or condition in a germ-cell which is essential to
the development of a particular feature, quality or manner of reaction of the
organism which arises from that germ-cell; a gene or factor.

The word ‘‘determiner,’’ as used in recent years, is not the
equivalent of ‘‘determinant 3,’’ which latter is correctly defined
in the dictionary in terms of Weismann’s complicated hypothesis.
“*Determiner,’’ ‘‘factor’’ and ‘‘gene’’ are now quite generally
used interchangeably without implication as to their fundamental
nature, ites i in the generic sense, as ‘‘that which determines.”’

Dominance. ‘‘In the cross-bred offspring of parents with marked mu-
tually antagonistic characteristics, the exhibition by such peste or its
descendants of one of these characteristics to the exclusion of the eh

Dominance. In Mendelian hybrids the capacity of a aada which is
derived from only one of the two generating gametes to develop to`an extent
nearly or quite equal to that exhibited by an individual which. has derived
the same character from both of the generating gametes. In the absence
of dominance the given character of the hybrid usually presents a ‘‘blend’’
or intermediate conldition between the two parents, but may present new
features not found in either parent.

There are several defects in the dictionary definition. In the
first place, the parents used in a given cross may not themselves.

No. 577] NOTES. AND LITERATURE 55

be homozygous, in which case some of their offspring will resemble
one parent and some the other; in such a case, according to the
dictionary, both of the contrasted characters would exhibit domi-
nance. The phrase ‘‘or its descendants’? would make it pos-
sible, in any case, to include both recessives and dominants, since
among the descendants of such cross-bred individuals there will
also be recessive individuals which ‘‘exhibit one of the character-
istics to the exclusion of the other.’’

Dominant. ‘‘(1) A marked parental character exhibited by a cross-bred
organism and its descendants. (2) The parent, cross-bred organism, or
descendant exhibiting such character. Parental characters latent in a cross-
bred organism, but actively evidenced by its dengensanie; are called reces-
sives, as are the descendants which exhibit them

Domin: (1) A character which exhibits Joi inance, i. e., that one of
two contrasted parental characters which appears in the individuals af the
first hybrid PEHE to the exclusion of the alternative, ‘‘recessive,’’ char-

ter. (2) individual possessing a dominant character, in contrast to
those ads which lack that character, which are called ‘‘ recessives,’’

An ‘‘extracted dominant,’’ as defined in the dictionary, is not
distinguishable from the pure homozygous dominant used in the
cross from which the dominant in question was ‘‘extracted,’’ as
no mention is made of the essential historical fact that it is of
hybrid origin and that its parent or other known ancestor did not
breed true to the same dominant character.

Factors. ‘‘ Latent rks unite which upon crossing give rise to
the new characters found in the h

Factor. An independently gubarstabts element of the genotype whose
pre. e makes possible any specific reaction or the development of any par-
ia unit-character of the organism which possesses that genotype; a gene
or determiner.

The limitation of the term ‘‘factor’’ to those cases in which new
characters appear in hybrids, is not in accord with present usage.
All the various characters of organisms are to an important degree
dependent upon the existence of genotypic factors, regardless of
the behavior of these organisms in crosses.

“¿<A minute hypothetical particle supposed to be the bearer of he-
reditary qualities.’’
n element of the genotype; a genetic factor; a determiner.

The treatment of this word in the dictionary is particularly
mischievous. When I introduced the word ‘‘gene’’ to English-
reading students, I said:* ‘‘This word is proposed by Dr.
Johannsen .. . to denote an internal something or condition
upon whose presence an elementary morphological or physio-
logical characteristic depends. The word ‘gene’ has the advan-

8 Am. Nat., 43, p. 414, 1909.

56 THE AMERICAN NATURALIST [ Vou. XLIX

tage that it does not assume by its form or derivation any hypoth-
esis as to the ultimate character, origin or behavior of the deter-
mining factor.’’ In adopting the word ‘‘Gen’’ in the German,
Johannsen said :* ‘‘Das Wort Gen ist völlig frei von jeder Hy-
pothese ; es driickt nur die sichergestellte Tatsache aus, dass jeden-
falls viele Eigenschaften des Organismus durch in den Gameten
vorkommende besondere, trennbare und somit selbständige
‘Zustände,’ ‘Grundlagen,’ ‘Anlagen’—kurz, was wir eben Gene
nennen datn- beun sind. .. . die Gene sehr vieler Eigen-
schaften glatt trennbar sind, während andere nicht oder nicht
glatt sich trennen. Dies alles erinnert an das Verhalten chemi-
scher Körper. Damit ist aber noch gar nicht gesagt, dass die Gene
selbst chemische Gebilde oder Zustände seien—darüber wissen
wir vorläufig noch gar nichts.’’ How different is all this from ‘‘a
minute hypothetical particle’’! It is obviously improper, there-
fore, to define a gene as a ‘‘minute particle.’’ Neither is it correct
to say that it is ‘‘supposed to be the bearer of hereditary qual-
ities.” It is only the something of unascertained nature, which
must lie at the foundation of any elementary hereditary quality.
The spelling ‘‘gene’’ is not even mentioned in the dictionary as a
variant, yet this was the original spelling and is now in practic-
ally universal use among geneticists, while no one uses ‘‘gen.’’

‘t A race of organisms different from another in its hereditary
qualities; contrasted with phenotype.’’

Genotype. The fundamental hereditary constitution or sum of all the
genes of an organism,

The unfortunate definition of ‘‘genotype’’ given in the dic-
tionary was current in America at the time when the dictionary
forms were probably closed, so that the editors are not in any way
to blame for the totally erroneous definition. The definition given
by the dictionary for ‘‘genotype’’ fits fairly well the word
**biotype.’’

Heredity. ‘‘The oo manifested by an organism to develop in the
likeness of a progenitor.

Here The distribution of genotypic elements of ancestors among
the descendants; the resemblance of an organism to its parents and other
ancestors with respect to genotypic constitution.

The results of modern experimental work on heredity show that
the definition given by the dictionary is entirely too restricted.
Heredity must be so defined that it may apply to characters mas
were never exhibited by any ancestor.

Heterozygosity. ‘‘In Mendelian inheritance, the state or condition due
to an organism having developed from a heterozygote.’’

4‘*Elemente der exakten Erblichkeitslehre,’’ 1. Aufl., 1909, pp. 124-125.

No. 577] NOTES AND LITERATURE 57

Heterozygosity. The condition of an organism due to the fact that it is
a heterozygote; the state of being heterozygous; the eet, to which an
individual is heterozygous.

Heterozygote. ‘‘A Mendelian hybrid resulting from the fusion of two
gametes that bear different allelomorphs of the same character =s which
in consequence does not breed true; contrasted with homozygote.

Hete A zygotic individual in which any given genetic factor
has been derwed from only one of the two generating gametes. Both eggs
and sperms produced by such an individual are typically of two kinds, half
of them containing the gene in question, the rest lacking this gene; conse-
quently the effspring of heterozygotes usually consist of a mixture of indi-
ee some of which possess the corresponding character while others
lack

etic ‘t Development em a zygote originating from a union
of two gametes of the same kind,’
zygosis, sis state of oh homozygous; the extent to which an

E

Homozygote. iy sels formed by the conjugation of two gametes of
the same stock; any animal or plant that receives and retains the dominant
or recessive characters of both its parents, and is therefore said to be true
to type, and breeds true to t

Homozygote. An individual in which any given genetic factor is doubly
present, due usually to the fact that the two gametes which gave rise to this
individual were alike with respect to the determiner, in question. Such an
individual, having been formed by the union of like gametes, in turn
generally produces gametes of only one kind with respect to the given
character, thus giving rise to offspring which are, “ this Com like
the parents; in other words, NaDa usually ‘‘breed true.’’ A
**nositive’’ homozygote with respect to any character contains a pair
determiners for that character, while a ‘‘negative’’ homozygote lacks this
pair of determiners.

‘‘Two gametes of the same stock’’ is ambiguous because of the
indefiniteness of the word ‘‘stock.’’ Many homozygotes receive
some dominant and some recessive characteristics of the two par-
ents; and what can be intended by the statement that a plant or
animal which receives certain characteristics also ‘‘retains’’
them? How could it do otherwise?

Hypostasis. (Not given a esis definition in the dictionary.) That
relation of a gene in which i ion fails to appear because of the
gee or Dans effect of ‘witli gene; contrasted with ‘‘ epistasis.’’

The corresponding adjective ‘‘hypostatic’’ is smo not given a
genetic “definition in the dictionary.

Mendelize. ‘‘To cause to follow Mendel’s law of inheritance.’’

Mendelize. To follow Mendel’s law of inheritance.

The word is rightly indicated in the dictionary, as an intran-
sitive verb; it is manifestly incorrect to define it by the use of a
transitive ark

58 THE AMERICAN NATURALIST [Vou, XLIX

Mutant. ‘‘That which admits of or undergoes mutation or change;
specifically, an individual or a species which shows significant changes in
form or character in a single generation.’’

Mutant. An individual possessing a genotypic character differing bg
that of its parent or those of its parents, and not derived from them
normal process of segregation.

The expression ‘‘significant changes’’ is ambiguous, since every
change is significant of something.

Mutate. ‘‘To ‘sport.’ ’’

Mutate. To undergo a change in genotypic character independently of
normal segregation,

The word ‘‘sport’’ which is used in the dictionary definition of
‘‘mutate’’ is defined thus: ‘‘To vary suddenly or spontaneously
from the normal type; said of an animal or plant or of one of its
parts.’’ It is well known that many such sudden and spontaneous
variations from the normal type are not due to mutations. The
word ‘‘mutation’’ is defined in the dictionary as ‘‘a permanent
transmissible variation in organisms, as distinct from fluctuation.”’
This definition is good as far as it goes, but should expressly ex-
clude transmissible variations which are due to normal segrega-
tion and recombination of determiners.

Phenotype. ‘‘A type or strain of organisms distinguishable from others
by some character or characters, whether their observable differences from
other organisms be due to their inherent hereditary differences or to the
direct action of the environment upon them: contrasted with genotype.’’

Pheno e ee type of an individual or group of individ-
uals, i. e, the sum of the externally obvious characteristics which an indi-
vidual possesses, or which a group of individuals possesses in common; con-
trasted with genotype.

‘‘Phenotype’’ and ‘‘genotype’’ are both abstractions; the qual-
ities which distinguish the phenotype are always capable of direct
observation, while those of the genotype can only be inferred from
the results of genetic experiments.

e and absence hypothesis, ‘‘in the Mendelian doctrine of in-
heritance, the theory that an allelomorphic pair of characters in every zygote
has two contrasted factors or determinants, one representing the por
character of the generated organism and the other denoting its absence.’

Presence and absence hypothesis. The hypothesis that any simple Men-
delian difference between two individuals, results solely from the presence
of a factor in the genotype of the one individual, which is absent from that
of the other. Presence and absence of unit-differences as a convenient
method of describing the results of genetic experiments should be carefully
distinguished from the presence and absence hypothesis. The method is-
purely objective and entirely free from hypothetical implications.

It will be noted that the dictionary definition of this phrase is
directly opposite in significance to the one here set forth.

No. 577] NOTES AND LITERATURE ` 59

Pure line (Not included in the dictionary.) A group of individuals
derived solely by one or more self-fertilizations from a common homozygous
ancestor. Sometimes erroneously applied to groups of individuals believed
to be genotypically homogeneous (a homozygous biotype or a clone) without
regard to their method of reproduction.

pulsion. (Not given a genetic definition in the dictionary.) Such a
relation between two genetic factors that both are not, as a rule, included
in the same gamete, referring especially to cases in which the factors in
question give rise to enon different characteristics; also called
‘‘spurious allelomorphism. '

Sex-limited inheritance. (Not defined in the dictionary.) The associa-
tion of the determiner for any unit-character, with a sex-determiner, in such
a manner that the two determiners are either generally included in the same
gamete, or that they are generally included in different gametes, This
method of inheritance is also called ‘‘seu-linked’’ inheritance by Prater
£ rgan and his students.

Segregate. ‘‘To become separated from the rest; specif., of Mendelian
cia’ = separate, by a numerical law, into dominants, hybrids and re-
cessives

Segre: wa ate. With reference to Mendelian unit-characters, to become sep-
arated through the independent distribution of the genetic factors before or
at the time of the formation of the gametes

The dictionary definition goes too far; the formation of domi-
nants, hybrids and recessives depends not alone upon the fact that
the factors segregate, but that the segregated factors recombine,
The word ‘‘segregation’’ receives a fairly satisfactory definition.

Unit-character. (Not included in the dictionary.) In Mendelian in-
heritance a character or alternative difference of any kind, which is either
present or absent, as a whole, in each individual, and which is capable of
becoming associated in new combinations with other unit-characters.

I have made no systematic study of the definitions of technical
terms in other related fields, but have noted incidentally that
there is no recognition in the New Standard Dictionary of the
generally familiar usage of the words ‘‘meristic’’ and ‘‘sub-
stantive’’ as applied to types of variation.

G. H. SHULL

PROFESSOR DE VRIES ON THE PROBABLE ORIGIN OF
ŒNOTHERA LAMARCKIANA

IN a recent paper Professor Hugo De Vries' has given us the
results of a second examination in 1913 of material from the
herbarium of Lamarck, and of other sheets of @nothera in the
collections of the Muséum d’Histoire Naturelle in Paris.

1De Vries, Hugo, ‘‘ The Probable Origin of Gnothera Lamarckiana Ser.,
Bot. Gaz., Vol. LVII, p. 345, 1914.

60 THE AMERICAN NATURALIST [Vor. XLIX

is a very important contribution for there have been some changes
in the mounting of important specimens in the herbarium of
Lamarck since the first studies by De Vries in 1895 and it was
not clear what material formed the subject of his discussion in
* Die Mutationstheorie.’’ As the result of this second examina-
tion (1913) there can be no misunderstanding of De Vries’s
conclusion as to what represents the type of Œnothera Lamarck-
iana Seringe, and we have also very positive opinions on the
identity of other interesting material in the collections at Paris.
Thanks to his descriptions and photographs of these sheets further
confusion will be impossible and botanists may now make for
themselves the observations that will in the end determine their
judgment of the soundness of Professor De Vries’s views and of
the value of the exceptions that may be taken to them.

I shall not at this time discuss in detail the queries which pre-
sented themselves on my reading of De Vries’s paper. The most
important of the points probably rest on facts that should be
shown by the material, but which have not been published in the
account of De Vries. I expected to have the data in question this
autumn but the European disturbances have necessarily upset
my plans and it may be very many months before I can take up
the matter. `

However, I will briefly say that De Vries’s identification of
the sheets under consideration are to me not convincing chiefly
for the following reasons. His account gives no description of
the pubescence of the sepals, stems, or capsules when present.
Yet pubescence is a character of great importance in the descrip-
tion of many species of @nothera. To illustrate the point, all
races of O. grandiflora Solander that I know have sepals and
capsules almost glabrous or very sparsely pilose and puberulent.
Lamarckiana on the contrary presents sepals and capsules with
a very heavy puberulent and pilose pubescence. Should any of
the specimens at Paris which De Vries has identified with the
Lamarckiana of his cultures present sepals or capsules lacking
the heavy pubescence of this plant the fact to me would be very
strong evidence that his identification was incorrect.

There are two sheets under consideration as standing for the
type of G@nothera Lamarckiana Seringe. De Vries regards one

as unequivocally representing the type specimen. I have for
various reasons placed the greater emphasis upon the other.
Both specimens as shown in photographs appear to have essen-
tially the same features as to their general morphology. Miss

No. 577] NOTES AND LITERATURE 61

Eastwood and M. Gagnepain who compared the two specimens
reported to me that they were very similar. Both were undoubt-
edly known to Lamarck since the two sheets bear his hand-
writing, and it is quite possible that Lamarck based his descrip-
tion on both specimens,

The general morphology of these specimens presents several
features that are not those of the Lamarckiana of De Vries’s
cultures. Chief among these are (1) the approximate branches,
(2) the foliage of narrower and more distinctly petioled leaves,
(3) the inflorescence more open and with narrower bracts, (4)
the buds more slender and tapering, and apparently with more
attenuated sepal tips, (5) the long delicate hypanthium. In
these features the specimens are closer to O. grandiflora than to
Lamarckiana. Such morphological characters, it is true, might
vary somewhat under different conditions of growth and with
the time of collection whether early or late in the season. The
pubescence should give us the stronger evidence of relationship
since pubescence would be little if at all affected by growth con-
ditions or by season. Of the pubescence on one of these speci-
mens I have Gagnepain’s statement that it is close to that of
grandiflora, but it is only fair to say that no @nothera specialist
has reported upon such a comparison as is desired.

Lamarck’s description of the capsules of his plant as short and
glabrous is a point of great importance. The capsules of De
Vries’s Lamarckiana are certainly not glabrous but they are
short. In my contention that Lamarck’s plant was a form of
O. grandiflora Solander I was at first forced to assume that
Lamarck must have described immature or partially pollinated
capsules. I have, however, this summer grown cenotheras from
Mississippi which have the rosettes, habit, foliage, inflorescence,
and flowers of grandiflora, but which developed glabrous short
capsules essentially of the same relative proportions as those of
Lamarckiana. It is immaterial what is the origin or genetic
history of these plants; systematically speaking they represent
short-capsuled forms of O. grandiflora. Thus we now know of
grandiflora-like types which even as to their capsules agree with
the description of Lamarck. De Vries does not seem to be dis-
turbed by the fact that the material of his cultures presents
capsules with a heavy puberulent and pilose pubescence while
Lamarck’s description specifies a capsule ‘‘glabre.’’

In summary I must say that my opinion remains auhani
with respect to the affinities of the plant deseribed by Lamarck,

62 THE AMERICAN NATURALIST | [Vou. XLIX

@nothera Lamarckiana Seringe. Whichever of the two speci-
mens considered above represents the type, or if both were con-
cerned in the description, the evidence is to me very strong that
Lamarck dealt with forms of O. grandiflora Solander. I can see
no proof or even reasonable evidence that the Lamarckiana of
De Vries’s cultures agrees with either of the specimens from
Lamarck’s herbarium. A final judgment, however, should not
be made until we have before us details respecting the pubescence
of the specimens known to Lamarck.

De Vries is very positive that two other sheets in the collec-
tions at Paris present specimens agreeing with his Lamarckiana.
The first of these (De Vries, Plate XVIII) is from the her-
barium of Abbé Pourret and shows material which seems to me
to offer very much the same difficulties to an identification with
De Vries’s Lamarckiana as do the specimens of Lamarck. The
foliage of lanceolate leaves clearly petioled, the slender tapering
buds, the long delicate hypanthium; these are not characters
representative of the plants from the cultures of De Vries. They
are characters of O. grandiflora Solander and should the pubes-
cence prove to be similar to this species I should not hesitate to
place these specimens of Abbé Pourret among the forms of
grandiflora. Until we know the facts of the pubescence, further
discussion is unwise, but it does not seem to me that De Vries’s
identification rests on good evidence.

The remaining sheet at Paris which De Vries (Plate XIX)
identifies with his Lamarckiana is a plant from the herbarium
of André Michaux. De Vries on historie grounds naturally
attaches importance to this sheet for if it could be established
as in agreement with his plants the fact would bear directly on
the problem of the origin of O. Lamarckiana. The flowers are
large and the buds rather stout as we find them in our cultivated
Lamarckiana, but the sepal tips are longer and the bracts much
narrower than in Lamarckiana. The most striking characters
of this specimen as shown in the photograph are the narrow
lanceolate leaves and the extraordinary length of their petioles.
That such a plant could be related to De Vries’s Lamarckiana
which has ovate-lanceolate leaves, sessile or almost sessile, seems
to me well nigh impossible. Of the pubescence De Vries tells
us nothing, yet the numerous buds on the specimen should make
it easy to determine this character and it may become a crucial
point in judging the possible or impossible relationships of the
plant.

No. 577] NOTES AND LITERATURE 63

In the discussion which must develop from the conclusions of
De Vries he has taken by far the more difficult position since he
attempts an identification of herbarium material with a type
very accurately known to us through widely cultivated living
forms. My argument is presented primarily against his identi-
fications. It is not in any degree necessary to my argument that
I should assign the sheets under consideration to definite species.
Whether this can be done for any of them time will tell and I
must repeat that as evidence the character of the pubescence
may prove of the greatest value. I am working on the hypothesis
that the specimens of Lamarck and that of Abbé Pourret are
forms of O. grandiflora Solander. As for the specimen of André
Michaux, so many remarkable forms of @nothera are coming
into the experimental garden from the southern and western
United States that I am quite unwilling to express at present
even a guess as to its affinities.

De Vries has welcomed my suggestion that the source of the
cultures of Carter and Company may have been not Texas, as
they state, but England. This possibility seems to me to offer
an important line of investigation of early British records and
collections, but at present the suggestion appears to me nothing
more than a working hypothesis, although well worthy of atten-
tion. Texas and the West have some wonderful large-flowered
cenotheras and Carter and Company may have obtained from
such sources a plant which later hybridizing with other forms
produced the Lamarckiana of our present cultures. That there
are American western species which will hybridize with Euro-
pean biennis and produce a synthetic Lamarckiana is I believe
established by my present studies with Œnothera franciscana
Bartlett.

In recent papers I have reported that first generation hybrids
of O. franciscana pollinated by the Dutch biennis have the
essential taxonomic characters of the small-flowered forms of
O. Lamarckiana. They differ from Lamarckiana in relatively
small plus or minus expressions of these characters. It was to be
expected that large F, generations would give a wide range of
variation or segregation of characters and that forms would
appear much closer to Lamarckiana than the parent F, plants.
This proved to be the case in F, cultures of last summer (1914)
totaling about 1,600 plants. Among these I obtained a number |
of individuals which were so close to the large-flowered Lamarck-
iana that flowering shoots could scarcely be distinguished as to

64 THE AMERICAN NATURALIST [Vor. XLIX

pubescence, foliage, inflorescence, buds, flowers, and capsules.
The rosettes were also Lamarckiana-like. Only in habit was it
somewhat difficult to match the symmetry of Lamarckiana.
That further selection in later generations is likely still to fur-
ther improve on the results of this synthesis seems altogether
probable. These studies will shortly be described in full.

I am well aware that a synthesis of a Lamarckiana-like hybrid
even should it throw in successive generations a series of marked
variants (mutants) will not be considered by De Vries and his
disciples as casting doubt on the validity of the ‘‘mutation’’ of
Lamarckiana. They will say that in this case the hybrid took its
mutating habit from one or both of the parents. Since in my
cross one of the parents is the Dutch biennis which Stomps has
shown can produce nanella, semi-gigas and sulfurea mutants, it
will be claimed that any behavior of my hybrids similar to
mutation will be due not to the mixing of diverse germ plasms,
i. e. to crossing, but will be merely a further expression of
mutating habits inherent in the germ plasm of at least biennis
if not also of franciscana.

This phase of the discussion may rest until we know the future
behavior of my hybrids and the possibilities of the Dutch biennis
-as a form capable of mutation. It is to be expected that Stomps
will carry out his very important studies on a scale that will
virtually exhaust the mutative possibilities of this species. Such
a study on a close-pollinated species of Œnothera so well known
-as the Dutch biennis will give, it seems to me, the safest data
that has yet been published by students of mutation among the
Gnotheras. It becomes a matter of great interest to know the
range of variants that such a type can produce. Similar studies
= among some of the wild American species should also be made.
The open-pollinated assemblage of forms to which Lamarckiana
belongs must always be open to suspicion of hybridization more
or less remote in time or distant in relationship. Only prolonged
experiment can establish an open-pollinated @nothera as free
from the taint of crossing.

It is, I trust, clear that one may believe very strongly that
nothera Lamarckiana is not safe material on which to base
experiments designed to test the mutation theory and yet remain
rebaptive to evidence that may come from other sources.

BRADLEY Moore Davis

UNIVERSITY OF PENNSYLVANIA,

_ October, 1914 .

VOL. XLIX, NO. 57

The American Naturalist

MSS. intended for m ier and books, etc., intended for review should be
sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Hud son, New York.

cles containing summaries of research work bearing on the
[n pubteation of organic evolution are pacar tg welcome, and will be given preference
n pu

bscription price is four dollars a ae Foreign postage is
2 Canadian twenty-five cents additional. The charge for ine: ae ‘is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS :
Lancaster, Pa. Garrison, N. Y.
NEW YORK: Sub-Station 84

A d-class matt April 2, 1908, + tha Paat A(R

Congress of Mareh 3, 1879.

t Lancaster, Pa., under the Act of

FOR SALE JAPAN NATURAL HISTORY SPECIMENS

ARCTIC, ICELAND and GREENLAND Perfect Condition and Lowest
BIRDS’ SKINS, tein Bird Skins, Oology, deh Marine
Well pees | Low Prices and others. Catalogue free. Correspond-

ae ; ence aera
iculars of ; .

Bi DINESEN , Bird Collector T. FUKAI, Nat
vi k North celana, Via — England

+f INVESTIGATION Facilities for research in Zoology, —
| Embryology, Physiology and, Bot

j avail-

THE
AMERICAN NATURALIST

VoL. XLIX February, 1915 No. 578

CYCLES AND RHYTHMS AND THE PROBLEM
OF “IMMORTALITY”? IN PARAMECIUM

PROFESSOR GARY N. CALKINS
CoLUMBIA UNIVERSITY

Tue recent brilliant work of Woodruff and Erdmann
has thrown a flash of light upon the old question of age
and death in protozoa and upon the problem of the sig-
nificance of conjugation. The long, successful cultivation
of Paramecium aurelia by Woodruff led Jennings to say:

The work of Woodruff demonstrates that the very limited periods
within which Maupas and Calkins observed degeneration has no signifi-
cance for the question as to whether degeneration is an inevitable
result of continued reproduction without conjugation. In other words,
it annihilates all the positive evidence for such degeneration, drawn
from work on the infusoria. It justifies the statement that the evidence
is in favor of the power of these organisms to live indefinitely, if they
are kept under’ healthful conditions. It shows that Weismann was
correct in what he meant by speaking of the potential immortality of
these organisms.

The same work of Woodruff led Minot to say in the
course of his German lectures:

Quite conclusive as to the absence of senescence are the experiments
of L. L. Woodruff, who has maintained a — race of Para-
mecium for five years without conjugation.?

Woodruff also makes the statement in different publica-
tions that the cells of his pedigreed race of Paramecium
1‘*Age, Death and Conjugation in the Light of Work on Lower Organ-
isms,’’ Pop, Sci. Mo., June, 1912, p. 568.
2 ‘í Modern Preble of Biology,’ 2018, p. 62.
65

66 THE AMERICAN NATURALIST [Vou. XLIX

aurelia possess the potentiality to perpetuate themselves
indefinitely by division under proper environmental con-
ditions. In short, his results have given almost the only
experimental evidence in support of the view, advocated
by Weismann, that protozoa are potentially immortal.

The importance of this generalization and the deduc-
tions from it are self evident, and it is unfortunate that
so many should have advanced it before the life history
of Paramecium aurelia was fully known. Woodruff is to
be congratulated, however, in that he, with Miss Erdmann,
has now worked out stages in the life history of this or-
ganism which go far in clearing up the discrepancy be-
tween his results and those obtained by Maupas and his
followers.

Woodruff has carried on his pedigreed race of Para-
mecium aurelia for more than seven years with a fairly
uniform division rate, subject, however, to occasional and
periodic fluctuations which he calls rhythms. These cor-
respond roughly to what I have termed cycles which end
in depression periods and, unless stimulated, by death;
in rhythms, however, Woodruff maintains, there is no evi-
dence of depression.* Recently Woodruff finds from a care-
ful study of material fixed during the low periods of his
division rate rhythms that there takes place a complete
nuclear reorganization, after which the organisms continue
to live with renewed vitality as shown by the ascending
division rate. This process consists in the disintegration
and probable absorption in the cytoplasm of the old macro-
nucleus, one or more divisions each of the old micronuclei,
degeneration of some of the products of these divisions,
and the ultimate reformation of functional macronuclei
and micronuclei from others. These are, in essence, the
important new facts which cytological study has revealed
in the life history of Paramecium aurelia, and some evi-
-83I would like to suggest to Professor Woodruff that he work out the
death rate, the data for which he can undoubtedly obtain from his records.

I venture to predict he will find that the death rate rises with the decline of
the division rate and during the low sweep of the rhythms.

No. 578] CYCLES AND RHYTHMS 67

dence is further adduced to i that similar processes
occur in Paramecium caudatum.

While there is little to criticize in regard to the facts as
described for this. remarkable process, there is room for
difference of opinion in regard to the conclusions which
Woodruff and Erdmann draw from them. In addition
to what Hertwig has already written about this work, I
would call attention particularly to their conclusions con-
cerning endomixis, parthenogenesis, conjugating and non-
conjugating lines, Gi life cycle, and the potential immor-
tality of Paramecium.

In regard to endomixis the authors state:

Since the process results in the dissemination of the material from the
old macronucleus and the so-called reduction micronuclei in the cell, it

tion of the cell. This involves a more profound intermingling of
nuclear and cytoplasmic substances than is possible during the typical
vegetative life of the cell. Since this intermingling occurs within a -
` cell we term this reorganization process endomixis. Endomixis is fol-
lowed by a slight acceleration of cell phenomena and a new rhythm is
initiated.*

Further on they add:

We would therefore put emphasis on molecular re eae as the
result common both to endomixis and to conjugation

Nearly forty years ago Engelmann Se conju-
gation in much the same way as a process of reorganiza-
tion of the cell:

. die Conjugation der Infusorien leitet nicht zu einer Fortpflanzung
durch ‘Eier, ‘Embryonalkugeln’ oder irgend welche andere Keime,
sondern zu einem eigenthiimlichen Entwickelungsprocess der conjugirten
Individuen, den man als Reorganization bezeichnen kann.®
In several places in the same publication Engelmann
speaks of physical and chemical changes as accompanying

t Woodruff and Erdmann, ‘‘A Normal Periodice Reorganization Process
without Cell Fusion in Paramecium,’’ Jour. Exp, Zool., Vol. 17, No. 4,
1914, p. 491. The italics are in the original.

5 Ibid, p. 491.

6‘‘Ueber Entwickelung und Fortflanzung ‘von Infusorien,’’ Morph.
Jahrb., 1, 1876, p. 628.

68 THE AMERICAN NATURALIST [Vou. XLIX

conjugation. A similar interpretation was given by
Calkins:

. it is now a well-known fact that in this process of reorganization
the old macronucleus fragments and ultimately disappears in the cyto-
plasm. This disappearance must give rise to a great increase in the
nucleo-protein content of the cell, therefore to a new chemical composi-
tion of the cell as a whole. We have recently shown that, under cer-
tain conditions, nucleo-proteins (especially the purines) have a mark-
edly stimulating effect on the rate of cell division.’

Now such intermingling is no more characteristic of this
process of asexual reorganization than it is of the reor-
ganization following conjugation. In both cases, as
Woodruff and Erdmann show, reorganization is effected
by the physical and chemical change of the old macro-
nucleus and portions of the old micronucleus or micro-
nuclei. The sole difference in these processes of reor-
ganization is not to be found in the molecular rear-
rangement of the cell, but, as Woodruff and Erdmann
state, in the presence after conjugation of a syncaryon
and the nuclei derived from it. This difference, how-
ever, does not amount to much in closely related pairs
in conjugation. Several observers have shown that
closely related individuals, even sister cells, of Para-
mecium may conjugate, and I have followed out through
360 generations the history of such an endogamous ex-
conjugant from a pair which came from the same ances-
tral cell not more than ten days prior to conjugation.
There can not be a great difference in the syncaryon re-
sulting from such a union, over the functional micronu-
cleus had it undergone asexual endomixis. In other
words, the excellent term endomixis does not indicate
phenomena peculiar to asexual reorganization in Para-
mecium, but applies equally well to the process of reor-
ganization following conjugation. The terms asexual en-
domixis and sexual endomixis may serve to distinguish
the process of intermingling during parthenogenesis and

after conjugation, respectively.

7<*The perp conic: Conjugation of Blepharisma wnavtans Jour.
Morph. Vol. 23, 1912, p. 685.

No. 578] CYCLES AND RHYTHMS 69

Woodruff and Erdmann limit the application of the
term endomixis to the process of reorganization without
conjugation:

We therefore have employed a new term “ endomixis” for the re-
organization process in Paramecium, in preference to parthenogenesis
which Hertwig applied when he incidentally noted some isolated stages
of the nuclear phenomena which we have elucidated.®
A new name can not alter the significance of a process or
phenomenon. Parthenogenesis, in its broad sense, is the
development of an individual from an egg without fertili-
zation. In the same sense that a Paramecium ex-conju-
gant develops into a new individual, so does a Parame-
cium after this process termed endomixis. Woodruff and
Erdmann say:

In parthenogenesis there is à chromatin reduction which occurs and
is compensated for either in the egg itself or in some later period of the
life cycle of the race.®
The authors are not very happy in selecting this feature
as distinguishing parthenogenesis from asexual endo-
mixis, for in most cases of recognized parthenogenesis
in metazoa chromatin reduction plays no part; for exam-
ple, the majority of parthenogenetic’eggs give off only
one polar body, thus retaining in the egg the diploid num-
ber of chromosomes; others, notably the aphids and phyl-
loxerans, do not undergo synapsis or chromatin reduction ;
some others it is true, give off both polar bodies and develop
with the haploid number of chromosomes as is the case
in bees (males), and in artificial parthenogenesis. As to
the significance of parthenogenesis neither polar body
formation nor chromosome reduction furnishes the key,
for in many cases the eggs are predestined to partheno-
genetic development long before the polar body nuclei
are formed.

In regard to the reducing divisions of the chromosomes
in Paramecium we know very little. Evidence has been
adduced to indicate that the chromosomes are divided

8 Ibid., p. 493.

9 Ibid., p. 492.

70 THE AMERICAN NATURALIST [Vou. XLIX

longitudinally in both the first and the second divisions
of the maturation process. The significance of the third
division is as obscure in Paramecium as maturation is in
some metazoan hermaphrodites.

In parthenogenesis, finally, we are dealing with a bio-
logical phenomenon, not with an interpretation of par-
thenogenesis by Winkler or Strasburger or any other
individual, and to interpret this highly significant phe-
nomenon in Paramecium solely in the light of such defi-
nitions, as Woodruff and Erdmann do (p. 493), does not
carry conviction, nor does it conceal the real significance
of the phenomenon. Asexual endomixis in Paramecium
is parthenogenesis and nothing else, as Hertwig origi-
nally maintained in connection with these same phenom-
ena. Nor, except for the protozoa, is it a ‘‘new type of
parthenogenesis’’ for, if we accept conjugation as equiva-
lent to fertilization, its analogue is shown by the majority
of parthenogenetic eggs.

In regard to conjugating and non-conjugating races of
Paramecium, Woodruff and Erdmann state: |

Thus it is proved that both the reorganization process and conjuga-
tion are potentialities of the same race—and therefore there is no evi-
dence for the view of Calkins (713) that conjugating and non-conjugat-
ing races of Paramecium exist, or that “apparently some paramecia
are potential germ cells, others are not.”1°
This is rather a sweeping generalization to. draw from one
pedigreed line in which conjugating animals appeared
only after six years in culture. If every Paramecium is
a potential germ cell, why was it that no pairs of conju-
gating aurelia were found during these six years? Or,
in Calkins and Gregory’s observations on the first 32
cells and the pure lines arising from them, all from a
single ex-conjugant, why was it that all lines from one
quadrant gave epidemics of conjugation whenever the test
was made during a period of six months, while all other
lines from the remaining three quadrants failed to give
a single pair when tested under identical conditions?

10 Ibid., p. 490. Italies in the original.

No. 578] CYCLES AND RHYTHMS 71

Woodruff and Erdmann maintain that ‘‘under just
the proper conditions’’ conjugation will occur; this, of
course, can not be denied, but the fact that under the
same conditions some lines will conjugate while otherg
will not shows a physiological difference between them
which can not be gainsaid. I have no paternal jealousy
whatsoever in regard to the terms ‘‘conjugating lines’’
and ‘‘non-conjugating lines,” and am entirely willing to
accept in their place any terms which indicate the physio-
logical difference that I wished to express. I know of
no terms that express the conditions adequately. Substi-
tute for them, if more suitable, such expressions as
‘‘always ready to conjugate’’ and ‘‘rarely ready to con-
jugate.’’ Our observations on the 32 lines certainly jus-
_ tify the statement that some lines in regard to conjuga-
tion, were always ready, while others were rarely ready.
Woodruff and Erdmann have paid no attention to the
physiological conditions which the (perhaps unfortunate)
expressions ‘‘conjugating lines’’ and ‘‘non-conjugating
lines’’ were meant to express. It is true that after ten
months all but four of the so-called non-conjugating lines
each furnished a few pairs of conjugating individuals,
just as Woodruff’s line did after six years, facts which
show that the terms ‘‘conjugating lines’ and ‘‘non-
conjugating lines’’ as applied to races of Paramecium, if
used at all, should be used only in respect to relative in-
tensity of conjugating power. In this sense Woodruff’s
race is a non-conjugating race. We have found, further-
more, that conjugating lines have a lower vitality as
measured by the division rate, and a much higher death
rate, than do non-conjugating lines, all but four of the
eight lines from the conjugating quadrant dying out
within three months as against four of the twenty-four
lines of the non-conjugating quadrants, while at the end
of twenty months only one conjugating line was alive and
sixteen non-conjugating lines, a mortality of 87.5 per cent.
for the former and 33.3 per cent. for the latter. In Para-
mecium it is conceivable that lines with a high conjugating

72 THE AMERICAN NATURALIST [Vou. XLIX

power have a less well developed power of asexual endo-
mixis than do lines that are relatively sterile, and this,
correlated with their reduced vitality, if conjugation were
prevented, would account for the death of all pedigreed
races prior to Woodruff’s, which, as Woodruff and Erd-
: mann now show, has a high power of asexual endomixis.
We are still justified, I believe, in maintaining the state-
ment—modified now by their description of asexual reor-
ganization—as quoted by Woodruff and Erdmann:

Woodruff’s Paramecium aurelia is evidently a Paramecium Methuse-
lah belonging to a non-conjugating line the life history of which is not
known in any ease.11

‘It is clear that the cycle emphasized by Maupas, Calkins and others
is merely a phantom which has continually receded as each successive
investigator has approached the problem with improved culture methods
until it has vanished with Woodruff’s race of (so far) 4,500 generations.
What remains then is the rhythm and in the light of the present study,
which demonstrates the underlying cytological phenomena of which it
is an outward physiological expression, the whole problem takes on a
new aspect. The cell automatically reorganizes itself periodically by a
process which, in its main features, simulates conjugation—but without
a contribution of nuclear material from another cell. Therefore it is
evident (as has been shown by this culture) that the formation of a
synearyon, whose components are derived from two cells, is not nec-
essary for the continued life of the cell—it has an internal regulating
phenomenon which is oF adequate to keep it indefinitely in a per-
fectly normal condition.??

Here we are brought up sharply to face the question
which every student of pedigreed infusoria since Maupas
has tried to solve. Woodruff and Erdmann conclude
from their observations that old age and natural death
do not occur in Paramecium and that the so-called ‘‘cycle’’
is non-existent. I would draw from their observations
exactly the opposite conclusions, viz., that the one appar-
ent exception among pedigreed races, to the rule of de-
pression and natural death in the absence of conjugation
or its equivalent, is now removed, and that Woodruff’s
culture is no more than a long series of cycles.

11 Ibid., p. 429.

12 Woodruff and Erdmann, p. 489.

No. 578] CYCLES AND RHYTHMS 73

We understand by a ‘‘cycle,’’ in the sense with which
the term was first employed by Calkins, a more or less
periodic alternation of high and low vitality as measured

y the division rate. The lowering division rate indi-
cates the approach of a period of depression which was
interpreted as the equivalent of old age in metazoa, since
it indicates a weakening in the chain of vital activities and
ends in death unless conjugation or its equivalent is
permitted. No one since Maupas, so far as I am aware,
has attempted to limit a cycle in terms of definite numbers
of generations or definite lengths of time. In 1904 I
stated:

The well-marked cycles, therefore, with periods of depression which
demanded stimulation of a decided character, were apparently of six
months duration, while intermediate cycles of less importance were about
three months long. . . . During the first three cycles the number of
generations was nearly the same (200, 198, and 193, respectively), the
last, on the other hand, was much less, the individuals dividing only
126 times.18
The period of six months, more or less, or 200 + genera-
tions were not regarded as measures of the cycle, and it
was understood at that time that conjugation or its equiva-
lent always inaugurates a new cycle. Woodruff in 1905
introduced the term ‘‘rhythm’’ to designate the lesser
periodic fluctuations which I had called ‘‘intermediate
eycles.’’ Since the entire substance of the much-discussed
problem of immortality in infusoria is bound up with this
question of the cycle, it is necessary to analyze the so-
called rhythms of Woodruff to see how they agree with or
differ from the so-called cycles. In Paramecium the
cycle consists of the history of a bit of protoplasm in an
ex-conjugant and its progeny from which conjugation or
its equivalent is excluded, until natural death of the en-
tire race ensues. If conjugation or its equivalent occurs
the old cycle is abandoned and a new one is started, and
there must be as many new cycles as there are times when
conjugation or its equivalent takes place. It is imma-

18 ‘Studies on the Life History of the Protozoa,’’? IV. Jour. Exp. Zool.,
Vol. I, 1904, p. 424,

74 THE AMERICAN NATURALIST [Vor. XLIX

terial, furthermore, whether such conjugation occurs be-
tween individuals of the same race, or between individuals
of diverse ancestry, the effect is the same in putting off
ultimate weakness and death. With repeated conjuga-
tions in such a race the ultimate death may be postponed
indefinitely, and this was the argument on which Weis-
mann’s revised theory of potential immortality was based.

Now it is exactly the same with Woodruff’s rhythms.
He finds in his long culture repeated instances of ascend-
ing and descending division rates in fairly regular alter-
nate succession. The descending division rate is stopped
by an ‘‘internal regulatory phenomenon, endomixis.’’’*
Woodruff and Erdmann, while showing that endomixis
is different from conjugation in the absence of a syn-
caryon, apparently accept it as equivalent to conjugation
in connection with vitality of the protoplasm:

Endomixis and conjugation may occur simultaneously in different
animals of the same culture, thus strongly suggesting that the same gen-
eral conditions lead to both phenomena—one animal meeting the con-
ditions one way and another by the other, and that both phenomena
fill essentially the same place in the economy of life of Paramecium
awrelia.®
Again they say:

Endomixis does initiate a new rhythm in the life history of Para-
mecium, i. e., a period of increased metabolic activity and therefore of
reproductive activity, and since its fundamental morphological features
are almost identical with those preliminary to the formation of the
stationary and migratory micronuclei in conjugation, it lends strong
support to the view that the dynamic aspect of conjugation is not
absent.1®
Throughout the long period of seven years the Para-
megin aureha Pope without conjugation: ‘‘has
, undoubtedly on the aver-
age once each meea (ibid., p. 495). Hertwig has
already shown, as I do above, that asexual endomixis is
parthenogenesis, and if, in connection with the problem
of vitality, this is equivalent to conjugation, then we are

14 Ibid., p. 497.

15 Ibid., p. 492; the italies at the end are mine.
16 Tbid., p. 496.

No. 578] CYCLES AND RHYTHMS 75

justified in saying that throughout the seven years Wood-
ruff’s Paramecium has undergone the equivalent of con-
jugation on the average once each month, and if it is
equivalent to conjugation, then his long culture of more
than 4500 generations has no bearing on the question of
old age and natural death in Paramecium.

Nothing in this work of Woodruff and Erdmann seems
more clearly and forcibly demonstrated than that the cycle,
this ‘‘phantom’’ of many investigators, resolves itself
into a demonstrated fact, and that Woodruff’s ‘‘rhythm’’
and Calkins’s ‘‘eycle’’ are but different names for the same
phenomenon. If natural death is a necessary end to jus-
tify our use of the term ‘‘cycle,’’ we may ask the perti-
nent question: What happened to those individuals which
did not undergo asexual endomixis in Woodruff’s long
culture? If they died, does not this fact indicate the end
of a cycle? If they underwent parthenogenesis, the
equivalent of conjugation, does not this fact indicate the
beginnings of new cycles? If they continued to live with-
out reorganization, evidence for which has never been
given by Woodruff, then there would be some justification
for our authors’ conclusion. To argue that it is the same
race which continues after asexual endomixis is to use
the same argument that Weismann used unsuccessfully,
viz., that an ex-conjugant is the same old individual since
no corpse has been formed and therefore the infusoria
are immortal.

The frequent statement made by Woodruff that his
long culture sustains the view that old age and need of
conjugation are not necessary attributes of living matter
are contradicted by these later results. For example, he
states in 1913:

Diese Untersuchung hat uns gezeigt, dass, unter günstigen äusseren
Umständen, das Protoplasma der zuerst isolierten Zelle mindestens die
Potenz hatte, ähnliche Zellen bis zu einer Zahl von 2%*4° und eine
Masse Protoplasma von mehr als 10*°°° mal der Masse des Erdballes
zu erzeugen. Dieses Resultat, glaube ich, bestätigt unzweifelhaft die
Annahme, dass das Protoplasma einer einzigen Zelle unter günstigen

76 THE AMERICAN NATURALIST [Vou. XLIX

äusseren Umständen ohne Hilfe von Konjugation oder einer künstlichen
Reizung imstande ist, sich unbegrenzt fortzupflanzen und zeigt ferner in
klarer Weise, dass das Altern und das Befruchtungsbediirfnis nicht
Grundeigenschaften der lebendigen Substanz sind.17

I am entirely in sympathy with Hertwig when he says,
in connection with this citation:

Nach meiner Ansicht sind die Resultate, zu denen in den unseren
Auseinandersetzungen zum Ausgangspunkt dienenden Artikel Wood-
ruff gemeinsam mit Rhoda Erdmann gelangt ist, mit den hier zitierten
Sätzen unvereinbar.1§

The discovery of parthenogenesis in the life cycle of
Paramecium aurelia by Woodruff and Erdmann clears up
the obscurity which has involved all theoretical discus-
sions following pedigreed culture work with infusoria,
and we now see with much clearer vision the probability,
first, that conjugation or its equivalent has primarily the
result, as originally interpreted by Biitschli, of offsetting
and overcoming the progressive weakening of vitality in
infusoria; second, that more or less definite cycles of
vigor and depression, ending in natural death unless con-
jugation or its equivalent supervenes, are characteristic
of all pedigreed races of infusoria; third, that physical
‘‘immortality’’ is true of Paramecium and other ciliates
only in the same sense that it is true of metazoa; fourth
and last, that Paramecium protoplasm is subject to the
same laws of physiological usury that apply to metazoa,
and undergoes phenomena which, in metazoa, we call old
age, and which, as in metazoa, ends in natural death un-
less conjugation, or its equivalent parthenogenesis, saves

e race.

CoLUMBIA UNIVERSITY

17 Dreitausend und dreihundert Generationen von Paramecium u. s. W.,’
Biol. Centr., Vol. 33, No. 1, 1913, p. 35.

18 ‘‘ Ueber Piithansgencels der Infusorien,’’ ete., Biol. Centr., Vol. 34,
No. 9, 1914, p. 577.

THE PHENOMENON OF SELF-STERILITY!

PROFESSOR E. M. EAST

BUSSEY INSTITUTION, HARVARD UNIVERSITY

In both animals and plants in which the two sexes have
been combined in the same individual, cases have been
found where self-fertilization is practically impossible.
This gametic incompatibility has been called self-sterility,
although the term is hardly proper as applied to normal
functional gametes that may fuse with their complements
in the regular manner, provided each member of a pair
has been matured in a separate individual.

In plants the phenomenon has been known since the
middle of the nineteenth century, in animals a correspond-
ing discovery was made in 1896 by Castle, the species
being one of the Ascidians, Ciona intestinalis. During
the eighteen years that have passed since Castle’s dis-
covery, Ciona has been studied on a large scale by Morgan
(1905), Adkins (Morgan, 1913), and Fuchs (1914). The
botanists, however, have lagged somewhat behind; for, in
spite of having been acquainted with self-sterility in
plants for over half a century, and having found over
thirty species where a greater or less degree of self-
sterility occurs from which to select material, very few
thorough investigations into the physiology of the subject
have appeared.

The main facts regarding fertilization in Ciona intesti-
nalis are about as follows:

1. Under uniform suitable conditions, individuals vary
in degree of self-sterility, it being exceptional to find an
animal that is perfectly self-sterile.

2. Self-fertility has never equaled cross-fertility, though
the possibility remains that some animals may be self-

ad by title at the thirty-second ya of the American Society of
Aae December 31, 191 bh:
il

78 THE AMERICAN NATURALIST [Vou. XLIX

fertilized as easily as they may be crossed with certain
particular individuals.

3. The ease with which the ova of any animal ‘‘A’’ may
be fertilized by the sperm of other individuals may vary.

Morgan (1913) concluded from his own work and that
of Adkins that there were wide differences in the compati-
bility of ova to different sperm. Fuchs (1914) maintained
that 100 per cent. of segmenting eggs can be obtained in
every cross if the ova are normal and a sufficiently con-
centrated sperm suspension is used. It is possible that
Fuchs is correct and that varying concentrations of sperm
suspension were the cause of Morgan’s and Adkins’s re-
sults, yet the possibility of differences in this regard in-
herent in the individual is not to be overlooked. It will
be seen later that I regard the matter as of great impor-
tance to the general subject.

4, A chemical basis for self-sterility is shown in Fuch’s
experiments by (a) the decrease in ease of cross-fertiliza-
tion after contact of ova with sperm from the same ani-
mal, and by (b) the difference in ease of self-fertilization
after various artificial changes in the chemical equilibrium
of the medium surrounding the ova.

From the botanical side various studies on the physiol-
ogy of self-sterility have appeared since such investiga-
tions were initiated by Hildebrand in 1866. At this time
itis necessary for us to consider only those of Jost (1907),
Correns (1912), and Compton (1913).

Jost was able to show that in self-sterile plants tubes
formed from their own pollen were so limited in their
development that fertilization did not occur, although the
necessary length of pollen tube was easily developed after
a cross-fertilization. He saw as the cause of these phe-
nomena the presence of ‘‘ individueller Stoffe.’’ Pollen
was indifferent to ‘‘Individualstoff’’ of the same plant,
but was stimulated by that of other plants.

Correns (1912), working with one of the bitter cresses,
Cardamine pratensis, obtained results to which he gave a
simpler interpretation. Starting with two plants, B and

No. 578] SELF-STERILITY 79

G, he crossed them reciprocally and tested 60 of the off-
spring by pollinating from the parents, on the parents,
and inter se. The back crosses of (BX G) or (GX B)
with B and with G apparently indicated four classes about
equal in size with reference to gametic compatibility:
(1) plants fertile with both B and G; (2) plants fertile
with B but not with G; (3) plants fertile with G but not
with B; (4) plants fertile with neither B nor G.

To these facts Correns gave a Mendelian interpretation
by assuming the existence of two factors each of which in-
hibits the growth of pollen tubes from like gametes. Rep-
resenting these factors by the letters B and G, it is clear
that types BB and GG could never be formed. The orig-
inal plants were supposed to be of classes Bb and Gg, re-
spectively. When crossed there resulted the four types
BG, Bg, bG and bg. Plants of types BG, Bg, and bG
should be self-sterile, while plants of the type bg should be
self-fertile. Plants BG should be fertile with plants bg,
plants Bg should be fertile with bG and bg, and plants bG
should be fertile with Bg and bg. As a matter of fact
Correns’s results were not clearly in accord with the
theory. Plants of the type bg were not self-fertile, and
the other classes of matings showed many discrepancies.
It is only fair to say, however, that the author recognized
some of these difficulties, but believed them to be due to
other inhibitors.

In a part of Compton’s (1913) work, a still simpler
interpretation of. self-sterility is offered, at least for a
particular case, that of Reseda odorata. Darwin’s origi-
nal discovery that both self-sterile and self-fertile races
of this plant exist was confirmed and the following results
obtained in crossing experiments. Self-sterile plants
crossed either with self-sterile or with self-fertile plants
gave only self-sterile offspring. Certain self-fertile
plants, however, gave only self-sterile offspring when self-
pollinated. Other self-fertile plants gave ratios of 3 self-
fertile to 1 self-fertile offspring when self-pollinated, and
ratios of 1:1 when crossed with pollen from self-sterile

80 THE AMERICAN NATURALIST [Vow XLIX

plants. For these reasons he regards self-fertility as a
simple Mendelian dominant to self-sterility in the case
studied. I believe Compton would draw no such sharp
line about self-sterility in general. In fact, he follows
Jost in suggesting the presence of a diffusible substance
in the tissues of the style and stigma which retards or
promotes pollen tube growth after self-pollination or
cross-pollination in some manner analogous to the mech-
anism that promotes animal immunity or susceptibility
after infection.

The only alternative general hypothesis has been pro-
posed by Morgan, and this can be discussed more advan-
tageously after the presentation of my own work, of which
only an abstract will be given at this time.

In 1909 I made a cross between a small red-flowered
Nicotiana, Nicotiana forgetiana (Hort.) Sand. and the
large white-flowered Nicotiana of the garden Nicotiana
alata Lk. and Otto. var. grandiflora Comes. All of the
plants of the F, generation appeared to be self-sterile.
Tests of Nicotiana forgetiana? have shown these plants
also to be self-sterile, but both self-fertile and self-sterile
plants of the other parent have been found. From data
gathered later, there seems to be no doubt that a self-
sterile plant of Nicotiana alata grandiflora was used in
the actual cross. This conclusion seems reasonable in
view of the fact that of over 500 plants of the F,, Fa F;
and F, generations tested, not a _— self-fertile plant
was found.

The plants of the F, generation were all vigorous and
healthy, and in spite of the fact that they resulted from a
species cross which Jeffrey claims always produces large
amounts of abnormal pollen, a large number of examina-
tions of pollen from different individuals showed from 90

2I thought originally that both of these species (East, 1913) were self-
fertile. Seed had been obtained from a carefully bagged inflorescence of
each species in 1909. Either the plant of N. forgetiana which gave this
seed was self-fertile—something that I have never been able to find since
that time—or there was an error in manipulation. At any rate, the plants
resulting from this seed were all self-sterile

No. 578] SELF-STERILITY 81

to 100 per cent. of morphologically perfect pollen grains,
a condition about the same as was found in the pure spe-
cies. To this statement there is one exception. A single
plant was found with only about 2 per cent. of good sound
pollen.

Several experiments were made in which crossing and
selfing was done on a large scale, using plants of the F,,
F, and F, generations which had segregated markedly in
size and were of at least 8 different shades of color. In
one of these experiments 20 plants of the F, generation
coming from 2 crosses of F, plants were used. It was
planned to make all possible combinations of these plants,
400 in all. This task proved overburdensome, however,
and in addition to the self-pollinations but 131 inter-
crosses were made with the following results.

1. Each plant was absolutely self-sterile.

2. Leaving out of consideration the plant with shrunken
imperfect pollen only two crosses failed. This failure of
1.5 per cent. of the crosses may have been due to im-
proper conditions at the time of the attempts, but as a
number of trials were made the possibility remains that
there is a small percentage of true cross-sterility.

3. Of the 129 successful inter-crosses, 4 produced cap-
sules with less than 50 per cent. of the ovules fertilized.
The remaining crosses produced full capsules. It is
barely possible that this result shows a slight variability
in ease of cross-fertilization, but I am more inclined to
believe that these 4 cases where a low percentage of fer-
tilized ovules were obtained were accidental.

Other crossing experiments of the same kind have cor-
roborated these results. Out of 120 inter-crosses, only 3
failed.

Later, something over 100 inter-crosses were made be-
tween 12 plants of an F, population resulting from cross-
ing two sister F, plants. Six of the attempts at cross-
fertilization—3 to 8 trials per plant being made—were
failures. These plants as well as others tested were com-

82 THE AMERICAN NATURALIST [Vou. XLIX

pletely self-sterile, and apparently there was cross-steril-
ity in about 6 per cent. of the possible combinations.

In the F, generation, 10 plants resulting from crossing
two sisters of the F, generation were selected for experi-
ment. Unfortunately, I was able to make only 58 inter-
crosses, 5 of which, almost 10 per cent., failed.

Back crosses have furnished another line of experiment,
though they have not been carried on as systematically as
were those of Correns. Nearly 85 back-crosses using
plants from the progeny of four combinations which: in-
cluded four individuals as parents, have been made. The
plants themselves all proved self-sterile, and in addition
5 of the back crosses failed.

When these experiments were begun I expected to find
that the facts would accord with a simple dihybrid Men-
delian formula similar to that which Correns later pro-
posed as an interpretation of his results, yet only by con-
siderable stretching and a vivid imagination will Cor-
rens’s data fit such an hypothesis, and my own data do
not fit at all. No self-fertile plants have been produced
by any combination, and cross-sterility is a possibility in
only from 1.5 to 10 per cent. of the combinations. Fur-
thermore, Correns’s idea of inhibitors appears unlikely
from some other data I have gathered with the help of
Mr. J. B. Park. Ten plants were involved in this experi-
ment. Pairs of plants were provided to furnish series
of selfed and crossed flowers. The pistils of these flowers
were fixed at regular periods after pollination, stained,
sectioned, and the pollen tubes examined. Fertilization
not later than the fourth day marked the end point of the
crossed series, the dropping of the flowers between the
eighth and the eleventh day ended the selfed series. As
the flowers on each plant had about the same length pistils,
curves of pollen tube development for both crossing and
selfing could be constructed. The pollen grains germi-
nated perfectly on stigmas from the same plant, from
1,200 to 2,000 tubes having been counted in sections of
single pistils. The difference between the development

No. 578] SELF-STERILITY 83

of the tubes in the selfed and the crossed styles is wholly
one of rate of growth. The tubes in the selfed pistils de-
velop steadily at a rate of about 3 millimeters per twenty-
four hours. There is even a slight acceleration of this
rate as the tubes progress. If the flowers were of an
everlasting nature one could hardly doubt but that the
_ tubes would ultimately reach the ovules, though this would
not necessarily mean that fertilization must occur. Since
the maximum life of the flower is about 11 days, however,
the tubes never traverse over one half of the distance to
the ovary. On the other hand, the tubes in the crossed
pistils, though starting to grow at the same rate as the
others, pass down the style faster and faster, until they
reach the ovary in four days or less.

From these facts it seems reasonable to conclude that
the secretions in the style offer a stimulus to pollen tubes
from other plants rather than an impediment to the de-
velopment of tubes from the same plant.

The whole question, therefore, becomes a mathematical
one, that of satisfying conditions whereby no stimulus is
offered to pollen tubes from the same plant, but a positive
stimulus is offered to tubes from nearly every other plant.

Morgan has given an answer to this question in a gen-
eral way. If I understand his position correctly, my own
conclusions are not very different from his, but are some-
what more definite. Morgan (1913) states that the re-
sults of Adkins and himself on Ciona intestinalis can best —
be understood by the following hypothesis:

The failure to self-fertilize, which is the main problem, would seem
to be due to the similarity in the hereditary factors carried by the eggs
and sperm; but in the sperm, at least, reduction division has taken
Place prior to fertilization, and therefore unless each animal was
homozygous (which from the nature of the case cannot be assumed
possible) the failure to fertilize can not be due to homozygosity. But
both sperm and eggs have developed under the influence of the total or
duplex number of hereditary factors; hence they are alike, i. e., their pro-
toplasmie substance has been under the same influences. In this sense,
the case is like that of stock that has long been inbred, and has come
to have nearly the same hereditary eee If this similarity decreases

84 THE AMERICAN NATURALIST [Vov. XLIX

the chances of combination between sperm and eggs we can interpret
the results.

I make this quotation to show Morgan’s viewpoint. It is
for him to say whether the following conclusions are ex-
tensions of his own or not.

The tolerably constant rate of growth of pollen tubes in
the pistils of selfed flowers, compared with the great ac-
celeration of growth of the tubes from the pollen of other
plants as they penetrate nearer and nearer to the ovary,
undoubtedly shows the presence of stimulants of great
specificity akin to the ‘‘Individualstoffe’’ of Jost. We
are wholly ignorant of the nature of these stimulants, but
I am inclined towards a hypothesis differing somewhat
from his. Experiments by several botanists, which I
have been able partially to corroborate, point to a single
sugar, probably of the hexose group, as the direct stimu-
lant. The specific ‘‘Individualstoffe’’ I believe to reside
in the pollen grains and to be in the nature of enzymes of
slightly different character, all of which, except the one
produced by the plant itself for the use of its own pollen
or by other plants of identical germinal constitutions,
can call forth secretion of the sugar that gives the direct
stimulus. At least this idea links together logically the
fact of the single direct stimulus and the need of ‘‘Indi-
vidualstoffe’’ to account for the results of the crossing and
selfing experiments. But whether or not this be the cor-
= rect physiological inference, the crossing and selfing ex-
periments call for a hypothesis that will account for no
stimulation being offered the tubes from the plant’s own
pollen, while at the same time great stimulation is given
the tubes from the pollen of nearly every other plant.

This is a straight mathematical problem, and it is
hardly necessary to say that it is insoluble by a strict
Mendelian notation such as Correns sought to give. This
is obvious to any one familiar with the basic mathematics
of Mendelism. On the other hand, a near Mendelian in-
terpretation satisfies every fact.

Let us assume that different hereditary complexes stim

No. 578] SELF-STERILITY 85

ulate pollen tube growth and in all likelihood promote fer-
tilization, and that like hereditary complexes are without
such effect. One may then imagine any degree of hetero-
zygosis in a mother plant and therefore any degree of
dissimilarity between the gametes it produces, without
there being the possibility of a single gamete having any-
thing in its constitution not possessed by the somatic tis-
sues of the mother plant. From. the chromosome stand-
point of heredity the cells of the mother plant are duplex
in their organization; they contain N pairs. The cells
of the gametes contain N chromosomes, one coming from
each pair of the mother cell; but they are all parts of the
mother cell and contain nothing that that cell did not con-
tain. These gametic cells can not reach the ovaries of
flowers on the same plant because they can not provoke
the secretion of the direct stimulant from the somatic cells
of that plant.

All gametes having in their hereditary constitution
something different from that of the cells of a mother
plant, however, can provoke the proper secretion to stim-
ulate pollen tube growth, reach the ovary before the flower
wilts and produce seeds. Such differences would be very
numerous in a self-sterile species where cross-fertilization
must take place; nevertheless like hereditary complexes:
in different plants should be found, and this should ac-
count for the small percentage of cross-sterility actually
obtained. It must be granted that this hypothesis satis-
fies the facts, but that is not all. It is admittedly a per-
fectly formal interpretation, but from a mathematical
standpoint,—granting the generality of Mendelian inheri-
tance,—it is the only hypothesis possible that can satisfy
the facts.

Let us now look into a few of the ramifications of the
subject. Examinations of the pistils that have been sec-
tioned after cross-pollination show a considerable varia-
tion in the rate of growth of individual pollen tubes,
though our curves of growth have been made by taking
the average rate of elongation. Is this variation a result

86 THE AMERICAN NATURALIST [Vou XLIX

of chance altogether or must we assume a differential rate
of growth increasing directly with the constitutional dif-
ferences existing between the somatic cells and the vari-
ous gametes? If we assume that any constitutional dif-
ference between the two calls forth the secretion of the
direct stimulus to growth, chance fertilization by gametes
of every type different from that of the mother plant will
ensue; if there is a differential rate, selective fertilization
will occur. This question can not be decided definitely at
present, but two different lines of evidence point toward
chance fertilization.

1. Flowers from a single plant pollinated by different
males show no decided difference in rate of fertilization.

2. Color differences are transmitted in expected ratios.

Further, it will be recalled that beginning with the F,
generation, sister plants crossed together have given us
our F, and F, populations, and that these F, and F, popu-
lations apparently have given a constantly increasing per-
centage of cross-sterility. This is what should be ex-
pected under the theory that a small difference in germ
plasm constitution is as active as a great difference in
causing the active stimulation to pollen tube growth. -
Breeding sister plants together in succeeding generations
causes an automatic increase of homozygosity as is well
known. This being a fact, cross-sterility should increase.
Such an increase in cross-sterility has been observed in
the F, and the F, generations, but it would not be wise to
maintain dogmatically that it is significant.

There are various questions, including the important
one of the origin of self-sterility, that can not be discussed
at this time. In conclusion, therefore, let us turn once
more to the phenomenon of self-sterility in Ciona intes-
tinalis. It seems to me that the hypothesis outlined above
has few, if any, drawbacks when applied to self-sterility
in plants. The question there, as far as we have gone, is
one of pollen tube growth, and the theory that the secre-
tion of the direct stimulant can be called forth only by a
gamete that differs in its constitution from the somatic

No. 578] SELF-STERILITY er. «|

cells between which the pollen tube passes, is logical. If
the same theory is to be extended to animals, however, it
follows that the external portions of the membranes of
the animal egg that have been shown by the wonderful in-
vestigations of Loeb and of Lillie to have such important
functions, must be functionally zygotic in character. I
am aware that this suggestion may be considered pretty
radical, but it certainly should be given consideration.
I do not like to draw an analogy between the animal egg
and a pollen grain, but it may be mentioned that in these
structures—surely comparable to the animal egg in the
fineness of their membranes and walls—both color and
shape are inherited as if they were zygotic in nature.
December 5, 1914.

LITERATURE CITED.

Castle, W. E. The Early Embryology of Ciona intestinalis Flemming (L.).
Bull. Mus. Comp. Zool., Harvard University 27, 201-280. 1896.

Compton, R. H. Phenomena and Problems of Self-sterility. New Phytolo-
gist, 12, 197-206. 1913.

Correns, ©. Selbststerilitiit und Individalstoffe. Peéstechr. d. mat.-nat.
Gesell. zur 84. Versamml, deutsch. Naturforscher u. Arzte, 1912.
Miinster i, W., pp. 1-32.

East, E. M. Inheritance of Flower Size in Crosses between Species of
Nicotiana. Bot. Gaz., 55, 177-188. :

Fuchs, H. M. On the Conditions of Self-fertilization in Ciona. Archiv.
f. Entwickl. d. Org., 40, 157-204.

pee i Action of ter Somat a on the Fertilizing Power of Sperm.

52. 1914.

~fruchtung von Corydalis cava. .
Jost, L. Zur Physiologie des Pollens. Ber. i deut. bot. Gesell., 23, 504-
515. 1905.

—— Ueber die PAET einiger Blüten. Bot. Zig., Heft V and VL
1907.
Morgan, T. H. Some Further Experiments on Self-fertilization in Ciona.

Biol. Bull., 8, Stags 1905.
SETIN Heredity ù d Sex. New York. Columbia Univ. Press, pp. ix + 1-

282. 1913 rek cited 217).

THE BLACK-AND-TAN RABBIT AND THE SIG-
NIFICANCE OF MULTIPLE ALLELOMORPHS

W. E. CASTLE AND H. D. FISH

Bussey Institution, Forrest Hiuus, Mass.

Ir is well known that the European rabbit has under-
gone great variation, and now exists in a large number of
domesticated varieties. Darwin and most other natural-
ists speak of this as ‘‘variation under domestication,’’
implying that domestication has caused the variation.
Modern genetic research, however, indicates that domesti-
cation has occasioned the preservation rather than the
origin of the fundamental variations involved. But to
what extent man through selection is able to modify the
fundamental variations which nature occasionally pro-
duces as sports is still an open question. Evidence is
nevertheless accumulating that certain of these funda-
mental variations may occur in two or more alternative
forms, and the question then arises (1) whether these al-
ternative forms have arisen independently by distinct acts
of mutation, or (2) whether one has arisen from another
by a process of secondary mutation, or (3) whether one
may not have been transmuted into another by a more or
less gradual process. Toward the testing of these sev-
eral hypotheses much genetic research is now being di-
rected. The first step to be taken is evidently to ascertain
in how many alternative forms the same fundamental
variation may occur and how these forms are inter-
related. A further step will be the attempt to produce
new alternative forms at will. It is our purpose, in this
paper, to discuss a newly discovered alternative form
(allelomorph) of the gray, or agouti, type of coat found
in wild rabbits. It occurs in the variety known as black-
and-tan.

This variety appears to have arisen from the wild gray,
or agouti, type without the loss of any known genetic fac-
tor, but by a modification in one. Simple loss of genetic
factors is believed by most students of genetics to have

88

No. 578] MULTIPLE ALLELOMORPHS 89

given rise to black, chocolate and albino varieties of rab-
bits and other rodents, but a hypothesis of this sort will
not fit the present case. No factorial loss can be detected,
but only a change in that genetic factor which has been
called the agouti or gray factor. Under the influence of
this factor, what would otherwise be a black variety be-
comes gray, and what would otherwise be sooty yellow
(‘‘tortoise’’ of the fanciers) becomes clear yellow (‘‘fawn’’
of the fanciers). This same factor converts chocolate
into cinnamon (Punnett, 1912). In every way, accord-
ingly, its influence on the coloration of rabbits is similar
to that of the agouti factor in guinea-pigs and mice.

But in mice Cuénot (1909) showed that the agouti factor
may assume three distinct forms allelomorphic to each
other, the effects of which are seen respectively in gray,
light-bellied gray, and in yellow mice. Entire absence of
agouti marking from the fur (non-agouti) forms a fourth
allelomorph in the series.

In guinea-pigs the agouti factor assumes two alterna-
tive conditions, the effects of which are seen in ordinary
(light-bellied) agoutis and in agoutis with ‘‘ticked’’ bel-
lies, respectively (Detlefsen, 1914). These two conditions
correspond closely in appearance and in order of domi-
nance to the light-bellied gray and the ordinary gray of
mice, the former being dominant in both cases. Non-
agouti is an allelomorph to both, as in mice.

The peculiarity of the agouti seen in black-and-tan rab-
bits is that it produces less extensive ticking of the fur
than does ordinary agouti. In a typical black-and-tan
rabbit the light-colored (yellowish) bands on the hairs,
which constitute the ‘‘ticking,’’ occur only sparingly on
the sides of the body, and not at all on the back or the
head. But the under side of the body, including the
throat and under surface of the tail, are light (yellowish
or whitish) and the back of the neck and inside of the ears
bear reddish or yellowish pigment, as in gray rabbits.
The typical black-and-tan rabbit of the fanciers has very
intense pigmentation which deepens the shade of the
“tan” (yellow) found on belly, sides, ete. But this inten-

90 THE AMERICAN NATURALIST [Vou. XLIX

sity is inherited independently of the agouti factor as
crosses with dilute colored varieties of rabbit show. For
a cross between black-and-tan and blue produces in F,
(1) blue-and-tans as well as (2) black-and-tans, (3) blacks,
and (4) blues. This result is strictly parallel with that
obtained by crossing intense gray rabbits with blue ones.
In that case there are produced (1) blue gray, (2) intense
gray, (3) black, and (4) blue young in F,. It is evident
that in each case a dihybrid cross is made and that the end
products are the same in the two series except for the dif-
ference in the agouti marking of varieties (1) and (2).
The natural conclusion is that. black-and-tan contains an
alternative form of agouti to that found in gray rabbits.
If so, it should be capable everywhere of substitution for
gray, wherever the latter occurs throughout the entire
series of color varieties, and indeed this appears to be the
case.

That the black-and-tan factor, like the ordinary agouti
factor, is independent of the extension-restriction pair of
allelomorphs is shown by a cross of black-and-tan with
sooty yellow (i. e., non-agouti yellow or ‘‘tortoise’’).
F, contains (1) black-and-tan, (2) black, (3) yellow
(‘‘fawn’’), and (4) sooty yellow (‘‘tortoise’’) young.
The first two are varieties with extended pigmentation,
and the second two are varieties with restricted pigmen-
tation; further, varieties (1) and (3) contain modified
agouti, but varieties (2) and (4) do not.

If a gray rabbit had been used, instead of a black-and-
tan, in making the cross just described, three of the four
varieties obtained in F, would have been indistinguish-
able from those enumerated, and the fourth one would
merely have been gray instead of black-and-tan. This
supports the view that black-and-tan is merely an alterna-
tive form of gray.

Further, we have evidence to show that the black-and-
tan form of agouti, like the agouti of gray rabbits, is inde-
pendent of the genetic factors which respectively produce
Dutch pattern, English pattern, and angora coat, since we

have been able to coe aa individuals in which black-and-

No. 578] MULTIPLE ALLELOMORPHS 91

tan was associated with each one of these Mendelizing
characters, as well as others in which it was not associated
with them. Finally Haecker (1912) has shown that black-
and-tan, like the ordinary form of agouti, is independent
of albinism, since when black-and-tans are crossed with
Himalayan albinos, not only these two varieties are ob-
tained in the F, generation, but also blacks. The propor-
tions in which these three varieties were obtained by
Haecker approximate the modified dihybrid ratio, 9 black-
and-tan: 3 black: 4 Himalayan. One of the two Men-
delian pairs concerned is color vs. albinism; the other and
independent one, black-and-tan vs. black.

It is known that if a gray rabbit is used, instead of a
black-and-tan one, in a cross with Himalayan albinos, the
same 9:3:4 ratio is obtained in F,, of grays, blacks, and
albinos, respectively. The observed results differ, in the
two cases, only in the substitution of gray for black-and-
tan, which is further evidence that it is only another form
of the same genetic factor.

Notwithstanding all this consistent and converging evi-
dence, it is possible that the modified form of agouti seen
in black-and-tan is not due to a changed agouti factor
itself, but to the modifying action of a factor associated
with it which partially inhibits its action. Here we must
consider two subordinate possibilities: (a) that the sup-
posed modifier is wholly independent of the agouti factor,
and (b) that it is coupled with the agouti factor. The
first possibility is readily disproved; the second one is not
so easily disposed of.

(a) If black-and-tan were due to the action of an inde-
pendent modifying factor associated with agouti, a cross
of black-and-tan with ordinary black should permit the
separation of agouti from its supposed modifier in a con-
siderable part of the F, gametes and F, zygotes, so that
we should expect F, to contain gray animals as well as
blacks and black-and-tans. But experiments started sev-
eral years ago at the Bussey Institution show that when
black is crossed with black-and-tan no gray offspring are
_ obtained either in F, or in F,, but a black-and-tans in

92 THE AMERICAN NATURALIST [Vow XLIX

F,, and black-and-tans and blacks in F,. This result
shows that black-and-tan is a simple dominant over black.
To establish the allelomorphism of black-and-tan with
gray the following experiments may be cited. A black-
and-tan rabbit heterozygous for black was crossed with
a pure-bred Belgian hare, which variety possesses the
genetic color factors of wild rabbits, including the ordi-
nary agouti factor. All the F, young were gray, closely
resembling Belgian hares, but proved to be genetically
of two types. For, when mated with black rabbits, some
of them produced gray young and black young, while
others (even when mated, as in some cases, with the same
black animals) produced gray young and black-and-tan
young. This result was quite what was to be expected if
_ gray, black-and-tan and black are mutually allelomorphic
conditions. On no other hypothesis which we can sug-
gest was it to be expected. For the black-and-tan parent
in the cross was known to be heterozygous for black. It
accordingly should form two sorts of gametes, black and
black-and-tan respectively, provided that these conditions
are allelomorphic to each other. The Belgian hare parent
was known to transmit gray in all its gametes. The com-
binations expected from the crdss are therefore of two
types, viz.: (1) gray combined with black, and (2) gray
combined with black-and-tan. It is well known that gray
and black are allelomorphs of each other, the former being
dominant. Zygotes of type (1), therefore, should pro-
duce gametes of two sorts, gray and black; and when
back-erossed with black should produce equal numbers of
gray young and black ones but no black-and-tan young.
We have tested 12 F, gray young from this cross (6 males
and 6 females) which are evidently of type (1). Mated
with black animals, they have produced 69 gray young,
and 65 black ones, but no black-and-tans. ©
-On the other hand 8 F, gray rabbits from the cross
under discussion have proved to be of type (2), producing
gray young and black-and-tan young but no black ones.
Together they have produced 44 gray and 51 black-and-
tan young, besides 14 other young (two litters) which

No. 578] MULTIPLE ALLELOMORPHS 93

were certainly not blacks, since they had light bellies,
but which died before attaining the age at which gray
can be distinguished from black-and-tan. It is certain
that among the 109 young produced by the 8 animals of
type (2) not a single one was black.

But if black-and-tan is not an actual allelomorph of
gray, black young as well as black-and-tans should have
been produced in the foregoing case. For if black-and-
tan is not allelomorphic with gray, or is due to an inde-
pendent inhibitor of gray, then an F, gray should produce
gametes of four sorts, rather than as indicated of two
sorts; t. e., gametes should arise which transmit both gray
and black-and-tan, and others which transmit neither gray
nor black-and-tan. The former sort possibly might not
be capable of immediate detection in the back-cross with
black, but the latter should be readily discovered since
they would necessarily produce black young (neither gray
nor black-and-tan). The total absence of black young
from the litters produced by type (2) matings therefore
indicates strongly that gray and black-and-tan are allelo-
morphs of each other.

(b) An alternative view, however, deserves. considera-
tion. If gray and black-and-tan are not actual allelo-
morphs, it is conceivable that they may each be closely
‘‘coupled’’ with a common structure in the germ cells and
so behave as allelomorphs under ordinary circumstances,
though not being such in reality. Or, what would give
the same practical result, gray and black-and-tan might
be supposed to contain the same agouti factor, but this
might be considered in one as closely coupled with a modi-
fying factor which made its action different. Neither
form of this hypothesis is capable of proof or disproof,
for which reason alone the hypothesis is unimportant, but
its probability grows less the larger the number of records
obtained which show no breaking of the supposed coup-
ling. Our cases are not as yet numerous enough to throw
much light on this question, but so many cases have
already been discovered in which characters assume three
or more mutually allelomorphic conditions and in which —

94 THE AMERICAN NATURALIST [Vou. XLIX

no evidence of coupled modifiers has yet been discovered,
that the existence of such assumed modifiers seems at
present doubtful.

Besides the triple or quadruple series of agouti allelo-
morphs now known for mice, guinea-pigs and rabbits, at
least three other Mendelian factors concerned in the pig-
mentation of rodents vary discontinuously in this way.

1. Castle (1905) and Punnett (1912) have shown that
the Himalayan rabbit possesses a form of albinism allelo-
morphic with that of ordinary albino rabbits, and that
both are allelomorphic to ordinary pigmentation. Guinea-
pigs show an even more extended series of albino allelo-
morphs (Castle, 1914, Wright, unpublished data).

2. Punnett (1912) has discovered in rabbits an alterna-
tive form of the ‘‘extension’’ factor, one in the presence
of which the agouti factor produces a less amount of tick-
ing than normally. He describes it as a darkened exten-
sion, č. e., as ordinary extension modified by a coupled
darkening factor. This is of course only an alternative
form of statement to saying that extension occurs in two
forms, for he discovered no cases in which the hypothet-
ical coupling was broken. The three allelomorphs in the
ease of Punnett’s rabbits were accordingly: 1, ordinary
extension; 2, darkened extension, and 3, restriction.

3. In still another Mendelian factor affecting the pig-
mentation of rodents discontinuous variation occurs at-
tended almost certainly by the formation of a series of
allelomorphs. Cuénot (1904) stated that white-spotting
in mice occurs in a graded series of conditions as regards
the amount or extent of the white areas. He found that
widely separated stages in the series Mendelize on cross-
ing, 7. e., that the segregates fluctuate about modal con-
ditions corresponding roughly with the conditions of
spotting found in the respective parents crossed, and he
concluded the number of allelomorphs which it would be
possible to find in the series to be indefinitely great. Sub-
sequent studies of the subject made by Little (1914) in
mice, and by Castle and Phillips (1914) in rats, have not
served to simplify the matter, and yet they confirm Cué-

No. 578] MULTIPLE ALLELOMORPHS 95

not’s general idea that a series of mutually allelomorphic
conditions of spotting exists. Unquestionably in rats,
‘*hooded’’ and ‘‘Trish’’ are such modal conditions of
spotting, allelomorphic with each other and with the un-
spotted or self condition (Doncaster, 1905; MacCurdy
and Castle, 1907; Castle and Phillips, 1914). The last-
named authors find that independent factorial modifiers
probably affect the extent of the spotting and yet .that,
aside from such modifiers, the spotting factor proper may
assume relatively stable allelomorphic conditions which
Mendelize when crosses are made between stages suffi-
ciently distinct. The point of especial interest in allelo-
morphic conditions of spotting is that they are not per-
fectly stable, but are capable of gradual and apparently
indefinite modification through the selection of fluctua-
tions either plus or minus. It would be premature to
conclude that similar fluctuations (though perhaps less
conspicuous ones) do not occur about the modal condi-
tions of other genetic factors which show allelomorphic
variation. The black-and-tan form of agouti certainly
fluctuates in the amount of ticking found on the sides of
the body and the head; doubtless some of this fluctuation
may be due to factors genetically distinct from the chief
allelomorphic factor concerned, but there is at present no
sufficient ground for supposing the chief factor itself to
be incapable of fluctuation. Indeed, it seems highly prob-
able, in the light of evidence already obtained, that the
present modal condition of the black-and-tan character is
one which has been attained only as a result of persistent
selection, and that.reversed selection will carry it back
appreciably nearer to the modal condition seen in gray
rabbits. Accordingly, it appears doubtful whether allelo-
morphs are themselves perfectly and permanently stable.
Moreover, the rapid increase of recognized allelomorphs
makes us wonder whether their number is limited and defi-
nite. Black-and-tan represents, on the whole, an inter-
mediate condition between black and gray. Is it not con-
ceivable that intermediates may yet be discovered be-

96 THE AMERICAN NATURALIST [Vou XLIX

tween black-and-tan and black, or between black-and-tan
and gray, or even that black-and-tan itself might be dis-
placed to such an intermediate condition by selection of
its fluctuations? Here are fruitful fields of inquiry to be
cultivated before we conclude with the exponents of ‘‘ex-
act’’ heredity that selection of fluctuations is useless and
that only mutations count in evolution.
December 26, 1914.

PAPERS CITED
Castle, W. E.
1905. Heredity of Coat Characters in Guinea-pigs and Rabbits. Publ.
No. 23, Carnegie Institution of Washington
1914, ere New Varieties of Rats and Gatnpa-piga and their Rela-
tions to Problems of Color Inheritance. AMER. NAT., 48, pp.

65-73.
Castle, W. E., and J. C. Phillips.
1914. Pi ebald Rats and Selection. Publ. No. 195, Carnegie Institu-
tion of Washington
Cuénot, L
1904. L’hérédité de la pigmentation, chez les souris (3me Note).
. de Zool. Exper. et gen., Notes et Rev., p. xlv.
1909. Recherches sur l’hybridation. Proe. Seventh Internal, Zool.
Congress, pp. 45-56. (Prize essay submitted to the Congress,
August 1907.)
Detlefsen, J. A.
1914. Publ. No. 205, Carnegie Institution of Washington.
Doncaster, L.
1905. On the Inheritance of Coat-color in Rats. Proe. Camb. Phil.
Soc., 13, p. 215
Haecker, V.
1912. Ueber Kreuzungsversuche mit Himalaya und Black-and-Tan
aninchen. Mitt. d. Naturf. Gesell. zu. Halle, 2 Bd., 1912,

pp. 1-4
Little, C. C.
1914. ‘*Dominant’’ and ‘‘ Recessive’? Spotting in Mice. AMER. NAT.,
48, pp. 74-82.

MacCurdy, H., and W. E. Castle
1907. Selection and Cross-breeding in Relation to the Inheritance of
oat-pigments and Coat-patterns in Rats and Guinea-pigs.
Publ. No. 70, Carnegie Institution of Washington.
Punnett, R. C. ‘
1912. Inheritance of Coat-color in Rabbits. Journ. of Genetics, 2,

DATA ON A PECULIAR MENDELIAN RATIO IN
DROSOPHILA AMPELOPHILA

JOSEPH LIFF? .

A Mutant with pink eyes was found by Professor T. H.
Morgan in the summer of 1910, in one of his culture
bottles which contained wild, red-eyed Drosophila. He
described it as follows :?

‘‘The pink eye is more translucent than the red eye, but
of about the same general tone. It lacks the dark fleck
seen in the red and vermilion eye when the eye is exam-
ined with a lens. This black fleck changes its position as
the lens travels over the eye. The pink eye, P, is witha
little experience easily distinguished from the other colors,
especially in newly hatched flies. When the fly gets old
the eye turns to a brown color very characteristic of this
type of eye.’’

Pink was found to be recessive to red. The Mendelian
expectation in the F,, viz., three red to one pink, gave the
following (Morgan, 1911):

TABLE I
5 pi Fe Proportion of
1 Red | Pink Red: Pink
Red tk Ki ee | 3,063 | 169 18:1
Tek Oo xed n Sioa LAS 237 5:1

The expectation in either case was 3:1, but the num-
bers realized were’ 18:1 and 5:1.

In the spring of 1912 I repeated this experiment under
the direction of Professor Morgan in order to find
whether or not the above ratio would persist. The results
have already been published (Morgan, 1912), but a brief
summary is here reproduced for reference:

1 From the Zoological Laboratory of Columbia University.

2 Jour, Exp. Zoology, Vol. 11, No. 4, November, 1911.

97

98 Drar AMERICAN NATURALIST [Vou. XLIX

TABLE II
Red 9 X Pink g — in F2 Pink X pg Red o> in Fə
Bottle | Red | Pink | Se st: | Goito te | Pink | ger gietgl
| | | |
Pee A ae eect a... 541 | 124 bast
B.....| 140 | 34 | 4.1:1 Bi 190 1 68 | Bed
C....] 375 | wW | 5.3: 1 PE 582 | 136 | -43:1

The above are records of mass cultures. When pairs
were used, the fluctuations in ratio were much more
marked. The records of 40 pairs gave an almost un-
broken series running from 1.8:1 up to 6:1. In seven
cases out of the 40 (18 per cent.) the pink flies exceeded
the expectation; 3 pairs (7.5 per cent.) gave a 3:1 ratio,
while in the remaining 30 pairs (75 per cent.) the pink
fell behind. The total number produced by these 40 pairs
was 4,056, of which 891 were pink—an average ratio of
3.98: 1, about the same as that shown in Table IT

In a second experiment the F, hybrids were back-
crossed to the pink. The expectation was 1:1. But the
records of 15 bottles of mass culture showed fluctuations
running from 1:1 up to 2.3:1. The total number counted
in these back crosses was 5,527, of which 2,391 were pink,
giving an average ratio of 2.31:1. The pink flies fell
behind again, and in about the same proportion as in the
normal cross.®

These remarkable fluctuations were observed at the
time the experiments were in progress, and it was sug-
gested that some environmental condition was responsible
for the results by either accelerating or retarding‘ the
development of the one or of the other variety. The fact
that all these experiments were performed at the same
time, and the bottles kept side by side in a room in which
a nearly constant temperature was maintained through-
out the winter, precludes the chance of a factor outside
the culture bottles operating here. Attention was there-

8 For a detailed account of these experiments see Morgan, 1912.

4It should be noted here that owing to the danger of overlapping of
generations, the bottles were discarded on the tenth day (counting from the
day the first F, emerged) regardless of the number of unhatched pupe

No. 578] DROSOPHILA AMPELOPHILA 99

fore directed to the condition of the food inside the bottles,
An examination seemed to indicate that those in which the
food was dry yielded the higher pink ratios. To test this,
two of the bottles in which conditions were normal, and
in which the F, had just begun to emerge, were made
‘‘wet’’ by the addition of a considerable amount of banana
juice. But they still showed a similar tendency to yield
a relatively higher proportion of pink.

To ascertain more definitely whether or not moisture
or dryness affected in any way the development of these
flies, a special experiment was arranged in which some
flies were bred in ‘‘dry’’ bottles, and some in ‘‘wet’’
bottles. In the first case, the banana was thoroughly
dried by means of filter paper which was discarded after
it had absorbed all the available moisture, and the banana
wrapped in fresh paper; in the second, banana juice was
added every second or third day, so that there was
throughout the experiment an abundant amount of wet
food in the bottles. The effect of this treatment is shown
in Tables III and IV:

. TABLE III
RECORD oF F, or A Cross RED BY PINK IN WHICH THE FLIES DEVELOPED IN
BOTTLES IN WHICH THE Foop was ‘‘Dry’’

Pink 9 x Red # Red 9 X Pink 7
Er: | Í j
Red Pink | P $ Red Piok | P ti
Bottle roportion Bottle | ote pee on
y ayle | Red: Pink | a | ole | Red : Pink
A....| o] 86| 25|18| 42:1 |a....| 91/109] 28| 32] 33:1
Boiss 73| 78| 20 4 s:r fig 76) 05] | 35) 26:1
Cau 182|164| 39 | 44| 43:1 |e..... 42| 14 | 15 | 3.0:1
D -Iola ptei S- he ‘| yet 117| 55 |46 | 2.7:1
TABLE IV

RECORD oF F, oP a Cross RED BY PINK IN WHICH THE FLIES DEVELOPED IN
BOTTLES IN WHICH THE Foop Was ‘‘WET’’ FROM THE BEGINNING

Pink 9 x Red # Red 9 X Pink ð

Red Pink Red | Pink |
Bottle |-——— ee Lae | eee pe hy
e|e|ele ; gia el?
A.. -dorioa w) 28:1 |a....|82|89|17/17| 50:1
B.. J2 964-46) aed Bold [b -Jiel 8 | 8 6| 20:1
e S41 93 2213] 37:13 le.. 2} 90} 32 6| 60:1

[Vou. XLIX

THE AMERICAN NATURALIST

100

92 |29 |89|\ 92
99 | 18 |°°°)***|"**"T e1990q yonpeg
ZFT | SFT |89| 94| E£ | 43 | 9S | £8 | 68) G6) EST | LOT; O | S | T8| 9! ST) FI Z4 |901| $8) 98| SOT! LST! LOT | Z6| 06) Z2/°** “T8301,
. s.sejesoeoje o eae ao. eee ele ee elewes I 0 Z 0 . BE Eco led & 0 I 0 Z Z I 0 eee eleae ode Il m
+ lo lolze dg lods lo detet do lolala izle lo le -d-t ot
+ ile ielea lect lets lerlo lp tcletoteloe lolz lo le lo ler ler fect dela d
Let jole reye eoa oraa ee n 1g tk p a a a 16 12 TES p
OS | HT Gerjo a eI to lo Jore re Ot 16 o Bt inde oe (oe 19 it 7 R RE
BI [Sl JSI ee eee Tr oo o o ee ee er 2 ie n e e ro
Sl |6 19 Jar eie e Oe OS |B Or | 018 168 0 18 16 16 e rn A Ss. ett 8 S
t.Je PSR e g ge e ee ee San a ie t 40 18 eo L 18 18 1016S Y
© |S Je pee e egia Teo 10D eee 104016 18 16 |e ai 19 1s 19) SN
O 10 1810 os “Sy poms -9 10 10 14 1 018 er a T eo 0 181 et 10.16 S E | o S
sesejososs|jsse fe ES sesjsssosjesos .. esfessoejooso pee Nae Sn A kS Ge RA A 0 0 Z ss.es|essjossjoeso 9 [=e]
ar far elaeodes tsr e leebeks dzz ls le zlizirls
6 -3 olele Itio lo 9 laisti o ee ee ee E eee ioe r na a a
8 TSI 19 1S | OIE ee re ee Te et erie ter. OF Tt et ier 2 >
$ 19 [bg] OL] 8 | SE] Ot 10-78 1 8t 16. [eee ee ati et ie ie ie ie et Stier ag
e a T Jao he oia Oog Ol 6 e eee A ee e g kee ee an
o jo TVS g8 FLIET T e eee 9 a ik 8 ie ee ee Te
818 fT LO 0 47 9 RTT ztia orior een e i E F 18 ee ee X
6 JTT e 10.) 8-1 ete 1018 Z ioo ee 16 eek 018 8 a O10 ee
O [Ft UG) oO re IZ 90 0 | Or or olore o 6 18S 10 108: 818. Oiri a N
e) 8 es © é 2 o A © ò Pidin s irio IS 1 P16) © é © 1/6} 16

yurd pou yurd pou yud pou yud pe yud ped yurd pou yud pew oun

b ayog SITIO 9 IPOT p Io 2 940g q ƏmMog v Əyog

ALIAILONGOUG AALLVINY WAHT, SMOHS TIVL IHL 'WIVNIE | uae
CIZA ANIA INO NV QAY ANO GANIVLNOD HOIHM AO HOVA STILLOQ 4 NI aaug ‘ANIA 4 ANV aay 4 ‘8,5 PI HO GuOOmYy
A Wav yi

No. 578] DROSOPHILA AMPELOPHILA 101

It is evident that neither dryness nor moisture has any
effect in rendering more favorable the conditions needed
for the emergence of either the red or the pink variety.
The results show that under unfavorable conditions, large
numbers of larve and pupe fail to develop, since the total
yield of each and every bottle is far below the normal
output; and those which do emerge are but chance sur-
vivals.

It was suggested that the reason the pink flies fell be-
hind the expected ratio, was the fact that the mutant
was weaker than the wild stock and therefore less likely
to come through the. larval and pupal stages. If this
were the case, they should always fall behind. In many
cases, however, they actually exceeded the expectation.
Furthermore, they always seemed to be just as vigorous,
and to live as long as the wild fly.

The hypothesis which was formed at this stage, and
which determined to a large extent the experiments which
followed was, that another factor not related to eye-color
was at work. Such a factor, if one is assumed to be pres-
ent, by its independent action might be responsible for
the disturbance in ratio. It might, moreover, be present
in the wild stock which originally gave rise to the pink,
since the wild fly is similarly, though less frequently,
affected. It is with the search for such factors that the
subsequent experiments will chiefly deal.

Before presenting the data, it will be well to point out
some of the possible sources of error which were to a
great extent eliminated.

I. The method usually employed in these experiments
is as follows: The flies, which are to be cross-bred, are
taken out of the culture bottles as soon as they hatch and
before they have time to mate. They are then put into
a clean, sterilized bottle in mass cultures of about five or
more pairs. There they remain till their offspring (F,)
are ready to emerge: 9-10 days in summer and 11-12 or
13 days in winter when the temperature is low. The F,
flies are placed in fresh bottles for a similar length of

102 THE AMERICAN NATURALIST [Vou. XLIX

time, and then removed. During the succeeding ten days
the F, are counted each day as they hatch. The bottle is
then discarded for fear of overlapping of generations;
for, the F, might mate and deposit eggs before removal.

It has, however, been observed that each time a bottle
is discarded a considerable number of pupe and even
larve remain behind. This being the case, it is possible
that the ratio we get does not always represent what
actually happens. In order to count the total output, it
was decided to transfer the flies to a second bottle on the
fifth day. All eggs deposited, during the five days that
the parent remained there, would thus have at least fifteen
days to develop. ‘It was hoped that, in this way, a more
representative ratio would be obtained.

II. It was shown (Tables III and IV) that large num-
bers of larvee fail to develop when a bottle becomes too
‘“dry’’ or too ‘‘wet.’’? Considerable care was taken to
avoid either of these conditions. If a bottle showed a
tendency to dry up, fresh food was immediately added;
when it was too wet, the moisture was absorbed by filter
paper.

III. The yield of a mass-culture bottle is always rela-
tively small as compared with that of the same number of
flies mated in pairs. This would indicate probably greater
mortality due to overcrowding. For this reason only
pairs were used in the later experiments.

In the first of these experiments pure stocks, both pink
and red, were used; for it was believed that if differences
existed other than the red-pink distribution, between the
two varieties, they would be more emphasized if hybridi-
zation had not been effected. The chief purpose, how-
ever, was to become familiar with the modes of behavior
of the races. The experiment follows:

A number of flies, both pink and red, were isolated
within one to six hours after hatching and the sexes kept
apart for 3-4 days, after which time they were mated,
red to red, pink to pink. Immediately after mating, which
took place within five minutes to two hours, the males

No. 578] DROSOPHILA AMPELOPHILA 103

were removed. One red and one pink of these females
were put into each of seven bottles. In this way the same
environmental conditions were secured for the eggs of
both. Five days later they were allowed to mate again
(not by the same males) and placed in a second set of
bottles. The same two females that were together in the
first set were also together in the second. There they
remained five more days. Counts of the flies that hatched
were made from day to day, and the bottles were emptied
as long as they continued to yield. The results are given
in Table V.

The records of these flies show several interesting and
suggestive facts. It will be noted, in the first place, that
the length of larval life varies through wide limits. Eacli
bottle contained eggs which were deposited during a
period of no more than five days. The hatching periods,

however, extended through eleven days in the first set of

bottles (April 25-May 6) and twelve days in the second
set (April 30-May 12). The flies which emerged first
consumed but ten days for development; those which
emerged last took at least sixteen days. This phenom-
enon was more marked where the number produced was
larger, suggesting that crowding may retard the devel-
opment of some individuals.

Of equal interest is the fact that the pink flies invariably
began to hatch from 24 to 48 hours later than the red. This
was true in the second set of bottles as well as in the first,
which proves that it was not due to late maturity of the
parents, for, at the time of transfer, they were in the
midst of their productive period.

Another point of interest is to be found in the fact that
the pink stock was, on the whole, less fertile than the red.
In the two bottles n and ‘g’ where the productivity of the
two was about equal the red, like the pink, were also low-
producers. This is significant, and will be referred to
later.

In order to test the above-mentioned facts, the follow-
ing periment: was performed. F, hybrids were mated

104 THE AMERICAN NATURALIST [Vou. XLIX

in pairs, and transferred during a period of twenty days
(May 16—June 4) 13 times, remaining in each bottle from
one to two days. Care was taken to count every fly of the
F, that hatched. The result follows:

ABLE VI
RECORD OF F, FLIES WHICH HATCHED FROM EGGS DEPOSITED DURING A
PERIOD OF 20 Days; DURING WHICH TIME THE PARENTS WERE
TRANSFERRED THIRTEEN TIMES

Pink 9 X Red g — in F2 Red ọ X Pink ð > in F?
ree Red Pink Proportion ae Red Pink *| 5. ortion
9 S 9 g Red; Pin 9 E g 3 Red : Pin
A. 235 | 209 95 2.3 :1 a 134 | 118 | 42 | 37 3.231
Bosch 220 | 213 | 84 | 75 WL A | b 105 | 108 | 67 | 64 16:1
Cas 162 | 168 | 65 | 58 rA aes | Cee 189 | 153 | 52 | 53 3.2:1
D....|202|119| 14 | 19 O72 5 ds 58| 11 8 6.5 :1

Here we have a group in which the pink ran sometimes
relatively ahead of the red. But the other extreme is also:
represented in pairs D and d. The numbers obtained are
in each case large enough to be significant.

This experiment was repeated on a larger scale in the
fall of 1912. Ten pairs were used for. each of these
crosses and they were continually transferred as long
as they lived. The records follow:

TABLE VII
F, RECORD FROM RED-EYED FLIES CROSSED TO PINK-EYED FLIES MATED IN
PAIRS. SHOWING THE TOTAL OUTPUT OF EACH F, PAIR
DURING ITS LIFETIME
PINK 9 X RED f > IN F,

DSe a reier e PoR Total | Total mye Lh tag Proportion
1 u
Pair | “Lived® |Transferred g | S | g| Red | Pink | Each Pair Red: Pink
A..| 45 27 |604578| 89| 81| 1,282 | 170 1,452 7.5:1
Bist 8# 15 169145) 30| 41| 314| 71 385 44:1
Cin ME 22 |320/305/116| 94| 625 | 210 | - 835 3.0 :1
Di Of 32 |658 681/203 221| 1,339 | 424 1,763 3.2:1
E.. 16 {218 218) 79 436 | 152 2.9 :1
ri p 28 |476435/153|144| 911 | 297 1,208 3.0 :1
a. m 17 [259248 83) 89] 507 | 172 679 |29+:1
E T 22 392,396 152/125] 788| 277 f 28+: 1
I G 32 [864805 258'281| 1,669 | 539 2,208 3.1:1
Juj 8 19 |357 358130124) 715 | 254 969 2.831

5 Lived only eight days. i i
6 The length of time a fly liveđ should not be taken as a criterion for
measuring its vigor. In most cases death is accidental.

No. 578] DROSOPHILA AMPELOPHILA 105

TABLE VIII
F, RECORD FROM RED-EYED FLIES CROSSED TO PINK-EYED FLIES MATED IN
PAIRS. SHOWING THE TOTAL OUTPUT oF EACH F, PAR
DURING ITS LIFETIME
RED 9 X PINK fd > IN F,

No.of | Red | |

Py A FHE | Total | Total Total No. Proportion
E rR Days Each Mee DR e| Be | eja] Hed | Pink | Produced (Red; Pink
ps | 28 20 30 2041 p 660 | 236 896 | 28%)
b. Si 13 11 1631 70, 58| 49| 333| 107 440 S45)
c. 7 5 74 90; 22| 28) 164 50 214 | 3a: i
a... 14 12 213/195) 58} 57| 408) 115 523 3.5+:1
é: 22 15 101 106) 33| 26; 207 59 266 g0: i
ie 21 16 215266 95| 78) 481| 173 654 2.8: 1
g. 37 24 erami 13,139 646 | 252 898 260:1
X 45 2 (433 444/179 169| $877 | 348 1,225 | 2.2+:1
i. 43 26 1546 533187 185) 1,079 | 372 1,451 2.0.21
L > 30 21 298 306 90| 83, 604 173 777 | Bs Fs ae |

A comparison of Tables VI, VII and VIII suggests the
possible presence of high and low pink-producing

‘*strains’’ in these stocks. To test this, some of the off-
spring of pair A (Table VII) in which the ratio was 7.5
of red to 1 of pink, were inbred for the F, in order to see
if the same ratio would persist. As there were among the
red both homozygous and heterozygous forms, they were
each mated to their pink sisters or brothers. This com-
bination would give with the former all red (since red is
. dominant) and with the latter a ratio of 1:1.

Pair F (Table VII) in which the ratio was ideal, 3:1,
was chosen for the control, and treated in like manner.

In this as in the preceding experiment the flies were
mated in pairs and transferred to fresh bottles every
second or third day. A peculiar thing happened. Out of
25 pairs taken from ‘‘A’”’ only two gave offspring; the
remaining 23 pairs were apparently sterile. It could not
have been due to bad banana or any other unfavorable
condition, for the flies had already been transferred five
times and no pupæ were found in any of the other bottles.
Furthermore, the 16 pairs of the control which ran paral-
lel to them, and were fed with the same food, did well.

106 THE AMERICAN NATURALIST [Vow XLIX

To find out whether these flies were actually sterile,
each of the 14 remaining pairs—9 having meanwhile
died—were separated and every individual mated to wild
red-eyed stock. The sterility of the pink flies, both male
and female, was found to be absolute, while all red of
both sexes were fertile.

As an additional test, some of the offspring of the last
cross were inbred en masse in order to extract the pink
flies which they would produce, since some of them were
heterozygous for eye color.

A small number of pink flies were obtained and mated
to their red brother and sisters: each pink female was
put in a bottle with 3 or 4 red males, and each pink male
with 3 or 4 red females. Out of 19 individuals thus tested,
only three were found to be fertile; the remaining 16 were
sterile.

These facts seem to indicate that some factor or group
of factors which make for sterility were present in the
‘*pink’’-containing gamete. The results are the more
significant since the hybrid fly, in which this condition
prevailed, produced a very low pink ratio. Of the control
in which 10 pairs were found to be heterozygous for eye-
color with an expectation of 1:1, the following results
were obtained:

TABLE IX
RECORD OF THE OFFSPRING OF F, FLIES OF A Cross oF PINK 9 By RED ĝ IN
WHICH A Ratio Was 3:1, as EXPECTED. (See pair F of Table VIII)
Red Heterozygous 2 by (Brother) Pink g

Red Pink Total No. i

Pair [Faek Liv Rod | Pink | Produced | Ror Pine
9 g 9 a ch

1 50 352 348 236 244 700 480 1,180 Leet

2 27 150 130 170 161 280 231 511 E21
3 18 51 85 59 50 136 109 245 12+:1

4 29 251 245 240 258 496 498 994 10:3

5 10 147 132 133 112 279 245 524 gS Fa |

The reciprocal cross gave the following:

No. 578] DROSOPHILA AMPELOPHILA 107

Pink Heterozygous Q by (Brother) Red ¢

| Red Pink |
._ |No. of Days! : Total | Total Pro —
Pair g | Produce
Each Lived Pe o 9 3 ed Pink by Eac Re

I 20 82 8 69 9 171 128 299 13+:1

| 123 124 94 107 247 201 448 L2: i
III 32 | 223 240 213 247 463 460 920 sR i aah S
IV 29 | 215 176 184 163 391 347 738 | Fs Gee |
y 40 216 178 101 136 394 237 631 | i he aoe |

It will be seen that all the red flies of the last cross were
heterozygous, and should give, on further inbreeding, a
ratio of 3:1. Four pairs out of five (one being sterile),
taken from Pair III of Table IX, gave the following:

TABLE X
RED 9 X Rep ¢ or Pair III, TABLE IX

No. of Red Pink Total No,
Pair Days Each ——— Total | Total | produced | Pro rtion

Dea g E 9 gi | Rel i, Pink | FIMA | : Pink
a 40 312 | 348 43 45 | 660 88 748 Cf BE
b 23 282 | 316 80 103 | 598 183 781 8.321
c 40 194 | 214 60 70 | 408 130 538 | 3.1:1
d 14 175 | 163 47 59 | 338 | 106 444 | 32:1

Pair ‘‘a’’ above gave the same result as pair “A”
(Table VII)—the ratio in each case being 7.5: 1. The off-
spring of the latter were found to contain a high percent-
age of sterile pink flies, owing to which the attempt then
made to test that ratio failed. It was therefore decided
to repeat the same experiment with the offspring of this
‘a’ pair. As in the former case, the heterozygous red
flies were picked out by crossing them to their pink
brothers and sisters. The expectation was again 1:1.
The records follow.

TABLE XI
REcorpD oF E1eut Pas, HETEROZYGOUS RED X TO ngs TAKEN FROM THE

OFFSPRING OF Pair ‘‘a,’’ TABLE
Pink 9 x Heterozygous Red $

shag — Pink Total | Total San No. | Proportion
Days Each roduced Poe
e iivet | ol @ | 6 Lo | Bet| Fmt |, by Meck | Bod t Fisk
r 27 179 176 79 88 | 355 | 167 522 21+:1
A. 12 155 153 83 66 | 308 | 149 457 204+ :1
g... 28 163 | 205 | 126 | 159 | 368 | 285 653 1.3 ii
Di s 243 | 222 | 206 | 219 | 465 | 425 | 890 | r1:1

108 THE AMERICAN NATURALIST [Vou. XLIX
The reciprocal cross gave:

Red Heterozygous 2 X Pink g

| i |
No, of Red Pink | Total No .|
n Total | Total | Proportion
eine g FOO eae ` FEF a | Red | Pink | b tucht | Red: Pink
a 28 282 236 210 260 518 470 988 | be Nees |
b 28 296 290 5 5 586 10 596 | 59.0: 1
e 12 144 158 136 152 302 288 590 1.0+:ł
d 25 Ta 113 63 98 185 161 346 }11i+:1

The pair with which we started (F, Table VII) gave
the ideal 3:1 ratio; but in each of the three generations
which were bred from its offspring (Tables IX, X, XI)
there appeared again the same fluctuations which were
observed in the preceding experiments, and with even
more striking emphasis. Among the offspring of the
same pair are found some that give a 3:1 ratio and some
that give a 7.5:1; in a second pair we have some giving
1:1, and one giving 59:1. The latter especially suggests:
the presence of a factor that actually inhibits the devel-
opment of the pink flies, and, moreover, that it is being
segregated in a mixed stock.

If the presence of such factor is assumed, we should be
able by inbreeding to select stocks in which it is present
and in which it is absent. For this reason the experi-
ment recorded above (Table V) was here repeated with
some modifications. It will be rémembered that in the
former, a fertilized red female and a fertilized pink male
were placed in each bottle and their offspring, the F,,
counted. In the following experiment, in order to secure
segregation, the F, were counted. Four virgin pairs, two
red and two pink, were taken out of the culture bottles and
mated separately. . Of the F, of each of the four, six pairs
were taken out, 24 pairs in all, and bred for the F,
Unlike the first experiment, the males in this case were
allowed to remain with the females throughout the ex-
periment. This insured sufficient sperms for the eggs.
Every second day the food was removed, together with the

No. 578] DROSOPHILA AMPELOPHILA 109

eggs deposited upon it, and fresh banana supplied. Each
two batches of eggs—one deposited by a red female, one
by a pink female—were placed together in one bottle so
that they might develop side by side and under the same
environmental conditions. The result of this experiment
is shown in Table XII.

TABLE XII
F, oF Two PINK Pairs A, B, AND Two RED PAIRS a, b, SHOWING SEGREGA-
TION OF PRODUCTIVITY. Eces or Al-al, BI-bI, ETC., WERE
DEVELOPED IN THE SAME BOTTLE

Pink Red
No, of No. of |
Days Total | Ave. F; Days | | pe | E
gong ym 7 s E Das er ieee’ ks duced | Day
AI | 44 |450 a0 896 | 20.40 aI | 20 233219 452 | 22.60
~ |All 169/224; 393 | 16.36 | 5 H o o S MO
„|I| 28 | 61) 67| 128 | 457| |a] 25 |63| 64) 127 sao
S| AIV| 30 |221|233 15.13 a aIV | 31 (272232) 504 | 1
MAV | 36 /331\345| 676 | 18.77 aV | 30 (179/171 350 “1166
AVI} 21 | 37) 33) 70 | 3.33 aVI | 22 152/138 390 | 17.73
BI | 29 {199/217 416 | 14.34 bI | 44 122104 226 | 5.10
m |B | 24 93| 1 7.75 |_,| bI | 28 |103| 77) 180
Bl} 35 | 31 a7 78 | 28g b TII | 302256 558 | 12.70
2@|BIV| 30 |15|13| 28 | 093| &|bIV | 35 | 22/34 56
MTBN | 283 |11 10 21 | 0.75 bV | 302298 595 1920
BVI! 18 8 1.00 BVI | 44 154139 293
—— |

Segregation with respect to productivity is here evi-
dent. Whether the low fertility? seen in so large a pro-
portion of these flies was due to an actually low egg-
production, or whether it was due to something which
prohibited development or to some defect in the germ
cell owing to which fertilization could not be effected, is
not known. That one of the latter possibilities is likely
to be realized here can be inferred from the work of Dr.
R. R. Hyde in this laboratory. He counted the eggs of
hundreds of individuals, and compared them with the
number of flies which emerged from them. According to

T The term ‘‘fertility’’ is used here, as defined by Hyde, to indicate the

number of eggs that complete development and give rise to ma
(See Hyde, Jour. Exp. Zcol., August, 1914, p. 185.)

110 THE AMERICAN NATURALIST [Vou. XLIX

his observations, only about 75 per cent. of the eggs of the
wild fly ever reach maturity, and in some of the muta-
tions no more than 25 per cent. of the eggs develop.

Another point of interest brought out in the last experi-
ment is the fact that the wild, red-eyed fly behaves in
exactly the same manner as the mutant pink fly. This
may be the reason for the observed shifting of the ratio
sometimes in favor of the one variety, sometimes in favor
of the other. It shows furthermore that it was not the
pink as such that caused the disturbance. The red also
might be similarly disturbed and perhaps by the same
agent or by another agent that affected the productivity
in the same way.

If the abnormally low number produced by some of the
pairs of Table XII be due to the inability of a large num-
ber of their eggs to develop, and if we assume this char-
acter to be transmissible, it must reappear in the F, of a
cross in which one of the parents possessed this factor,
i. e., a large number of individuals, one quarter of the
output, should fail to develop. This would be in accord-
ance with Mendelian principles. A number of crosses
were therefore made in various combinations with the
individuals taken from Table XII. The results follow:

TABLE XIII
F, or 16 PAIRS oF A Cross RED BY PINK IN WHICH THE PARENTS CAME
FROM AT, aI, TABLE XII, THE AVERAGE DAILY PRODUCTIVITY OF
WHIcH WAS 22 AND 20, RESPECTIVELY
A. Pink Q (Productivity 20 Per Day) X Red g (Prod. 22 Per Day)

No. of Days Red Pink Total | Total | Total No. | proportion
ch à

Pair = ‘ P ý r Red | Pink p Red : Pink
1 17 301 110 04 595 214 2.00e2 1
2 32 4 457 60 | 147 941 1,248 | 301:1
3 32 570 181 189 | 1,074 370 1,444 2.90: 1
4 32 559 548 20. „l 405 1,512 PATA a i
5 8 125 126 251 5 326 8,89: 1
6 18 261 243 98 89 504 187 691 27071
7 21 391 | 449 | 155 | 142 | 840| 297| 1,137 | 2.86:1
8 32 673 700 203 | 236 | 1,373 439 1,812 3.10: 1
Total number produced by 8 pairs....... 6,685 | 2,294 | 8,979

Average proportion, 2.91 : 1

No. 578] DROSOPHILA AMPELOPHILA 111

B. Red Q (Productivity 22 Per Day) X Pink g (Prod. 20 Per Day)

Red | Pink

| | |
i No.of. odes vines 2 Total | Total | Total No. | proportion
Pair elspamey m | z | Eor Red | Pink | o oe | Red: Pink
I 30 415 | 395 | 135 | 127 | 810] 252| 1,062 | 3.21:1
II 14 132 121 41 73 253 | 114 207 4 2.22: 1
III 24 317 337 97 94 654 191 | 845 | 3.42:1
Vv 30 484 | 378 | 119 20 | 862) 239) 1,101 | 3.60:1
24 392 418 96 140 810 | 2 1,046 | 3.42:1
y 30 656 592 198 20 248 | | 646 | 3.14:1
VII 30 566 647 182 09 | 1,213 391 | 1,504 | 2.10: I
VIII 24 438 451 132 153 889 1286] 1,174 | 3.10:1_
Total produced by 8 pairs................. | 6,739 | 2,106 | 8,845 |

Average proportion, 3.24 : 1

If the proportion of red to pink, realized in the F,,
depends upon the relative fertility of the two parents
which form the cross, we should get in this case, where
the parents were supposedly equally fertile, the ideal 3:1
ratio. The records, however, show considerable fluctua-
tions. Nevertheless, these results are perfectly in accord
with our hypothesis. Looking back to Table XII, which
furnished the parents of this cross, the explanation is
obvious. The averages per day for AI—AVI were 20, 16,
4, 15, 18 and 3, respectively. Similarly, al-aVI gave 22,
0.7, 5, 10, 16, 11 and 17, respectively. It is therefore
reasonable to assume that among the offspring of AI
(productivity 20) and al (productivity 22) individuals
should be found which would repeat the series. Fluctua-
tion is, therefore, to be expected. The average of many
such pairs, however, should be 3:1. The proportion ob-
tained was 2.91:1 in one case; 3.24:1 in the other, or a
general average of 3.08: 1.

It should also be noted here that in this as well as in
the subsequent experiments, wherever eight pairs are
recorded, they are not the offspring of one, but of two
distinct crossings of one pair each which were made at
the same time; that pairs 1-4, 5-8; I-IV, V-VIII, respec-
tively, were brothers and sisters. More than one line is
thus represented in each case. With these facts in mind,
we may pass on to the remaining experiments,

112 THE AMERICAN NATURALIST [Vow XLIX

TABLE XIV
F, of 12 PAIRS OF a Cross RED BY PINK IN WHICH THE PARENTS WERE AI
AND bII (TABLE XII) THE AVERAGE DAILY PRODUCTIVITY OF
WHIcH WAS 20 AND 6, RESPECTIVELY
A, Pink 2 (Productwity 20 Per Day) X Red g (Productivity 6)

No. of Da Red Pink Total No,
Pair | Each Was. Total | Total | Produeed | Proportion
ie ee ci eae
1 32 355 331 117 95 686 212 898 Sze f1
2 28 340 373 126 132 713 258 971 2.761
3 19 330 288 120 124 618 244 862 2.52: 1
4 32 544 562 188 188 | 1,106 376 1,482 2.94 :1
5 23 574 578 203 214° | 1,152 417 1,569 Trd o e D
6 14 196 194 66 59 390 125 515 3.12: 1
7 32 648 603 222 246 | 1,251 468 1,719 2.6754
8 32 429 374 141 144 803 285 1,088 ASL Sy
Total number produced by 8 pairs.......-. | 6,719 | 2,385 | 9,104
Average proportion, 2.84 : 1
B. Red 2 (Productivity 6 Per Day) X Pink g (Productivity 20)
' |No. of Days ya Fisk | Total | Total | Total No. | proportion
Pair | Each W | : Produced N
sir yg we d as 9 3 9 | Red | Pink fey wie Red ; Pink
I 28 488 482 176 IF 970 | 353 1,323 2.74: 1
II 6 83 78 21 30 161 51 212 3.15 <1
HE 28 463 465 164 153, 928 | 317 1,245 2.93: 1
IV 21 375 383 114 116 758 | 230 988 8.20 ri
Total number produced by 4 years....... 2,817 | 951 3,768

Average proportion, 2.96 : 1

In most of these pairs the pink slightly exceeded the
3:1 expectation. In the few in which they fell behind,
the red (if we assume fertility to be the cause) might have
been of a higher fertility than the pink, as has been
explained. As a group, however, they give a proportion
somewhat below 3:1. _

In the next cross, the red fly was the more fertile. The
results are given in the following table:

No. 578] DROSOPHILA AMPELOPHILA 113

TABLE XV
F, of A Cross RED BY PINK IN WHICH THE RARENTS WERE BIV AND al
(TABLE XII) THE AVERAGE DAILY PRODUCTIVITY OF WHICH
AND 22, RESPECTIVELY
A. Pink 9 (Productivity 1 Per Day) X Red g (Productivity 22)

of Red | Pink Total | Total
Pair | D. 'E h ; a ota zo a Si Proportion
Days OB vod 9 | P | 9 p Red | Pink ts 0 iue È : Pink
1 22 398 | 405 | 120 | 105 225 | 1,028 | 3.52 :1
2 28 512 | 477 | 107 | 125 | 989| 282] 1,221 | 4.26:1
3 22 420 | 389 | 104 | 135 | 809| 239| 1,048 | 3.34:1
4 18 340 | 06 192 3.33 :1
5 19 420 | 428 | 129 | 134 | 852] 262] 1,115 | 3.24:1
6 19 372 | 396 | 100 04 | 768 72 | 3.76:1
7 17 198 | 220 | 418| 141| 559 | 3.00:1
8 174 | 212 62 62 | 386 510 | 3.10:1
Total number produced by 8 pairs......... 5,665 | 1,620 | 7,285
Average proportion, 3.50 : 1
B. Red Q (Productivity 22 Per Day) X Pink 3 (Productivity 1)
No. of Days Red Pink | Total | Total | Total No. | proportion
Pair eo Was Produced
red 9 a g e Red Pink by Eact : Pink
I 31 264 | 262 | 101 | 87 | 526| 188| 714 | 2.80:1
II 27 353 | 350 703| 165 4.26: 1
Ill 31 430 | 403 112 1,042 1
IV 31 420 | 111 | 128 9| 1,086 | 3.54:1
vV 28 471 | 122 | 147 | 957| 269 3.60 : 1
VI 573 | 568 39 | 154 | 1,1 1,434 | 3.90:1
VII 28 116 | 127 |1,010| 243| 1,253 | 4.15:1
VII 24 | 502 | 501 | 134`| 140 |1,003| 274| 1,277
Total produced by 8 paits.........-,., 20%- 7,020 | 1,880 8,900

Average proportion, 3.73 : 1

Of the 16 pairs of this cross only one gave less than 3: 1.
the remaining 15, the proportion was, in each case,
considerably higher than 3:1. It will be noted that of
all 16 pairs that one was the least fertile. This would
indicate, on the hypothesis suggested, that the gamete
containing the ‘‘red’’ factor did not have relatively as
high a potential of tortility as did the parent which pro-
duced it.
A comparison of Tables XIV and XV shows that we
have two distinct groups: one in which the extracted pink
exceed the expectation, and one in which they fall behind

114 THE AMERICAN NATURALIST [VoL. XLIX

the expectation. Yet the method employed in each case
was the same; the history of each is the same. The only
difference is to be found in the fact that in the one case
the pink came from a more fertile parent; in the other,
the red.

The offspring of pairs ‘‘7’’ (Table XIV) and ‘‘2’’
(Table XV) in which the ratios were 2.67: 1 and 4.26: 1,
respectively, were inbred for the F,. Fifteen pairs were
taken from each, but as there were among the red both
homozygous and heterozygous flies, only eight gave pink
in each case. The results follow:

TABLE XVI

RECORD OF 8 PAIRS HETEROZYGOUS RED-EYED F, OF PAIR ‘‘7’’? (TABLE XIV)
IN WHICH THE Ratio Was 2.67: 1

Red Pink Total No.

No. of Total | Total | produced b Proportion

Pair ae a 2 z | Red | Pink Each | Red: Pink
1 14 270 279 80 78 158 707 3.47 :1
2 14 159 167 58 42 326 1 426 3.26: 1
3 14 7 2 1 12 99g: 1
4 14 135 143 46 49 278 95 373 2.93 : 1
5 14 1 157 49 50 340 99 439 3.23: 1
6 14 172 54 45 42 326 87 413 3.74 :1
7 1 142 137 43 49 279 92 371 3.00: 1
8 14 12 16 103 28 131 2:67 : 1

TABLE XVII

RECORD OF 8 PAIRS HETEROZYGOUS RED-EYED F, OF PAIR ‘‘2’’ (TABLE XV)
IN WHICH THE RATIO OF RED TO PINK Was 4.26: 1

Red Pink Total No.
pais | Be Teta | Tote | rete | apron
Q g Q a

1 14 295 308 100 603 197 800 3.06 :1
2 14 392 326 97 105 718 2 920 55:1
3 14 281 95 9 181 750 3.14: 1
4 14 110 32 214 279 3.31: 1
5 14 167 170 70 59 337 129 4 1
6 14 314 277 82 86 591 168 759 3.52: 1
ri 14 160 32 41 329 7 402 50

8 14 133 146 49 Al 90 369 3.10: 1

These results are inital, in that they show that the
original ratios, which their parents gave, were lost.

No. 578] DROSOPHILA AMPELOPHILA 115

The fact that it has been found possible by proper
manipulation to get a group in which the ratio fluctuated
in one direction only, even if it was not as marked as was
hoped it would be, indicates that the disturbance is due to
an internal, and not to an external, cause. This was
further emphasized by the distinct tendency for segrega-
tion, as was to be expected if there were some hetero-
zygous individuals. It was also suggested in another
way. No matter how short-lived or how long-lived a pair
was; whether it was transferred once, twice, or even
twenty times, the ratio of red to pink did not vary
throughout its life when the yields of the several bottles
were compared with one another.

TABLE XVIII TABLE XIX
F, or Four Pars (A, B, C AND F, or Pams A, B, C anD D or
D) Rep? X PINK ĝ IN WHICH TABLE XVIII

THE MALES WERE CROSSED Expectation 3: 1

EACH TO SEVERAL oF His
OWN DAUGHTERS
Expectation 1:1

Pi F,of No.of | Total | Total | Proportio
> $ 7 Tea Pak ko oon Phir Pair Red Pink Red: Pink
1 45 | 2.15:1 8.) G 3ni
2 105 71 | 146:1 A 2 158 | 55 | 3.00:1
ax 3 231 161 | 1.43:1 96 21 | 4.60:1
4 171 | 123 40:1
5 Sit ti 1.10 :1
1 110 80 | 1.38:1 1 140 | 43 | 3.27:1
2 263 | 251 | 1.00:1 B 2 215 | 72 | 3.00:1
BX 3 242 | 198 | 128:1 3 | 1% | 7 | 109:1
4 183 | 143 | 1.28:1
5 231 | i89 | 122:1
6 162 |: 102 | 1.62:1
1 %1 | 113 78:1 1 50 | 21 2.38: 1
2 257 | 116 | 2.22:1 c 2 125 | 43 | 290:1
“Cx 3 255 1. 1 3 122 | 49 | 2.49:1
4 206 | 153 | 1. 1 4 07 | 288:1
5 943 | 107: | 1.28:1
6 125 1.27:1
5 —
179 1.30 :1 26 | 6 | 433:1
DX 2 | 997: | 191 |-220:1 D 2 112 | 34 p ago:
107 44: 3 194 | 74 | 2.62:1

116 THE AMERICAN NATURALIST [Vou. XLIX

The proportion of red to pink was found to bear a
direct relation to the relative ‘‘fertility’’ of the parents
which produced the hybrid. This suggests a causal rela-
tion between the two.

In dealing with ‘‘fertility’’ the difficulty that one en-
counters is, that the offspring of any pair may, with
respect to this character, differ from either parent, and
also differ amongst themselves, forming a graded series
running from the most to the least fertile. An individual
taken from such a population is an indefinite quantity and
will often defeat the purpose of the experiment. In order
to simplify this as far as possible, the following experi-
ment was planned:

Four red-eyed, virgin females were each mated to a
pink male. Each male was again crossed to several of his
own daughters. The records are given in Table XVIII.
As a control a number of F, pairs were bred in each case.
The records are given in Table XIX.

A graphic representation of all pairs recorded in
Tables VI-XIX, except for the several very unusual
ratios, is given in Figs. 1 and 2.

—

‘ l
UES |
3.0:1
Ee,
j
| 35:1

4.6:1

ie

n
oc hind B
3 3
oi +

Fig.1

m Leil

Fig. 1 contains 99 pairs in each of which the expected
ratio was 3:1, with a total population of 82,607. They
are distributed as follows:

No. 578] DROSOPHILA AMPELOPHILA 117

No. of Ratios of
Pairs Red : Pink

: eee
Rn Se ee eee
ge
$]
<4
@
w
=i
pæd

Total..

a]

1.0:1
L

Fig.2
Fig. 2 contains 37 pairs of back-crosses; expected ratio
1:1; total population 17,008. They are distributed as
follows: :

118 THE AMERICAN NATURALIST [Vou. XLIX

No. of Ratios of
Pairs Red : Pink®
LOSI
gave 1.1:
gave 1.2:
gave 1.3:
gave 1.4:
gave 1.5:

A
E

=

[e]

D H Om wM a OON O
09
($)
4
©
n
D
pi pa pn d ped pad, had pad d pd. l pat

Total. .37

To these should be added:

1 pair which gave 9.7: 1, Table V
1 pair which gave 6.5: 1, Table VI

1 pair which gave 7.5: 1, Table VII
1 pair which gave 7.5: 1, Table X
1 pair which gave 59.0: 1, Table XI

Except for the several detached pairs at the extreme
limits, Fig. 1 shows a normal curve. A disturbance of
0.5 in either direction (less than 10 per cent.) is quite
within the limits of experimental accuracy. The larger
disturbances, ratios of 6 or 7:1, and also the first results
reported by Morgan (711 and ’12) are yet to be explained.
These are too large to be attributed to experimental error.

The data presented in the foregoing pages show that
there has been a marked improvement in the ratio of pink
to red since 1911. In one ease only (1913) was the dis-
turbance greater than those of Morgan (59:1). The
remaining very marked disturbances were between 6 and
10:1. And these appeared so infrequently that in mass-
cultures their presence would hardly have been felt.

A corresponding improvement has also been observed
in the fertility of the pink-eyed race between 1912 and
1913. This is seen on comparing Tables V and XII. In
the first, the fertility of the pink was much lower than
that of the red; in the second (about one year later), it
was as high.

Hyde (’14) showed that in some races of Drosophila

8 None gave a ratio of less than one red to one pink.

No. 578] DROSOPHILA AMPELOPHILA 119

ampelophila, the number of eggs failing to reach matu-
rity is between 25 per cent. and 75 per cent. of the total
output; and concluded that this peculiarity probably be-
haves as a Mendelian recessive factor. More recently,
Morgan (’14) describes recessive lethal factors in Droso-
phila, which he defines, ‘‘as any factor that brings about
the death of the individual in which it occurs, provided
that its effect is not counteracted by the action of its
normal allelomorph.’’

In the light of this evidence, the following conclusions
suggest themselves: ie

1. The original pink-eye mutant was heterozygous for
some non-sex-linked factor which, in the homozygous state,
acts like Morgan’s lethal. This factor was, in the course
of time, to a large extent eliminated, as is to be expected
if the individuals homozygous for it are more likely to
die. The chance of such homozygous forms appearing
again, has thereby been much reduced. This is borne
out by, and also explains, the improvement in the pink
race.

2. A similar recessive, though not necessarily the same
factor, might also be present in some individuals of the
wild, red-eyed stock. Hyde’s work mentioned above
gives weight to this assumption—which is not at all an
unreasonable assumption in a species as unstable as this,
judging by the vast number of mutations reported. For
this reason, the red sometimes fall behind the expected
ratio. 7

3. The mode of action of these lethals shows that they
are linked to the ‘‘pink’’ factor or to its normal ‘‘red’’
allelomorph. This will be clear from the following
analysis:

Of the flies recorded in Fig. 2, one parent was RP (with
gametes R and P); the other was PP (with gametes P
and P). The zygotes resulting from these gametes almost
invariably give fewer PP’s than RP’s. In other words,
the homozygous forms run behind the heterozygous forms.
The relation between these two classes may also be sup-
posed to hold in the F, cross (Fig. 1). Here, however,
the reds (RR and RP) run relatively less often ahead of

a

120 THE AMERICAN NATURALIST [Vow. XLIX

PP. This must be due to a deficiency in the homozygous
RR flies. In other words, the results taken all together
(Figs. 1 and 2) show that the disturbance is brought
about by factors (in the third chromosome) which in the
homozygous state act as lethals or perhaps as semi-
lethals. Random introduction of one or two or no lethals
may be assumed, as follows:

(A) If the lethal is introduced by the ‘‘pink-bearing”’
tna the homozygous pink will be depressed in
the

on ‘Te introduced by the ‘‘red-bearing’’ chromosome,
the homozygous red will be depressed in the F,.

(C) If two lethals, both of which are identical, are in-
troduced at the same time, one by the red and one by the
pink, all classes will be equally depressed,’ and the re-
sults as far as concerns the F, ratio will be the same as if
there were no lethals present, i. e., the 3:1 ratio will be
realized.

(D) If two lethals that are different are introduced
at the same time, one by the red and one by the pink, both
the homozygous classes (RR and PP) will be depressed,
but not the RP. There would be somewhat fewer pinks
than expected in the F,.

I wish to acknowledge my indebtedness to Professor
T. H. Morgan, whose kind attention and suggestions both
throughout the foregoing experiments and in the prep-
aration of the present report, were invaluable. I also
wish to express my appreciation to Mr. H. J. Muller to
whom I owe some suggestions concerning the ph pb
tion of the results.

“LITERATURE CITED

Hyde, R. R. Fertility and Sterility in Drosophila ampelophila. Jour.
. Zool., Vol. 17, No. 2, August, 1914,

Morgan, T. H. An Attempt to Analyze the Constitution of the Chromo-
somes on the Basis of Sex-Limited Inheritance in Drosophila. Jour.
Exp. Zool., Vol. 11, No. 4. .

Two Sex-Linked Lethal Factors in Drosophila and their Influence on the
Sex-Ratio. Jour. Exp. Zool., Vol. 17, No. 1, July, 1914.

9 ‘í Crossing-over”’ is ignored as the character of the results is not changed
thereby.

SHORTER ARTICLES AND DISCUSSION

SELECTION, SUGAR-BEETS AND THRIPS

A DISCOVERY of great importance to students of genetics has
recently been made by one of the plant-breeders' of the U. S.
Department of Agriculture, viz., that beets are regularly cross-
pollinated and that an important agent in the process is a minute
inconspicuous insect, so small that it readily passes ‘‘through the
meshes of fine silk chiffon.’’

To understand fully the theoretical importance of this dis-
covery one need only recall the large attention given to the
sugar-beet in recent adverse criticisms of the selection-theory.
De Vries in his ‘‘Mutationstheorie,’’ p. 72, cites the case of the
sugar-beet as showing the most systematic, refined and elaborate
selection known for any cultivated plant, and yet as being with-
out any permanent effect in raising the sugar content of the beet.
For, although the average sugar content of the beet has by syste-
matic selection been practically doubled in the last 60 years,
De Vries holds the improved racial condition to be unstable and
thinks that the improved race would within a few generations
revert to its old level of sugar-content if the selection were dis-
continued. His reason for thinking so is the familiar fact that
the offspring of the best selected beets are on the average not
quite so good as their selected mother-beets, but show a tendency
to regress downward toward the old level of sugar content. It
should be pointed out, however, that in reality regression is not
toward the original average of 7 or 8 per cent. sugar-content, but
toward an average twice as high as this. For De Vries’s varia-
tion polygon (l. c., Fig. 22) for the sugar content of 40,000 beets
shows a nearly kyara probability curve about a mode at
15.5 per cent. It is to be supposed therefore that regression
would occur toward this condition from both the upper and the
lower halves of the frequency polygon, rather than toward the
old average condition of 7-8 per cent., which, according to the
data of DeVries, is now rarely if ever seen in the improved race.
To have doubled the average sugar-content of the beet is cer-
tainly something of an achievement for selection; the form of

1 Shaw, Harry B., ‘‘Thrips as Pollinators of Beet Flowers,’’ Bull. No.
104, U. S. Dept. Agr., July 10, 1914.

121

122 THE AMERICAN NATURALIST [Vou. XLIX

the variation polygon indicates that the change is permanent, so
far as ordinary racial characters have permanency.

But why, it may be asked, has selection not achieved more in
this case? Why should the descendants of, say, a 25 per cent. beet
not score better than this? There are probably several reasons
why. (1) Physiological reasons probably offer obstacles. A beet
can not be formed which is all sugar. There has to be present
in the beet a machinery for manufacturing the sugar. Perhaps
25 per cent. is an impossibly high average for a race of beets.
(2) Perhaps the exceptional 25-per-cent. beet owes its extra
sweetness in part to environmental causes which are not per-
manent. In that case the extra sweetness is ‘‘ somatic rather
than germinal,’’ as we should say in the case of an animal.

(3) Finally the discovery that beets are never self-fertilized,
but in every generation are cross fertilized, explains why im-
provement of the beet through selection is so slow and tedious a
process. What progress could the animal breeder expect to make
if he were able to select only the dams, but never the sires, for his
flocks? This is the condition which confronts the plant breeder
in attempting to improve the sugar beet. The animal breeder
is often chided with the small numbers which his experiments
yield as compared with the enormous numbers which an ex-
periment with plants may produce, but the animal breeder has
at least this satisfaction that when the animals are securely
penned there need be no uncertainty about pedigrees.

The careful observations of Shaw show that thrips, so common
in the blossoms of plants and yet so minute as easily to escape
notice and to penetrate within silk nets and under paper bags,
may be a cause of unsuspected cross-pollination and unaccount-
able ‘‘ mutation ’’ in the breeding of cereals and other plants.

W. E. CASTLE
BUSSEY INSTITUTION,
October 24, 1914

A NOTE ON MULTIPLE ALLELOMORPHS IN MICE

Proressor T. H. Morean has recently published in this jour-
nal the results of some of his experiments on color inheritance in
mice. In this paper he offers material which he considers ‘‘evi-
dence establishing”’ a series of multiple allelomorphs. His series
consist at present of four forms, ‘‘yellow, gray white-belly, gray

No. 578] SHORTER ARTICLES AND DISCUSSION 123

gray-belly and black.’’ The essential point of his conclusion is
that no more than two of these conditions can be transmitted by
any one animal.

The fact that Cuénot in his series of classic papers on color
inheritance in mice (1902-1911) recognizes these same four types
as forming a group of allelomorphs is not mentioned by Morgan,
whose paper, without knowledge of Cuénot’s work, might well be
taken to contain ‘‘the evidence establishing this series of allelo-
morphs’’ as he himself considers that it does. Since Morgan
appears to have overlooked Cuénot’s work with these forms, it
may be interesting to give a brief statement of Cuénot’s results.

As early as 1903 Cuénot recognized that albinos, potentially
yellows, when crossed with black gave besides yellow offspring
either black or agouti young, but not both. This is, of course,
evidence that yellow, agouti and black are all allelomorphic to
one another. In 1904 he gives formule (p. 46) showing that he
considers this to be the case. At the same time he gives the.
ratios produced by crossing an albino potentially a heterozygous
gray (agouti) with a yellow carrying black, but no agouti, and
albinism. For present purposes the albinism in the cross is
negligible. Cuénot recognized that the ratio expected from this
cross was 2 yellow, 1 black and d agouti (gray). He obtained
34 yellow, 20 black and 16 agouti; the calculated numbers being
38:19:19. Sturtevant (1912) in discussing the allelomorphism
or coupling of black, agouti and yellow in mice has also over-
looked Cuénot’s results, for in mentioning the cross of a hetero-
zygous agouti with a yellow carrying black, he states “‘appar-
ently Morgan is the only one who has reported such a cross. He
obtained 4 yellows, 5 agoutis and 1 black.’’

To return to Cuénot’s work; in 1907 he made a report on the
hereditary behavior of the white bellied agouti variety (gris à
ventre blanc) which he considers allelomorphic to yellow, agouti
and black. On page 10 in speaking of determinants he says:
“Il y en a le même nombre dans les races unicolores et dans la
race grise; ces races diffèrent, non pas par la quantité de leurs
déterminants mais par la qualité.’’ This is essentially the idea
underlying multiple allelomorphism. Later in the same paper
he says of G, the agouti determinant ‘‘. . . il présent un grand
nombre de mutations: G’, N et J.” (@' = white bellied gray;
N =black and J=yellow.) On page 13 he tabulates the vari-
eties, in order of their dominance, yellow, white-bellied agouti,

124 THE AMERICAN NATURALIST [Vou. XLIX

agouti and black. Morgan reached the same order of dominance
in 1911 and has récently (1914) recorded them, beginning with
black, as follows:

b= black,
BS&=eray gray-belly,
BY = gray white-belly,
BY = yellow.

In 1908 Morgan published certain facts concerning the inher-
itance of the white-bellied gray pattern. Cuénot at once (1908)
publicly called Morgan’s attention to the similarity of their
material and added facts which showed that he had already
investigated the inheritance of this same pattern in 1907.
Morgan later acknowledged its similarity.

In 1911 Cuénot states plainly (p. 47) : ‘‘Les souris jaunes sont
characterisées par un déterminant J, allélomorphe a G, G’ et N,
at qui les domine tous dans les croisements... il n’y à que les
` zygotes renfermant J dominant un autre déterminant allélo-
morphe (G, G’ ou N) qui peuvent évolouer.’’

Morgan’s 1914 paper adds several detailed matings and records
the testing of yellows of both sexes. However, in most respects,
his work corroborates the pioneer experiments of Cuénot and does
so in such detail that he falls into the same error as did Cuénot
in considering ‘‘black’’ as a necessary member of the allelo-
morphie series. This is obviously incorrect for the whole series
of allelomorphs exists equally well in forms utterly lacking the
ability to produce black pigment as some of Morgan’s experi-
ments showed. The true series of allelomorphs is yellow, white
bellied agouti, gray-bellied agouti and non agouti (not black).

BUSSEY INSTITUTION,
October 19, 1914

LITERATURE CITED
Cuénot, L.
1902. La loi de Mendel et 1’héredité de la pigmentation chez les souris.
Arch. Zool. Exp. et Gén. (3), Vol. 10, Notes et revue, p. xxvii
1903. 2me note. Arch. Zool, Rap. et Gén. (4), Vol. 1, Notes et revue,

p. xxxiii.

1904. 3me note. Ibid., Notes et revue (4), Vol. 2, pp. xlx-lvi.

(1905. 4me note. Ibid., Notes et revue (4), Vol. 3, pp. exxiii—cxxxii.

1907. 5me note. Ibid., Notes et revue (4), Vol. 5, p. 1.

1908. 6me note. Sur quelque anomalies apparentes de proportions
incised Ibid. (4), Vol. 6, Notes et revue, pp. vii—xv.

No. 578] SHORTER ARTICLES AND DISCUSSION 125

1911. 7me note. Les déterminants de la couleur chez les souris, étude
comparative. Ibid., Notes et revue (4), Vol. 8, pp. xl-Ivi.
Morgan, T. H.
1908. Some Experiments in Heredity in Mice. Sei, N. S., Vol. 27,
p. 493.
1911. The Influence of Heredity and of Environment in Determining
the Coat Color in Mice. Annals N. Y. Acad. of Sct., Vol. 21,
pp. 87-117.
1914. Multiple Allelomorphs in Mice. Am. Nat., Vol. 48, pp. 449—458.
Sturtevant, A. H
912. Is There Association Between the Yellow and Agouti Factors in
Mice? AM. Nat., Vol. 46, pp. 368-371,

ON THE TIME OF SEGREGATION OF GENETIC
FACTORS IN PLANTS

In @nothera lamarckiana, Geerts (3) has observed that two mi-
crospores of each tetrad abort. From the results of reciprocal
crosses, De Vries (9) concluded that there was a segregation of
genetic factors between the aborted and unaborted pollen-grains.
In my crosses of Stizolobium species (2), half of the pollen-grains
abort in a random manner in the anthers of the F, hybrids; and
I can only explain the results of the breeding work on the hypoth-
esis that there is a segregation between the four microspores of
each tetrad. Hence I conclude that the segregation does not take
place before the cell-divisions which form the pollen-mother-cells,
but takes place in the divisions which form the microspores. In
other words, segregation occurs here, not among the cells of the
diploid generation, but at the moment of formation of the indi-
viduals of the haploid generation.

In the ovules of Stizolobium crosses I have shown that there
is a random segregation of aborted and normal embryo-saes; and
this agrees with the observations of Geerts on the functional mega-
sporesof O.lamarckiana. If somatic segregation occurred, there
would be a segregation of whole ovaries or parts of ovaries with
ovules all aborted or all normal, causing a distribution which
would differ markedly from the binomial distribution demanded
by a random segregation according to the law of chance. I have
shown that with lots of ovules each, the distribution of the
aborted and normal ovules corresponds to the binomial (1 + 1)”.
Hence segregation can not have taken place before the formation
of the nucellus of the ovule.

In many species and varieties of Citrus, as Strasburger (7)

. 125

126 THE AMERICAN NATURALIST [Vou. XLIX

and Osawa (4) have proved (and as I can confirm), embryos
are formed from the tissue of the nucellus adjacent to the em-
bryo-sac. I have also shown (1), as Strasburger suspected, that
a similar mode of formation prevails in certain varieties of Man-
gifera indica. In F, hybrids between certain Citrus species (8),
these adventive embryos do not show segregation; and the ad-
ventive embryos of a mango variety give plants nearly constant
to that variety. Hence segregation had not taken place when
the cells surrounding the megaspore-mother-cell were formed.*
The same conclusion follows on the work of Ostenfeld (5) and
Rosenberg (6) with certain Hieracia.
JOHN BELLING
FLORIDA AGRICULTURAL EXPERIMENT STATION

REFERENCES
1. Belling, J. Report of Florida Agricultural Experiment Station for 1908.
Pp. ex—cexxv, Plates VII-X, 1909.
- Belling, J. The Mode of Inheritance of Semi-sterility in Certain Hybrid
P Eie Zeitschrift für induktive Abstammungs- und Vererbungslehre,
Bd. XII, S. 303-342, 1914.

bo

3. Beiträge zur Kenntnis der Cytologie und der partiellen
Sterilität von Gnothera Lamarckiana. Récueil des Trav. Bot. Néerl.,
Vol. 5, pp. 93 ff., 1909.

4. Osawa, I. Gridlogies! and Experimental Studies in Citrus. Jour. Coll.

Agr. Imp. Univ, eed Vol. 4, No. 2, pp. 83-116, 1912.

. Ostenfeld, C. H. Further Studies on the Apogamy and Hybridization
of the Hieracia. piesa ind. Abst. u. Vererbungslehre, Bd. III, S.
241-285, 1910.

- Rosenberg, O. Cytological Studies on the Apogamy in Hieracium. Bot.
Tidskrift, Vol. 28, 1907.

7. Strasburger, E. Veber Polyembryonie. Jenaische Zeitschr. f. Naturwiss.,

Ba. 12, S. 647, 1878.

Swingle, W. T. New Types of Citrus Fruits. Proc. Florida State
Hortieultural Soe. for 1910, pp. 36-42. Plates I-VIII, 1910.

9. De Vries, H. Gruppenweise Artbildung, 1913.

oO

a

oO
+

1 Somatic segregation is not the only available hypothesis in the eases of
the double Matthiola and Petunia. For the double stock may have one half
or less of its pollen-grains ineffective for fertilization (compare Correns on
the pollination of Mirabilis jalapa, in Ber. Deutsch. Bot. Ges., Bd. 18, S.
422-435); and in the ev doubleness may be incompletely dominant, as
in the greenhouse earnatio

NOTES AND LITERATURE

REPULSION IN WHEAT?

THE evidence was furnished by the (F,) of the cross: Smooth
Black X Rough White. ‘‘Smooth Black” is a wheat obtained
from the (F,) of the Rivet X Fife cross and it breeds quite true.
Its glumes are absolutely glabrous and of a burnished black color.
“Rough White’’ is the well-known Essex Rough Chaff Wheat.
The glumes are very hairy and of the ordinary white color. The
(F,) sorted into the following classes:

Rough Rough Smooth Smooth
Black White Black White
120 43 47 3

The expectation for the 1:3:3:1 repulsion is:
109.8 49.9 49.3 3.3

‘“‘Blackness’’ is probably not a simple character for in the (F,)
various degrees of it occur—the patches of it on the glumes being
of various sizes and intensities of color. There is evidence that
it is closely connected with the ‘‘gray’’ color of Rivet glumes.

F. L. ENGLEDOw

THE DETERMINATION OF THE BEST VALUE OF THE
COUPLING-RATIO FROM A GIVEN SET OF DATA!

Mr. G. N. Couns has suggested in this journal? a general
method for determining the value to assign to the coupling-ratio —
for a given set of data. He has worked out the value of a coeffi-
cient of association for the whole series of possible integral ratios
1:1:1:1, 2:1:1:2, etc., and then used the observed value of the
same coefficient to decide which ratio gives the best agreement
with the facts. The method is very simple, but does not lead to
the value which is the most advantageous in a certain sense. If
F,, F,, F,, F,, are the set of theoretical frequencies for a given
value of the ratio and if F,’, F, F,’, F,’, are the observed fre-

1‘*A Case of Repulsion in Wheat,’’ by F. L. Engledow, St. John’s Col-

lege, Cambridge (Proc. Camb. Phil. Soe., Vol. 18).
1F, L. Engledow and G. Udny Yule (Prog. Cambridge Phil. Soc., XVII,
6).

127

128 THE AMERICAN NATURALIST [Vou. XLIX

quencies, and if x =X (F' — F )?/F, then the probability p that
in random sampling deviations of equal or greater improba-
bility will arise is a function of x? which decreases continu-
ally as x? increases. The best value of the ratio will then be
that value which makes p a maximum or x? a minimum. The
problem taken in the note is to determine this value. Un-
fortunately the solution is not a simple one, depending on an |
equation of the fourth degree. A few cases are, however, taken
as illustrations and the question of probable error is discussed.
The recognized fact that, especially when the coupling-ratio is
high, its value may receive considerable alteration without
greatly altering the closeness of agreement between theory and
fact, receives additional emphasis from some of the results given
and makes it clear that considerable caution must be used before
attaching importance to the precise values of high ratios.

F. L. E ann G. U. Y.
2 AM. Nat., XLVI. i

ve

THE | |
AMERICAN |

NATURALIS

Sciences with

p $

Re
Mis

on ge
ET E ate
oei
AUR AN
ee

Sor

The American Naturalist

MSS intended for erences and books, etc., intended for tisay should be
sent to the Editor of THE AMI ac NATURALIST, Garrison-on-Hud a re
ho summaries of research work n the
blems of idea evelaton oe especially welcome, and will be pote pope
in i publication
ne hundrea reprints of Spe iag are supplied to authors free of charge.
Further reprints will be supplied
bscriptions and sivertis arte should be sent to the gee The
subscription price is four dollars a year oreign postage is s and
papaia poeter robert She cents additional. The e for single pei is
forty cents. The advertising rates are Four Dollars for

THE SCIENCE PRESS
Lancaster, Pa. Garrison, N. Y.
Sse _ NEW YORK: Sub-Station 84

Entered as seco ter, April 2, 1908, at th t Office at Lancaster, Pa., under the Act of
ane Congress of stephen

Sec eee FOR SALE | JAPAN NATURAL HISTORY SPECIMENS
ARCTIC. ICELAND an EENLA Perfect Condition and Lowest Prices
foie jem og anos cai se ‘Specialty: Bird Skins, Oology, Paiol, Marine
A E d Animals and others. Catalogue free. Correspond-
ence Se ;

T. FUKAI, Naturalist, T
i _ Konosu, Saitama, Japan

THE
AMERICAN NATURALIST

VoL. XLIX March, 1915 No. 579

MUTATION EN MASSE
HARLEY HARRIS BARTLETT *

Dure the writer’s experiments with @nothera two
different species have been discovered of which certain
strains give rise by mutation to large numbers of dwarfs.
In both cases the dwarfs occur in far greater numbers
than experience would lead one to expect, even in the most
actively mutant strain. Similar, although not exactly
parallel, phenomena have been observed by both de Vries
and Davis in certain hybrid @notheras, but not, as far as
the writer knows, in any unhybridized species. Since the
cultures have now been continued long enough so that
there can be no doubt as to the accuracy of the observa-
tions, the least complicated of the two cases is here placed
on record. It concerns @nothera Reynoldsii sp.nov. (A
technical diagnosis of this species will be published else-
where.) The seeds from which the cultivated strain arose
were collected at Knoxville, Tennessee, in the fall of 1910,
by Dr. E. S. Reynolds, then connected with the botanical
department of the University of Tennessee.

Fig. 1 is a diagram showing the size and relationship
of the cultures of @nothera Reynoldsii which have thus
far been grown. No diversity was found in the small
F, and F, cultures, of only ten and five plants, respec-
tively, which were grown in 1911 and 1912. The F,
generation of twenty-six plants, grown in 1913, exhibited

‘1 From the Bureau of Plant Industry, U. S. Department of Agriculture,
Offce of Plant Physiological and Fermentation Investigations. Published
by the permission of the Secretary of Agriculture.

129

130 THE AMERICAN NATURALIST [ Vou. XLIX

a most unexpected segregation into three marked types,
forma typica, reproducing the parental form, and two
dwarf types, mut. semialta and mut. debilis, so named
because of their resemblance to the two classes of dwarfs
which de Vries? obtained from Œ. nanella X Œ. biennis.
Mut. semialta is about half as tall as f. typica and has a

P
LHO
MLD
E; IROI WILD SEED

491, lO PLANTS
ALL F TxPICA

L ARONI NO. 89, E IVCA aiie 0.05, F
A PA 105 PLANTS kh PLANT:
004 F: TYPICA

LITUTATTON

SF rman 89-3, F TIPICA A
1973,
lod 33} 79 PLANTS, CLASSIFIED AD

29 EIVA | 2 rer semua | 18 UT CERES
J l

4g PROT 89-3 9-3743, £F EKPA i
og 1914, 28 PLANTS iy ea aoe EE [VOT SLUTS Fy PROM 89-93-25, UT. DEBILIS
sr LUF, AP PLANTS
E

gram showing the size and relationship of the cultures of @no-
Wet oe 1911-1914,

very dense and showy inflorescence in which the fruits
and flowers are very little smaller than in the parent
form. The leaves, however, are decidedly reduced. Mut.
debilis is more variable in size than mut. semialta, but
averages about half as high as the latter. Its fruits and
flowers are somewhat reduced, but by no means propor-
tionately to the size of the plant. The leaves, on the
contrary, are much more reduced than those of mut.
semialta. The inflorescence is not as dense, but often
longer.

The unlooked-for occurrence of these types in the F;
of 1913 led to the duplication in 1914 of both the F, and
F, generations from seeds which had been left over from
former years. In 100 additional F, plants of the mutant

2‘‘Gruppenweise Artbildung,’? pp. 241-244.

No. 579] MUTATION EN MASSE 131

strain there were 99 plants of f. typica and one mutation
of a quite different type from either mut. semialta or
mut. debilis. The original F, culture had consisted of
26 plants, including two of f. typica, 16 of mut. semialta

Fig. 2. A random sample ae six plants from the Fs culture of 1913. No. 4
is f. typica; the rest are all mut. semialta.
and eight of mut. debilis. In the supplementary culture
of 53 sister-plants, grown in 1914, there were 27 plants
of f. typica, 16 of mut. semialta and ten of mut. debilis.

132 THE AMERICAN NATURALIST [Von XLIX

Fic. 3. Adjacent plants of mut. ogc se the left No. 89-3-25, chosen as
parent of one of the F, progenies) and mut. semialta (on the right; No, 89-3-24,
chosen as parent of st of the F, pando dich The small labels on the plants
are 10 cm. long. (Reduction same as in Fig. 4.)
There can therefore be no doubt that the F, was an essen-
tially uniform generation and that the F, was the first
generation to throw the two dwarf types, except perhaps
as rare mutations, which were not detected on account
of the small size of the cultures. In this connection it
may be remarked that the mutations of Gnothera Rey-
noldsii can not be detected in very young cultures with
any degree of precision. Up to the time the rosettes are
set out in the garden, after they have been started in the
greenhouse in pots, they show no consistent differences
among themselves. It happens that six seedlings of the
1913 F, were photographed before any diversity what-
ever had been detected in the culture. They must there-
fore be considered a random sample from the 26 plants.
All turned out to be mut. semialta except one, which was
f. typica. The photograph is reproduced as Fig. 2.

At maturity the contrast between the classes is very

No. 579] MUTATION EN MASSE 133

striking, and leaves no room for doubt as to the proper
classification of any individual. Fig. 3 shows adjacent
plants of mut. debilis and mut. semialta; Fig. 4 mut.
debilis and f. typica. Figs. 5, 6 and 7 show branches of
the three forms, on the same scale of reduction.

The F, generation, grown in 1914, consisted of the
progenies of two plants of each of the three types. The
two externally identical parent plants of f. typica (there

EN

ae |

Adjacent plants of mut. debilis (in front) and f. typica (behind ;

IG. 4. j
No. 89-3-13, chosen as parent of one of the F, progenies

). The small labels on
the plants are 10 em. long. (Reduction same as in Fig. 3.

134 THE AMERICAN NATURALIST [ Von. XLIX

were only two in the F, of 1913) proved to be of very
different genetic constitution. The progeny of one, num-
bering 100 plants, were all strictly like the parent, show-
ing not the slightest deviation from f. typica. The other
progeny, however, repeated the diversity of the F, gen-
eration, containing five plants of f. typica, 13 of mut.
semialta and five of mut. debilis in a culture of 24 plants.
This progeny, also, included one plant of a third dwarf

5. Mnothera Reynoldsii f. typica. Branch from F, plant No. 89-38-18,
chosen as parent of one of the F, cultures. The entire plant is shown in Fig. 4.

mutation, which will be referred to below as mut. bilonga.

The two F, plants of mut. semialta which were used as
parents gave very similar progenies, consisting of mut.
semialia and mut. debilis. In one case the numbers were
41 of mut. semialta and five of mut. debilis in a total of
46; in the other case, 83 of semialta and four of mut.
debilis in a total of 87.

No. 579] MUTATION EN MASSE 135

The two progenies from mut. debilis parents, contain-
ing 85 and 43 plants, respectively, were all mut. debilis
like the parents, except that each progeny contained one
individual of mut. bilonga. Before discussing the latter
mutation it may be well to capitulate.

1. The individuals of f. typica are of two kinds, (a)
those which do not throw dwarfs, and (b) those which
throw from 60 per cent. to 80 per cent. of dwarfs.

Fic, 6. Mut. semialta. Branches of Fs plant No. 89-3-23.

2. The dwarfs are of two kinds, one of which, mut.
semialta, is intermediate between f. typica and the ex-
treme dwarf, mut. debilis.

3. Mut. semialta reproduces itself in the greater part
of its progeny, but throws a small number (seemingly
about 7 per cent.) of mut. debilis.

4. Mut. debilis does not throw either f. typica or £
semialta. It comes true, except for the fact that it rarely
throws mut. bilonga. .

Mut. bilonga is by far the most interesting of the vari-
ants of Gnothera Reynoldsii. Tt has occurred once as a

136 THE AMERICAN NATURALIST [ Von. XLIX

primary mutation from f. typica, and twice as a secondary
mutation from mut. debilis. Although mut. debilis seems
to be an extreme recessive, derived from f. typica either
by the simultaneous or by the successive loss of two
factors for height, it throws mut. bilonga, which shows
a return to the stature of mut. semialta. In fact, mut.
bilonga would be identical with mut. semialta if it were
not for the difference in the length of the fruits. It has
already been stated that in both mut. semialta and mut.
debilis the fruits are by no means as reduced in size as
the foliage and stems. It seems almost as though the

IIG. 7. Mut. debilis. Branches of Fs plant No. 89-3-12. The entire plant is
shown in Fig. 4.

process of mutation, which results in the formation of
either of these dwarfs, does not involve the factors deter-
mining fruit size. In other words, the slight reduction in
size seems not to be due to a modification of the hereditary
qualities of the plant, but rather to diminished nutrition.
If this explanation is the true one, the fruits of mut.

No. 579] MUTATION EN MASSE 137

e,

~
y
D

|

a

à

# eet

fs
\>

i

e

shown in Figs. 4

debilis are small for the same reason that the late autumn
fruits on weak lateral branches of f. typica are smat.
When, by mutation to mut. bilonga, mut. debilis reassumes
the stature and foliage size of mut. semialta, there 1s a
modification of the characters which determine the length
of the fruit. Not only is the stature doubled, and the
length of the leaves doubled, but the length of the fruit
is also doubled. Mut. bilonga is to all outward appear-
ance the same as mut. semialta, except that the fruits are
twice as long. Thus, we have the anomalous situation
that mut. bilonga, a dwarf type, is characterized by the

138 THE AMERICAN NATURALIST [Vou. XLIX

Branches of mut. debilis (left; No. 89-3-21-1) and mut. bilonga
(right; No. 89-3-21-85) ee ce ng in both foliage and fruits. The
two forms were sisters in the progeny of one of the original examples of mut.
debilis which appeared in the Pe of 1913.
longest fruits in the subgenus Onagra. Txeptionai fruits
are 70 mm. long; the average length of six hand-pollinated
_ fruits was 62 mm. By way of comparison, it may be
stated that the length of the average fruit of f. typica is
about 33 mm., and that the longest is 38 mm. None of
the immediate allies of @nothera biennis have longer
fruits than those of O. Reynoldsii f. typica, although
there are allies of O. muricata in which the fruits average
as long or longer. There is no species, however, in which
the fruit length of mut. bilonga is even approached. Here
we have an apparent case of progressive mutation, which
will be tested out as soon as possible. Mut. bilonga has
not thus far been carried into a second generation. Both
it and the two other dwarfs are completely self-fertile,

No. 579] MUTATION EN MASSE 139

and furnish an abundance of good seed. It is planned to
make a biometrical study of fruit length next year, when
the second generation of mut. bilonga will be available.

In Fig. 8 the two branches on the left are mut. semialta;
the two on the right mut. bilonga. The plants which fur-
nished the material belonged to an F, culture from f.
typica, containing five plants of f. typica, 13 of mut.
semtalta, five of mut. debilis and one of mut. bilonga.
In Fig. 9, on the contrary, the contrast is between sister-
plants of mut. debilis and mut. bilonga in the progeny of
mut. debilis. A comparison of the figures will show the
identity of mut. bilonga from the two sources.

The phenomenon presented by @nothera Reynoldsii,
called mutation en masse for want of a better name, seems
of sufficient interest to justify this preliminary paper.
The fact that it appears in one of the short-styled, self-
pollinating species makes it of especial interest. An ex-
planation can hardly be attempted until the interrelation-
ships of the various derivations have been worked out by
a series of crosses. Nevertheless, it seems clear that
mutation en masse bears a certain degree of resemblance
to Mendelian segregation. The fundamental mutation
which causes the diversity possibly occurs in only one of
the two gametes in a generation preceding the one in
which diversity becomes manifest. It is masked by the
dominance of the parental characters transmitted through
the other gamete. Segregation then occurs in the follow-
ing generation. No explanation suggests itself for the
enormous surplus of dwarfs in the progenies exhibiting
diversity, unless perhaps it is that the results are com-
` plicated by selective germination or selective mortality.
At any rate, the ratios thus far obtained do not conform
to any Mendelian expectation. Larger cultures, to be
grown next year, may prove more enlightening.

To the mutationist, the most interesting problem pre-
sented by Gnothera Reynoldsii is the origin of mut.
bilonga from mut. debilis, involving, as now seems prob-
able, the omin of a new character.

>

THE ALBINO SERIES OF ALLELOMORPHS IN
GUINEA-PIGS

SEWALL WRIGHT

BUSSEY INSTITUTION

Aupinism is one of the most familiar color conditions
found in mammals. In all cases it has proved to be a
simple Mendelian recessive to the pigmented condition.
This was demonstrated for guinea-pigs by Castle and
Allen in 1903. In the present paper evidence will be pre-
sented showing that in guinea-pigs there are two grades
of pigmentation, intermediate between full intensity and
albinism, which form with these a series of four allelo-
morphs with dominance in the order of increasing pig-
mentation.

The most highly pigmented condition is also the most
familiar. To this condition, which we may call intensity,
belong those types of guinea-pigs which show in the fur
intense black pigmentation or the intense orange-yellow
known as ‘‘red,’’ or both. Examples are the blacks, the
reds, the golden agoutis and the black-and-red tortoise-
shells of the fanciers. All of these have black eyes.

The second condition in intensity and order of domi-
nance contains color varieties which have long been
known. In these black is reduced to a sepia-brown color
much like human brown hair, known very inappropriately
as ‘‘blue.’’ Red is reduced to yellow or cream. The
eye color remains black. Thus we have blues, creams,
silver agoutis and blue-and-cream tortoise-shells in place
of the four types mentioned as of the intense condition.
That these four so-called dilute types, as well as others-
not mentioned, differ from the intense types by the same
factor or factors has long been known. It is shown by
the fact that one can start with any of these and by crosses
with the appropriate intense variety produce any other

140 ;

No. 579] ALBINO SERIES OF ALLELOMORPHS 141

dilute variety in the second generation at latest. The
writer has done this for all of the types mentioned by
starting with the cream variety.

The third condition is not so familiar as the others. It
has appeared at the Bussey Institution only in the de-
scendants of three guinea-pigs brought from Peru by Pro-
fessor Castle in 1911. Among these black is reduced to
sepia (or ‘‘blue’’), indistinguishable from the ‘‘blue’’ of
the dilutes. Red is reduced to white. Nota trace of yel-
low pigment has been found in guinea-pigs with this al-
lelomorph. One of the most striking features of this
condition is the glowing red color of the eyes, easily dis-
tinguishable from the black eyes of the intense and dilute
guinea-pigs as well as from the pink eyes of the albinos.
There is a deficiency of pigment in both retina and iris.
Because of this feature this condition will be known as
red-eye. It was announced as an allelomorph of albinism
by Castle (1914). Permission has very kindly been given
the writer to present in this paper data on the red-eye
condition, in the work on which he has been associated.
With the red-eye factor, the blacks, reds, golden agoutis
and black-and-red tortoise-shells become red-eyed blues,
red-eyed whites, red-eyed silver agoutis and red-eyed blue-
and-white tortoise-shells, respectively. These four varie-
ties have all been obtained by crossing one of them, the
red-eyed silver agouti, with various stock guinea-pigs
and extracting the different combinations in F,. The
red-eyed white is an interesting variety thus derived.
Red-eyed whites have been tested by crosses with reds and
= creams and have been shown conclusively to be the red-
eyed representative of these varieties, such crosses having
given 9 reds, 3 creams, 6 red-eyed whites and 6 albinos
only. This red-eyed white demonstrates most forcibly
the complete inhibition of yellow in the presence of the
red-eye factor.

In the albino condition black disappears from the coat
except in patches on the nose, ears and feet, and occa-
sionally some sootiness on the back. In this connection
it is interesting to note that nose, ears and feet are gener-

142 THE AMERICAN NATURALIST [ Vou. XLIX

ally the most highly pigmented regions in the dilute and
red-eye conditions. In the albinos, yellow disappears en-
tirely, just as in the red-eyes. The eyes are pink, due
to the loss of all pigment from the iris and retina. The
blacks and golden agoutis are replaced by sooty albinos;
the reds, by clear albinos; and the black-and-red tortoise-
shells, by albinos in which nose, feet and ears are sooty
or white, depending on the location of the spots.

The effects of the four allelomorphs on the appearance
of eye and fur may be tabulated as follows:

Effects of On Eye Color On Black Fur | On Yellow Fur
Intensity Black Black | Red
iluti Black Sepia | Cream
Red-eye Red Sepia _ White
binism White White
(sepia points)

It should be added that dilution of pigmentation may be
produced by other factors than members of the albino
series of allelomorphs. In the foregoing discussion such
factors have been assumed to be absent. It may be said,
however, that by starting with variations which owe their
dilutions to factors which are independent of the albino
series, doubly dilute varieties have been produced on in-
troducing the dilution or red-eye allelomorphs of albi-
nism. The effects of dilution, red-eye and albinism on
brown pigment are parallel in all cases to their effects on
black.

It should be said that the sepia (or ‘‘blue’’) due to the
dilution factor, or to the red-eye factor, varies through a
wide range which intergrades with black. Thus Castle .
(1905) and Sollas (1909) recognized that both intense and
dilute forms of pigmentation occur commonly in guinea-
pigs, but did not suggest any factorial explanation be-
cause of this intergrading. The inheritance of these fluc-
tuations is at present under investigation.

If intensity, dilution, red-eye and albinism are allelo-
morphs, gametes should always carry one, but only one,
of the four. Zygotes must always have two representa-

No.579] ALBINO SERIES OF ALLELOMORPHS 143

tives from the series, never more or less, which two may,

of course, be alike. Thus with dominance in the order of
increasing pigmentation intense guinea-pigs should be
- homozygous, or else carry recessive dilution, recessive
` red-eye or recessive albinism, but never more than one of
these; dilute guinea-pigs should be homozygous, or carry
recessive red-eye or recessive albinism, but never both;
red-eyes should be either homozygous or carry recessive
albinism; and finally albinos should always be homozy-
gous and never have the power of transmitting intensity,
dilution, or red-eye to their descendants. All of these
types have been obtained and tested with results in har-
mony with expectation. In this paper only a few of the
most critical crosses will be given, reserving a more de-
tailed discussion for a later paper.

Red-eye, as mentioned before, has occurred only in the
descendants of certain pigs brought from Peru. No al-
binos appeared in the pure stock. In the pure races, red-
eye behaved as a simple recessive. Thus intense by in-
tense gave 63 intense and 19 red-eye young, while red-eye
by red-eye gave only red-eyes, 28 in number.

As red-eyes had never appeared in our stock guinea-
pigs, it was natural to expect that any stock pig crossed
with red-eye, an apparently recessive condition, would
give only intense young. As a matter of fact, however,
numerous red-eye young appeared in F,. The next ques-
tion was whether one kind of stock differed from another
in its power of bringing about this apparent reversal of
dominance. A study of the records soon showed that
albinism had something to do with the matter, as the fol-
lowing tabulation indicates:

Intense Red-eye Albino
Red-eye X intense ...........- 48 22 20
Modave < Alino... cs cis 0 80 32

The difference in result of the two sorts of matings is
obviously significant. With an intense parent there were
more than 50 per cent. intense young and comparatively
few red-eye young. With an albino parent there were no

144 THE AMERICAN NATURALIST [Vou. XLIX

intense but numerous red-eye young. The intense ani-
mals and the albinos used were of the same stock and
hence could differ consistently only as regards albinism.
Clearly red-eye is not inherited independently of albinism.
That the apparent reversal of dominance is not due
merely to the presence of recessive albinism is shown by
the fact that F, red-eyes crossed inter se produced 89 red-
eyes, 36 albinos, but no intense young. Among these F,
red-eyes, some were demonstrated to be free from re-
cessive albinism. This complete inability of albinos to
transmit the intense allelomorph of red-eye in crosses
with the latter can only be interpreted in one way, aside
from linkage hypotheses. The intense allelomorph of
red-eye must also be an allelomorph of albinism. Thus
red-eye must either be an allelomorph of albinism or be
albinism itself, genetically, plus a modifying factor. The
latter rather improbable hypothesis has been definitely
eliminated. The pure South American stock under this
hypothesis must be homozygous for all such modifying
factors, since no albinos have appeared among them. The
hypothetical modifying factor must be a dominant unit
factor to account for the results given for F, and F,.

Some pure South American intense animals were crossed
with stock albinos. They produced F, intense young,
which must (under the hypothesis) be heterozygous for
albinism and for the modifying factor. These F, in-
tense young were back-crossed with stock albinos. There
should be both red-eyes and albinos among the young,
25 per cent. of each, if red-eye is albinism associated with
a modifier. But no red-eyes actually appeared. There
were 14 intense and 25 albinos. The chance that no red-
eyes would appear in 39 young is (3/4), or .000,001.
Thus the hypothesis that red-eye is albinism plus a modi-
fier may be dismissed.

On the view that red-eye is a dominant allelomorph of
albinism, the results above are easily explained. Red-
eye by albino or by red-eye should give only red-eyes and
albinos. The F, intense, which must carry recessive al-

No. 579] ALBINO SERIES OF ALLELOMORPHS 145

binism because of the stock albino parent, should not be
able to transmit red-eye. As for hypotheses of linkage,
it need only be said that no results have been obtained
which require them. The critical crosses have all been
made reciprocally as regards sex.

The fact that dilutes are more or less intermediate be-
tween intense and red-eye varieties suggested the fol-
lowing experiments which were designed to demon-
strate at once whether there was any relation in inheri-
tance between albinism and dilution. Dilutes were crossed
with albinos from certain stocks which for years had given
only intense and albino young, but no dilutes. Second,
intense pigs from these same stocks, which had given only
intense and albino young, were crossed with albinos from
dilute stock. If intensity and dilution form a pair of
allelomorphs which segregate independently of the pair,
color and albinism (as is the case in mice and rabbits),
these two crosses must give identical results. In each
case color is introduced by one parent, albinism by the
other; the intensity of certain stocks by one parent, the
dilution of certain stocks by the other parent. In fact
identical results should be obtained regardless of
whether dilution is due to a unit factor or multiple fac-
tors, or even whether its inheritance is Mendelian or not,
povided only that it is inherited ind i yofalbinism.
As it happened, these two crosses gave strikingly differ-
ent results. The first cross, viz., dilute by albino from
intense stock, gave only dilutes, 37 in number, aside from
albinos. The second cross, intense from intense stock
by albino from dilute stock, gave only intense young, 49
in number, aside from albinos. These different results
can only be explained by assuming that a member of the
albino series of allelomorphs, recessive to intensity, is
essential to dilute animals. Thus genetically, dilution
may be albinism plus a modifying factor, or it may be
red-eye plus a modifying factor, or it may be a new allelo-
morph. The last has proved to be the case. A large
number of intense animals, at least half of which under

146 THE AMERICAN NATURALIST [ Vou. XLIX

the first hypothesis would be expected to be heterozygous
both for albinism and the modifying factor, have been
crossed with albinos. None of them have given both
albinos and dilutes. It is of course expected under the
hypothesis of allelomorphism that no intense animal
should transmit both albinism and dilution. The view
that dilution is red-eye plus a modifying factor has also
been eliminated. Red-eye has never appeared in our
stock guinea-pigs, which must therefore be pure for any
modifying factor which might changea red-eye to a dilute.
But red-eyes of any generation crossed with stock albinos
have never given dilutes in 105 young. Therefore dilu-
tion can not be red-eye plus a modifying factor. The only
remaining hypothesis is that dilution is an allelomorph of
red-eye and albinism. It is dominant to both since dilute
by dilute has often given red-eyes and albinos, while the
latter varieties crossed inter se have never given dilutes.

Thus we have four allelomorphs corresponding to four
grades of pigmentation. The existence of such series has
a bearing on the nature of unit factors. The results could
be explained by perfect coupling, but such an explanation
seems highly arbitrary where the characters fall into a
natural physiological series. The series seems to suggest
that we have four variations in some one entity. Fur-
thermore, while we have only four such variations at the
Bussey Institution, it may well be that others exist else-
where, forming perhaps a continuous series. Such varia-
tions in this factor probably do not occur frequently.
When they do occur they probably take place by distinct
steps. The rather frequent occurrence of albinos in wild
species, without intermediates, indicates that variation
from one extreme to the other in the condition of the
factor may take place.

A point which has a bearing on the physiology of pig-
ment is the fact that members of the albino series of allelo-
morphs do not cause diminution in quantity of pigment,
merely as pigment, but affect yellow pigment differently
from black. This is seen most clearly in the red-eyes, in

No.579] ALBINO SERIES OF ALLELOMORPHS 147

which yellow is completely inhibited, while black is only
slightly affected.

Finally the question is raised whether anything similar
to this can be found among other mammals. Albinos are
found in many mammals as well as in lower animals. Asto
multiple allelomorphs, the case of the Himalayan rabbit
is well known and compares well with the guinea-pig
series. The Himalayan rabbit with its pink eyes and
white fur with dark patches on nose, ears and feet is com-
parable to the guinea-pig albino. The complete albino
rabbit recessive to the Himalayan is lower in the series
than anything known in guinea-pigs. The dilution of the
blue rabbit as well as that of the blue mouse and maltese
cat is of a different type from the guinea-pig dilution.
As Miss Sollas has shown, the pigment is clumped in-
stead of uniformly decreased in quantity. The effect is
slate-blue instead of sepia-brown. Mr. H. D. Fish has
made crosses (unpublished data, to which I refer with his
permission) which show as expected that rabbit dilution
is inherited wholly independently of albinism.

In man, we have albinos which are probably comparable
to guinea-pig and rabbit albinos. A study of the enor-
mous collection of data in the Monograph of Pearson, Net-
tleship and Usher convinces one that albinism in man is
recessive. But as Pearson points out, there are many
grades of albinism and each grade tends to maintain its
identity in inheritance. Among negroes there are albinos
with blue irises, red pupils, white skin and nearly white
hair. There are also darker grades, as with brown skin,
eyes and hair. There is no sharp line anywhere between
the complete albinos and the so-called xanthous types.
In the white races albinos pass into the extreme blonds in
a continuous series. In fact, study of records convinces
one that in some cases the same factor may produce well-
marked albinism, with red eyes, nystagmus and photo-
phobia in one member of a family, but merely extreme
blondism in another. It is worthy of note that human
light hair resembles closely the sepia of dilute guinea-

148 THE AMERICAN NATURALIST [ Vou. XLIX

pigs, and not at all the slate blue of dilute rabbits, mice
and cats. Thus, while I have not been able to find any
critical evidence, the suggestion seems worth making that
a series of allelomorphs of albinism may be in part re-
sponsible for differences in intensity of human pigmen-
tation.

Summing up: the results in guinea-pigs and rabbits sug-
gest that there is a hereditary factor in mammals, which
may exist with stability at different stages of divergence
from the normal; that divergence from the normal in the
factor tends to produce in the animal a corresponding re-
duction in the quantity of melanin pigment throughout
the body, conspicuously in fur, skin and eyes, of which re-
duction the limit is complete albinism; that in this reduc-
tion the qualitative differentiation of the pigment is a
factor, in that yellow pigment is affected more strongly
than black and its threshold of complete inhibition is
reached with less divergence of the factor; that in the re-
duction the location of the pigment is a factor, in that there
is less tendency toward reduction at the extremities—feet,
ears and nose—than elsewhere in skin and fur; that,
finally, any stage of divergence is dominant to any stage
‘more remote from the normal.

In conclusion, I wish to thank Professor Castle for the
opportunity to carry on this work and for numerous
suggestions during its progress.

December 24, 1914

LITERATURE CITED.

Castle, W. E. 1905. Heredity of Coat Characters in Guinea-Pigs and
Rabbits. Carnegie Institution of Washington. Publication

No. 23.
1914. Some New Varieties of Rats and Guinea-Pigs and their Rela-
tion to Problems of Color Inheritance. AMER, NAT., Vol. 48,
65

p. 65.

Castle, W. E. and G. M. Allen. 1903. The Heredity of Albinism. Proc.
Am. Ac. Arts and Sci., Vol. 38, No. 21, p. 603.

Pearson, K., E. Nettleship, and C. H. Usher. 1913. A Monograph on
Albinism in Man. Parts I and IV. Dulau & Co., London.

Sollas, I. B. J. 1909. Inheritance of Color and of Supernumerary Mam-
mae in Guinea-Pigs, with a Note on the Occurrence of a
Dwarf Form. Reports to the Evol. Com. of the Roy. Soc.,
No. 5.

PROGRESSIVE EVOLUTION AND THE ORIGIN
OF SPECIES?

PROFESSOR ARTHUR DENDY

THE opening years of the present century have wit-
nessed a remarkable development of biology as an ex-
perimental science, a development which, however full
of promise it may be for the future, for the time being
appears to have resulted in a widespread disturbance
of ideas which have themselves only recently succeeded
in gaining general acceptance. The theory of organic
evolution, plainly enough enunciated at the close of the
eighteenth and the beginning of the nineteenth century
by Buffon, Lamarck, and Erasmus Darwin, remained un-
convincing to the great majority of thinking men until
the genius of Charles Darwin ‘not only brought together
and presented the evidence in such a manner that it could
no longer be ignored, but elaborated a logical explana-
tion of the way in which organic evolution might be sup-
posed to have taken place. Thanks to his labors and
those of Alfred Russel Wallace, supported by the power-
ful influence of such men as Huxley and Hooker, the
theory was placed upon a firm foundation, in a position
which can never again be assailed with any prospect of
success.

This statement is, I believe, entirely justified with re-
gard to the theory of organic evolution itself, but the
case is very different when we come to investigate the
position of the various subsidiary theories which have
been put forward from time to time with regard to what
may perhaps be termed the modus operandi, the means
by which organic evolution has been effected. It is in
this field that controversy rages more keenly than ever
before. Lamarck told us that evolution was due to the

1 Address of the president of the section of zoology, British Association
for the Advancement of Science, Australia, 1914.

149

Do THE AMERICAN NATURALIST [VoL. XLIX

accumulated results of individual effort in response to a
changing environment, and also to the direct action of
the environment upon the organism. Darwin and Wal-
lace taught us that species originated by the natural se-
lection of favorable variations, and under the influence
of Weismann’s doctrine of the non-inheritance of ac-
quired characters the theory of natural selection is in
danger of becoming crystallized into an inflexible dogma.
In recent years de Vries has told us that species arise
by sudden mutations, and not by slow successive changes,
while one of the most extreme exponents of ‘‘Mendel-
ism,’’ Professor Lotsy, lately informed us that all species
arise by crossing, and seriously suggested that the ver-
tebrate type arose by the crossing of two invertebrates!

This curious and many-sided divergence of opinion
amongst expert biologists is undoubtedly largely due to
the introduction of experimental methods into biological
science. Such methods have proved very fruitful in
results which at first sight seem to be mutually contra-
dictory, and each group of workers has built up its own
theory mainly on the basis of observations in its own
restricted field.

Professor Bateson has said in his recently published
‘‘ Problems of Genetics’’:

When ... we contemplate the problem of evolution at large the
hope at the present time of constructing even a mental picture of that
process grows weak almost to the point of vanishing. We are left
wondering that so lately men in general, whether scientific or lay, were
so easily satisfied. Our satisfaction, as we now see, was chiefly founded
on ignorance.”

In view of this striking pronouncement on the part
of one who has devoted his life with signal success to
the experimental investigation of evolutionary problems,
the remarks which I propose to lay before you for your
consideration to-day may well appear rash and ill-ad-
vised. I cannot believe, however, that the position is
really quite so black as it is painted. We must perforce
admit that the divers theories with regard to the work-

2** Problems of Genetic,’’ p. 97.

No. 579] PROGRESSIVE EVOLUTION 151

ing of organic evolution cannot all be correct in all their
details, but it may be that each contains its own elements
of truth, and that if these elements can but be recognized
and sorted out, they may perhaps be recombined in such
a form as to afford at any rate a plausible working
hypothesis. We must bear in mind from the outset that
in dealing with such a complex problem many factors
have to be taken into account, and that widely different
views on the question may be merely one-sided and not
necessarily mutually exclusive.

I take it there are three principal facts, or groups of
facts, that have to be accounted for by any theory
of organic evolution:

1. The fact that, on the whole, evolution has taken
_ place in a progressive manner along definite and di-
vergent lines.

2. The fact that individual animals and plants are
more or less precisely adapted in their organization and
in their behavior to the conditions under which they have
to live.

3. The fact that evolution has resulted in the existence
on the earth to-day of a vast number of more or less
well-defined groups of animals and plants which we call
species.

The first of these facts appears to me to be the most
fundamental, and at the same time the one to which
least attention is usually paid. The great question, after
all, is, Why do organisms progress at all instead of re-
maining stationary from generation to generation? To
answer this question it is not necessary to go back to
the beginning and consider the case of the first terres-
trial organisms, whatever they may have been, nor are
we obliged to take as illustrations the lowest organisms
known to us as existing at the present day. We may
consider the problem at any stage of evolution, for at
each stage progress is, or may be, still taking place.
We may even begin by considering what is usually re-
garded as the highest stage of all, man himself; and
indeed this seems the most natural thing to do, for we

152 THE AMERICAN NATURALIST [Vov. XLIX

certainly know more about the conditions of progress in
man than in any other organism. I refer, of course, at
the moment, not to progress in bodily organization, but
to progress in the ordinary sense of the word, the prog-
ress, say, of a family which rises in the course of a few
generations from a position of obscure poverty to one
of wealth and influence. You may perhaps say that such
a case has no bearing upon the problem of organic evo-
lution in a state of nature, and that we ought to confine
our attention to the evolution of bodily structure and
function. If so, I must reply that you have no right
to limit the meaning of the term evolution in this man-
ner; the contrast between man and nature is purely ar-
bitrary; man is himself a living organism, and all the
improvements that he effects in his own condition are
part of the progress of evolution in his particular case.
At any rate I must ask you to accept this case as our
first illustration of a principle that may be applied to
organisms in general.

If we inquire into the cause of the progress of our
human family I think there can be only one answer—
it is due to the accumulation of capital, or, as I should
prefer to put it, to the accumulation of potential energy,
- either in the form of material wealth or of education.
` What one generation saves is available: for the next,
and thus each succeeding generation gets a better start
in life, and is able to rise a little higher than the pre-
ceding one.

Every biologist knows, of course, that there are many
analogous cases amongst the lower animals, and also
amongst plants. The accumulation of food-yolk in the
egg has undoubtedly been one of the chief factors in the
progressive evolution of animals, although it has been
replaced in the highest forms by a more effective method
of supplying potential energy to the developing off-
spring. It may indeed be laid down as a general law
that each generation, whether of animals or of plants,
accumulates more energy than it requires for its own
maintenance, and uses the surplus to give the next gen-

No. 579] PROGRESSIVE EVOLUTION 153

eration a start in life. There is every reason to believe
that this has been a progressive process throughout the
whole course of evolution, for the higher the degree or
organization the more perfect do we find the arrange-
ments for securing the welfare of the offspring.

We cannot, of course, trace this process back to its
commencement, because we know nothing of the nature
of the earliest living things, but we may pause for a
moment to inquire whether any phenomena occur
amongst simple unicellular organisms that throw any
light upon the subject. What we want to know is—How
did the habit of accumulating surplus energy and hand-
ing it on to the next generation first arise?

Students of Professor H. S. Jennings’s admirable
work on the ‘‘Behavior of the Lower Organisms’’ will
remember that his experiments have led him to the con-
clusion that certain Protozoa, such as Stentor, are able
to learn by experience how to make prompt and effective
responses to certain stimuli; that after they have been
stimulated in the same way a number of times they make
the appropriate response at once without having to go
through the whole process of trial and error by which it
was first attained. In other words, they are able by
practise to perform a given action with less expenditure
of energy. Some modification of the protoplasm must
take place which renders the performance of an act the
easier the oftener it has been repeated. The same is,
of course, true in the case of the higher animals, and we
express the fact most simply by saying that the animal
establishes habits. From the mechanistic point of view
we might say that the use of the machine renders it more
perfect and better adapted for its purpose. In the
present state of our knowledge I think we cannot go be-
yond this, but must content ourselves with recognizing
the power of profiting by experience as a fundamental
property of living protoplasm.

It appears to me that this power of profiting by ex-
perience lies at the root of our problem, and that in it

154 THE AMERICAN NATURALIST [ Vou. XLIX

we find a chief cause of progressive evolution. Jennings
speaks of the principle involved here as the ‘‘ Law of the
readier resolution of physiological states after repeti-
tion,’’ and, similarly, I think we must recognize a ‘‘Law
of the accumulation of surplus energy’’ as resulting
therefrom. Let us look at the case of the accumulation
of food-yolk by the egg-cell a little more closely from
this point of view. Every cell takes in a certain amount
of potential energy in the form of food for its own use.
If it leads an active life either as an independent or-
ganism or as a constituent part of an organism, it may
expend by far the greater part, possibly even the whole,
of that energy upon its own requirements, but usually
something is left over to be handed down to its imme-
diate descendants. If, on the other hand, the cell ex-
hibits very little activity and expends very little energy,
while placed in an environment in which food is abun-
dant, it will tend to accumulate surplus energy in excess
of its own needs. Such is the case with the egg-cells
of the multicellular animals and plants. Moreover, the
oftener the process of absorbing food-material is re-
peated the easier does it become; in fact, the egg-cell es-
tablishes a habit of storing up reserve material or food-
. yolk. Inasmuch as it is a blastogenic character, there
can be no objection to the supposition that this habit
will be inherited by future generations of egg-cells. In-
deed we are obliged to assume that this will be the case,
for we know that the protoplasm of each succeeding gen-
eration of egg-cells is directly continuous with that of
the preceding generation. We thus get at any rate a
possibility of the progressive accumulation of potential
energy in the germ-cells of successive generations of
multicellular organisms, and, of course, the same argu-
ment holds good with regard to successive generations
of Protista.

It would seem that progressive evolution must follow
as a necessary result of the law of the accumulation of
surplus energy in all cases where there is nothing to

No. 579] PROGRESSIVE EVOLUTION 155

counteract that law, for each generation gets a better
start than its predecessor, and is able to carry on
a little further its struggle for existence with the en-
vironment. It may be said that this argument proves
too mueh, that if it were correct all organisms would by
this time have attained to a high degree of organization,
and that at any rate we should not expect to find such
simple organisms as bacteria and Amebe still surviving.
This objection, which, of course, applies equally to other
theories of organic evolution, falls to the ground when
we consider that there must be many factors of which we
know nothing which may prevent the establishment of
progressive habits and render impossible the accumula-
tion of surplus energy. Many of the lower organisms,
like many human beings, appear to have an inherent in-
eapacity for progress, though it may be quite impossible
for us to say to what that incapacity is due.

It will be observed that in the foregoing remarks I
have concentrated attention upon the storing up of re-
serve material by the egg-cells, and in so doing have
avoided the troublesome question of the inheritance of
so-called acquired characters. I do not wish it to be
supposed, however, that I regard this as the only direc-
tion in which the law of the accumulation of surplus
energy can manifest itself, for I believe that the accu-
mulation of surplus energy by the body may be quite as
important as a factor in progressive evolution as the
corresponding process in the germ-cells themselves.
The parents, in the case of the higher animals, may sup-
ply surplus energy, in the form of nutriment or other-
wise, to the offspring at all stages of its development,
and the more capital the young animal receives the better
will be its chances in life, and the better those of its own
offspring.

In all these processes, no doubt, natural selection plays
an important part, but, in dealing with the accumulation
of food material by the egg-cells, one of my objects has
been to show that progressive evolution would take place

156 THE AMERICAN NATURALIST _[Vou. XLIX

even if there were no such thing as natural selection, that
the slow successive variations in this case are not chance
variations, but due to a fundamental property of living
protoplasm and necessarily cumulative.

Moreover, the accumulation of surplus energy in the
form of food-yolk is only one of many habits which the
protoplasm of the germ-cells may acquire in a cumu-
lative manner. It may learn by practise to respond with
increased promptitude and precision to other stimuli
besides that of the presence of nutrient material in its
environment. It may learn to secrete a protective mem-
brane, to respond in a particular manner to the presence
of a germ-cell of the opposite sex, and to divide in a
particular manner after fertilization has taken place.

Having thus endeavored to account for the fact that
progressive evolution actually occurs by attributing it
primarily to the power possessed by living protoplasm
of learning by experience and thus establishing habits
by which it is able to respond more quickly to environ-
mental stimuli, we have next to inquire what it is that
determines the definite lines along which progress mani-
fests itself.

Let us select one of these lines and investigate it as
fully as the time at our disposal will permit, with the
view of seeing whether it is possible to formulate a
reasonable hypothesis as to how evolution may have
taken place. Let us take the line which we believe has
led up to the evolution of air-breathing vertebrates.
‘The only direct evidence at our disposal in such a case is,
of course, the evidence of paleontology, but I am going
to ask you to allow me to set this evidence, which, as you
know, is of an extremely fragmentary character, aside,
and base my remarks upon the ontogenetic evidence,
which, although indirect, will, I think, be found sufficient
for our purpose. One reason for concentrating our at-
tention upon this aspect of the problem is that I wish
to show that the recapitulation of phylogenetic history
in individual development is a logical necessity if evolu-
tion has really taken place.

No. 579] PROGRESSIVE EVOLUTION 157

We may legitimately take the nucleated Protozoon
cell as our starting point,.for, whatever may have been
the course of evolution that led up to the cell, there can
be no question that all the higher organisms actually
start life in this condition.

We suppose, then, that our ancestral Protozoon ac-
quired the habit of taking in food material in excess of
its own requirements, and of dividing into two parts
whenever it reached a certain maximum size. Here
again we must, for the sake of simplicity, ignore the
facts that even a Protozoon is by no means a simple
organism, and that its division, usually at any rate, is a
very complicated process. Each of the daughter-cells
presently separates from its sister-cell and goes its own
way as a complete individual, still a Protozoon. It
Seems not improbable that the separation may be due
to the renewed stimulus of hunger, impelling each cell
to wander actively in search of food. In some cases,
however, the daughter-cells remain together and form a
colony, and probably this habit has been rendered pos-
sible by a sufficient accumulation of surplus energy in
the form of food-yolk on the part of the parent render-
ing it unnecessary for the daughter-cells to separate in
search of food at such an early date. One of the forms
of colony met with amongst existing Protozoa is the
hollow sphere, as we see it, for example, in Spherozoum
and Volvox, and it is highly probable that the assump-
tion of this form is due largely, if not entirely, to what
are commonly called mathematical causes, though we
are not in a position to say exactly what these causes
may be. The widespread occurrence of the blastosphere
or blastula stage in ontogeny is a sufficiently clear indi-
cation that the hollow, spherical Protozoon colony
formed a stage in the evolution of the higher animals.

By the time our ancestral organism has reached this
stage, and possibly even before, a new complication has
arisen. The cells of which the colony is composed no
longer remain all alike, but become differentiated, pri-

158 THE AMERICAN NATURALIST [ Von. XLIX

marily into two groups, which we distinguish as somatic-
cells and germ-cells respectively.

From this point onwards evolution ceases to be a
really continuous process, but is broken up into a series
of ontogenies, at the close of each of which the organism
has to go back and make a fresh start in the unicellular
condition, for the somatic cells sooner or later become
exhausted in their conflict with the environment and
perish, leaving the germ-cells behind to take up the run-
ning. That the germ-cells do not share the fate of the
somatic cells must be attributed to the fact that they take
no part in the struggle for existence to which the body
is exposed. They simply multiply and absorb nutriment
under the protection of the body, and therefore retain
their potential energy unimpaired. They are in actual
fact, as is so often said, equivalent to so many protozoa,
and, like the protozoa, are endowed with a potential
immortality.

We know that, if placed eee suitable conditions, or
in other words, if exposed to the proper environmental
stimuli, these germ-cells will give rise to new organisms,
like that in the body of which they were formerly en-
closed. One of the necessary conditions is, with rare
exceptions, the union of the germ-cells in pairs to form
zygotes or fertilized ova; but I propose, in the first in-
stance, for the sake of simplicity, to leave out of account
the existence of the sexual process and the results that
follow therefrom, postponing the consideration of these
to a later stage of our inquiry. I wish, moreover, to
make it quite clear that organic evolution must have
taken place if no such event as amphimixis had ever
occurred.

What, then, may the germ-cells be expected to do?
How are they going to begin their development? In
endeavoring to answer this question we must remember
that the behavior of an organism at any moment de-
pends upon two sets of factors—the nature of its own
constitution on one hand, and the nature of its environ-

No. 579] PROGRESSIVE EVOLUTION 159

ment on the other. If these factors are identical for
any two individual organisms, then the behavior of these
two individuals must be the same. If the germ-cells of
any generation are identical with those of the preceding
generation, and if they develop under identical condi-
tions, then the soma of the one generation must also be
identical with that of the other. Inasmuch as they are
parts of the same continuous germ-plasm—leaving out
of account the complications introduced by amphimixis
—we may assume that the germ-cells of the two genera-
tions are indeed identical in nearly every respect; but
there will be a slight difference, due to the fact that those
of the later generation will have inherited a rather larger
supply of initial energy and a slightly greater facility
for responding to stimuli of various kinds, for the
gradual accumulation of these properties will have gone
a stage further. The environment also will „be very
nearly identical in the two cases, for we know from ex-
periment that if it were not the organism could not de-
velop at all.

Throughout the whole course of its ontogeny the or-
ganism must repeat with approximate accuracy the
stages passed through by its ancestors, because at every
stage there will be an almost identical organism exposed
to almost identical stimuli. We may, however, expect
an acceleration of development and a slight additional
progress at the end of ontogeny as the result of the
operation of the law of the accumulation of surplus
energy and of the slightly increased facility in respond-
ing to stimuli. The additional progress, of course, will
probably be so slight that from one generation to the
next we should be quite unable to detect it, and doubtless
there will be frequent backslidings due to various causes.

We can thus formulate a perfectly reasonable explana-
tion of how it is that the egg first undergoes segmenta-
tion and then gives rise to a blastula resembling a hol-

3 This is, of course, a familiar idea. Compare Driesch, ‘‘ Gifford Lec-
tures,’’ 1907, p. 214.

160 THE AMERICAN NATURALIST [ Vou. XLIX

low protozoon colony; it does so simply because at every
stage it must do what its ancestors did under like condi-
tions. We can also see that progressive evolution must
follow from the gradual accumulation of additions at
the end of each ontogeny, these additions being rendered
possible by the better start which each individual gets
at the commencement of its career.

Let us now glance for a moment at the next stage in
phylogeny, the conversion of the hollow spherical proto-
zoon colony into the cclenterate type of organization,
represented in ontogeny by the process of gastrulation.
Here again it is probable that this process is explicable
to a large extent upon mechanical principles. Accord-
ing to Rhumbler,* the migration of endoderm cells into
the interior of the blastula is partly due to chemotaxis
and partly to changes of surface tension, which decreases
on the inner side of the vegetative cells owing to chem-
ical changes set up in the blastoceel fluid.

We may, at this point, profitably ask the question, Is
the endoderm thus formed an inherited feature of the
organism? The material of which it is composed is, of
course, derived from the egg-cell continuously by re-
peated cell-division, but the way in which that material
is used by the organism depends upon the environment,
and we know from experiment that modifications of the
environment actually do produce corresponding modifi-
cations in the arrangement of the material. We know,
for example, that the addition of salts of lithium to the
water in which certain embryos are developing causes
the endoderm to be protruded instead of invaginated,
so that we get a kind of inside-out gastrula, the well-
known lithium larva.

It appears, then, than an organism really inherits from
its parents two things: (1) a certain amount of proto-
plasm loaded with potential energy, with which to begin
operations, and (2) an appropriate environment. Ob-

4 Quoted by Przibram, ‘‘ Experimental Zoology,’’ English Trans., Part
I, p. 47.

No. 579] PROGRESSIVE EVOLUTION 161

viously the one is useless without the other. An egg
can not develop unless it is provided with the proper en-
vironment at every stage. Therefore, when we say that
an organism inherits a particular character from its
parents, all we mean is that it inherits the power to pro-
duce that character under the influence of certain en-
vironmental stimuli.© The inheritance of the environ-
ment is of at least as much importance as the inheritance
of the material of which the organism is composed. The
latter, indeed, is only inherited to a very small extent,
for the amount of material in the egg-cell may be almost
infinitesimal in comparison with the amount present in
the adult, nearly the whole of which is captured from
the environment and assimilated during ontogeny.

From this point of view the distinction between soma-
togenic and blastogenic characters really disappears, for
all the characters of the adult organism are acquired
afresh in each generation as a result ‘of response. to en-
vironmental stimuli during development. This is clearly
indicated by the fact that you cannot change the stimuli
without changing the result.

Time forbids us to discuss the phylogenetic stages
through which the ecelenterate passed into the cϾlomate
type, the cœlomate into the chordate, and the chordate
into the primitive vertebrate. We must admit that as
yet we know nothing of the particular causes that de-
termined the actual course of evolution at each succes-
sive stage. What we do know, however, about the in-
fluence of the environment, both upon the developing
embryo and upon the adult, is sufficient to justify us in
believing that every successive modification must have
been due to a response on the part of the organism to
some environmental change. Even if the external con-
ditions remained practically identical throughout long
periods of time, we must remember that the internal
conditions would be different in each generation, because

5 Compare Dr. Archdall Reid’s suggestive essay on ‘‘ Biological Terms as
(Bedrock, January, 1914).

162 THE AMERICAN NATURALIST [ Vor. XLIX

each generation starts with a slightly increased capital
and carries on its development a little further under in-
ternal conditions modified accordingly.

At this point it may be asked, Is the response to en-
vironmental stimuli a purely mechanical one, and, if so,
how can we account for the fact that at every stage in
its evolution the organism is adapted to its environment?
We shall have to return to this question later on, but it
may be useful to point out once more that there is good
reason to believe—especially from the experimental
work of Jennings—that the response of even a unicellu-
lar organism to stimuli is to a large extent purposive;
that the organism learns by experience, by a kind of
process of trial and error, how to make the response most
favorable to itself under any given change of conditions;
in other words, that the organism selects those modes
of response that are most conducive to its own well-
being. Under the term response to stimuli we must, of
course, include those responses of the living protoplasm
which result in modifications of bodily structure, and
hence the evolution of bodily structure will, on the whole,
be of an adaptive character and will follow definite lines.
There is good reason for believing, however, that many
minor modifications in structure may arise and persist,
incidentally as it were, that have no seen as adap-
tations.

One of the most remarkable and distinctive features
of the lower vertebrates is the presence of gill-slits as
accessory organs of ‘respiration. These gill-slits are
clearly an adaptation to aquatic life. When the ances-
tors of the higher vertebrates left the water and took to
life on land the gills disappeared and were replaced by
lungs, adapted for air-breathing. The change must, of
course, have been an extremely gradual one, and we get
a very clear indication of how it took place in the sur-
viving dipnoids, which have remained in this respect in
an intermediate condition between the fishes and the
amphibia, possessing and using both gills and lungs.

No. 579] PROGRESSIVE EVOLUTION 163

We also know that even the most highly specialized
air-breathing vertebrates, which never live in water and
never require gills or gill-slits at all, nevertheless pos-
sess very distinct gill-slits during a certain period of
their development. This is one of the most familiar il-
lustrations of the law of recapitulation, and my only
excuse for bringing it forward now is that I wish, before
going further, to consider a difficulty—perhaps more ap-
parent than real—that arises in connection with such
cases.

It might be argued that if gill-slits arose in response
to the stimuli of aquatic life, and if these stimuli are no
longer operative in the case of air-breathing vertebrates,
then gill-slits ought not to be developed at any stage of
their existence. This argument is, I think, fully met by
the following considerations.

At any given moment of ontogenetic development the
condition of any organ is merely the last term of a series
of morphogenetic stages, while its environment at the
same moment—which, of course, includes its relation to
all the other organs of the body—is likewise merely the
last term of a series of environmental stages. We have
thus two parallel series of events to take into considera-
tion in endeavoring to account for the condition of any
part of an organism—or of the organism as a whole—at
any period of its existence:

E, E, E, ... En environmental stages,
M,M.,M, ... Mn morphogenetic stages.

Ontogeny is absolutely conditioned by the proper cor-
relation of the stages of these two series at every point,
and hence it is that any sudden change of environment
is usually attended by disastrous consequences. Thus,
after the fish-like ancestors of air-breathing vertebrates
had left the water and become amphibians, they doubt-
less still had to go back to the water to lay their eggs,
in order that the eggs might have the proper conditions
for their development.

164 THE AMERICAN NATURALIST [Von XLIX

Obviously the environment can only be altered with
extreme slowness, and one of the first duties of the
parent is to provide for the developing offspring con-
ditions as nearly as possible identical with those under
which its own development took place. It is, however,
inevitable that, as phylogenetic evolution progresses,
the conditions under which the young organism develops
should change. In the first place, the mere tendency to
acceleration of development, to which we have already
referred, must.tend to dislocate the correlation between
the ontogenetic series and the environmental series.
Something of this kind seems to have taken place in the
life-cycle of many hydrozoa, resulting in the suppression
of the free medusoid generation and the gradual degen-
eration of the gonophore. But it is probably in most
cases change in the environment of the adult that is re-
sponsible for such dislocation.

To return to the case of the amphibians. At the
present day some amphibians, such as the newts and
frogs, still lay their eggs in water, while the closely re-
lated salamanders retain them in the oviducts until they
have developed into highly organized aquatic larve, or
even what is practically the adult condition. Kammerer
has shown that the period at which the young are born
can be varied by changing the environment of the parent.
In the absence of water the normally aquatic larve of
the spotted salamander may be retained in the oviduct
until they have lost their gills, and they are then born
in the fully-developed condition, while, conversely, the
alpine salamander, of which the young are normally born
in the fully-developed state, without gills, may be made
to deposit them prematurely in water in the larval, gill-
bearing condition. :

There can be no doubt that the ancestral amphibians
laid their eggs in water in a completely undeveloped con-
dition. The habit of retaining them in the body during
their development must have arisen very gradually in
the phylogenetic history of the salamanders, the period

No. 579] PROGRESSIVE EVOLUTION 165

for which the young were retained growing gradually
longer and longer. It is obvious that this change of
habit involves a corresponding change in the environ-
mental conditions under which the young develop, and
in cases in which the young are not born until they have
reached practically the adult condition this change di-
rectly affects practically the whole ontogeny. We may
say that the series

E, E, E ... En has become

Bi Big Mie css duns
and as the change of environment must produce its ef-
fect upon the developing organism the series

M, M, M; ...° Mn will have become
Ag. Pe Cade Fe

We must remember that throughout the whole course
of phylogenetic evolution this series is constantly length-
ening, so that what was the adult condition at one time
becomes an embryonic stage in future generations, and
the series thus represents not only the ontogeny, but
also, though in a more or less imperfect manner, the
phylogeny of the organism.

The character of each stage in ontogeny must depend
upon (1) the morphological and physiological constitu-
tion of the preceding stage, and (2) the nature of the
environment in which development is taking place. We
can not, however, distinguish sharply between those two
sets of factors, for, in a certain sense, the environment
gradually becomes incorporated in the organism itself
as development proceeds, each part contributing to the
environment of all the remainder, and the influence of
this internal portion of the environment ever becoming
more and more important.

The whole process of evolution depends upon changes
of environment taking place so gradually that the neces-
sary self-adjustment of the organism at every stage is
possible. In the case of our amphibia the eggs could

166 THE AMERICAN NATURALIST [ Von. XLIX

possibly undergo the first stages of development, the
preliminary segmentation, within the oviduct of the
parent just as well as in the water, for in both cases
they would be enclosed in their envelopes, and the
morphological differences between the early stages in
the two cases might be expected to be quite insignificant.
But it must be the same at each term of the series, for
each term is built upon the foundation of the preceding
one, and the whole process takes place by slow and im-
perceptible degrees.

It is true that by the time we reach the formation of
the vestigial gill-slits in the embryo of one of the higher
vertebrates the environmental conditions are very dif-
ferent from those under which gill-slits were developed
in their aquatic ancestors. But what then? Are not
the gill-slits also very different? The changed environ-
ment has had its effect. The gills themselves are never
developed, and the gill-slits never become functional;
moreover, they disappear completely at later stages of
development, when the conditions of life become still
more different and their presence would be actually det-
rimental to their possessor. The embryo with the ves-
tigial gill-slits is, as a whole, perfectly well adapted to
its environment, though the gill-slits themselves have
ceased to be adaptive characters. They still appear be-
cause the environmental conditions, and especially the
internal conditions, which have now become far more
important than the external ones, are still such as to
cause them to do so.

I think the chief difficulty in forming a mental picture
of the manner in which evolution has taken place, and
especially in accounting for the phenomenon of recapitu-
lation in ontogeny, which is merely another aspect of
the same problem, arises from attempting to take in too
much at once. There is no difficulty in understanding
how any particular stage is related to the corresponding
stage in the previous generation, and the whole series of
stages, whether looked at from the ontogenetic or from

No. 579] PROGRESSIVE EVOLUTION 167

the phylogenetic point of view, can be nothing else but
the sum of its successive terms.

It will be convenient, before going further, to sum up
the results at which we have so far arrived from the
point of view of the theory of heredity. We have as yet
seen no reason to distinguish between somatogenic and
blastogenic characters. All the characters of the adult
animal are acquired during ontogeny as the result of
the reaction of the organism to environmental stimuli,
both internal and ‘external. All that the organism ac-
tually inherits is a certain amount of protoplasm—en-
dowed with a certain amount of energy—and a certain
sequence of environmental conditions. In so far as
these are identical in any two successive generations the
final result must be identical also, the child must re-
semble the parent; in so far as they are different the
child will differ from the parent, but the differences in
environment can not be very great without preventing
development altogether.

So far, it is clear, there has been no need to think of
the germ-cells as the bearers of material factors or de-
terminants that are responsible for the appearance of
particular characters in the adult organism; nor yet to
suppose that they are, to use the phraseology of the
mnemic theory of heredity, charged with the memories
of past generations. They have been regarded as simple
protoplasmic units, and the entire ontogeny has appeared
as the necessary result of the reaction between the or-
ganism and its environment at each successive stage of
development. This can not, however, be a complete ex-
planation of ontogeny, for if it were we should expect all
eggs, when allowed to develop under the same conditions
from start to finish, to give rise to the same adult form,
and this we know is not the case. We know also, from
observation and experiment, that the egg is in reality by
no means a simple thing but an extremely complex one,
and that different parts of the egg may be definitely cor-
related with corresponding parts of the adult body. It

168 THE AMERICAN NATURALIST [ Vou. XLIX

has been demonstrated in certain cases that the egg con-
tains special organ-forming substances definitely located
in the cytoplasm, and that if these are removed definite
parts of the organism into which the egg develops will
be missing. We know, also, that the nucleus of the germ-
cell of either sex contains—at any rate, at certain
periods—a number of perfectly well-defined bodies, the
chromosomes, and these also have been definitely cor-
related in certain cases with special features of the adult
organization.

Before we can hope to complete our mental picture of
the manner in which organic evolution has taken place,
if only in outline, it is evident that we must be able to
account for the great complexity of structure which the
germ-cells themselves have managed to acquire, and also
to form some idea of the effect. of this complication upon
the development of both the individual and the race.

We must consider the origin of cytoplasmic and nuclear
complications of the egg separately, for they appear to
be due fundamentally to two totally distinct sets of fac-
tors. In the first place we have to remember that during
oogenesis the egg-cell grows to a relatively large size
by absorbing nutrient material from the body in which
it is enclosed. It is this nutrient material that is used
for building up the deutoplasm or food-yolk. There is
good reason for believing that the character of this
nutrient material will change, during the course of evo-
lution, pari passu with the changing character of the or-
ganism by which it is supplied. Doubtless the changeis of a
chemical nature, for we know from precipitin experiments
that the body fluids of closely allied species, or even of
the two sexes of the same species, do exhibit distinctly
recognizable differences in chemical composition. It
also appears highly probable, if not certain, from such
experiments as those of Agar upon Simocephalus, that
substances taken in with the food, which bring about
conspicuous modifications of bodily structure, may at
the same time be absorbed and stored up by the egg-cells

No. 579] PROGRESSIVE EVOLUTION 169

so as to bring about corresponding changes in the adults
into which the eggs develop.

There seems therefore to be no great difficulty in com-
prehending, at any rate in a general way, how the egg
may become the repository of definite chemical sub-
stances, organ-forming substances if we like to call them
so, possibly to be classed with the hormones and en-
zymes, which will influence the development in a particu-
lar manner as soon as the appropriate conditions arise.

Unfortunately, time will not allow of our following up |
this line of thought on the present occasion, but we may
notice, before passing on, that with the accumulation of
organ-forming substances in the egg we have introduced
the possibility of changes in bodily structure, to what-
ever cause they may be due, being represented by cor-
related modifications in the germ-cells, and this is doubt-
less one of the reasons why the germ-cells of different
animals are not all alike with regard to their potentiali-
ties of development.®

We now come to the question of how the nucleus of the
germ-cell acquired its great complexity of structure. f
We are not concerned here with the origin of the dif-
ferentiation into nucleus and cytoplasm and the respec-
tive parts played by the two in the life of the cell. The
problem which we have to consider is the complication
introduced by the sexual process, by the periodically re-
curring union of the germ-cells in pairs, or, as Weis-
mann has termed it, amphimixis. This is well known
to be essentially a nuclear phenomenon, in which the so-
called chromatin substance is especially concerned, and
it is a phenomenon which must have made its appear-
ance at a very early stage of evolution, for it is exhibited
in essentially the same manner alike in the higher plants
and animals and in unicellular organisms.

Let us suppose, for the sake of argument, that when
amphimixis first took place the chromatin of each germ-

6 Compare Cunningham’s ‘‘ Hormone Theory of Heredity ’’ (Archiv für
Entwicklungsmechanik der Organismen, Bd. XXVI, Heft 3).

170 THE AMERICAN NATURALIST [Vou. XLIX

cell was homogeneous, but that it differed slightly in dif-
ferent germ cells of the same species as a result of ex-
posure to slightly different conditions during its past
history. What would be likely to happen when two dif-
ferent samples of chromatin came together in the zygote?
The result would surely depend upon the interaction of
the complex colloidal multimolecules of which the chro-
atin is composed. Various possibilities would arise.
(1) The two samples might differ in such a way as to
act as poisons to one another, disturbing each other’s
molecular equilibrium to such an extent that neither
could survive. This is possibly what happens when an
ovum is fertilized by a spermatozoon of a distinct
species, though there are, of course, exceptions. (2)
They might be so alike as to be able to amalgamate
more or less completely, so that there would simply be
an increase of chromatin of possibly more or less modi-
fied constitution. (3) They might continue to exist side
by side, each maintaining its own individual character.

In the third case the union of the two different samples
would give rise to a mass of chromatin of twofold na-
ture, and repetition of the process from generation to
generation would, as Weismann has shown, result in
ever-increasing heterogeneity, until the chromatin came
to consist of a great number of different concrete par-
ticles, each of which might conceivably differ from all
the others. But when two heterogeneous masses of
chromatin meet in the zygote there may be all sorts of
mutual attractions and repulsions between the different
colloidal multimolecules, for all three of our supposed
cases may arise simultaneously, and thus the results may
become extremely complicated.

The chromatin of the germ-cells in all existing or-
ganisms is undoubtedly heterogeneous, and this hetero-
geneity may be to some extent visibly expressed in its
arrangement in more or less multiform chr omes
during mitosis. We may provisionally accept Weis- —
mann’s view that these chromosomes are themselves

No. 579] PROGRESSIVE EVOLUTION : 171

heterogeneous, being composed of chromomeres or ids,
which in their turn are composed of determinants.

All this complexity of structure may be attributed to
the effects of oft-repeated amphimixis, a view which is
supported in the most striking manner by the fact that
the nucleus in all ordinary somatic cells (in animals and
in the diploid generation of plants) has a double set of
chromosomes, one derived from the male and the. other
from the female parent, and by the well-known phe-
nomenon of chromatin reduction which always precedes
amphimixis.

When we approach the problem of heredity from the
experimental side we get very strong evidence of the
existence in the germ-plasm of definite material sub-
stances associated with the inheritance of special char-
acters. Mendelian workers generally speak of these
substances as factors, but the conception of factors is
evidently closely akin to that of Weismann’s hypothetical
determinants. The cytological evidence fits in very well
with the view that the factors in question may be definite
material particles and it is quite possible that such par-
ticles may have a specific chemical constitution to which
their effects upon the developing organism are due.

From our point of view the interesting thing is the
possibility that arises through the sexual process of the
permutation and combination of different factors de-
rived from different lines of descent. A germ-cell may
receive additions to its collection of factors or be subject
to subtractions therefrom, and in either case the result-
ing organism may be more or less conspicuously
modified.

By applying the method of experimental hybridization
a most fruitful and apparently inexhaustible field of re-
search has been opened up in this direction, in the de-
velopment of which no one has taken a more active part
than the present President of the British Association.
There can not be the slightest doubt that a vast number
of characters are inherited in what is called the Mende-

172 ` THE AMERICAN NATURALIST [ Vou. XLIX

jan manner, and, as they are capable of being separately
inherited and interchanged with others by hybridization,
we are justified in believing that they are separately
represented in the germ-cells by special factors. Im-
portant as this result is, I believe that at the present
time there exists a distinct danger of exaggerating its
significance. The fact that many new and apparently
permanent combinations of characters may arise through
hybridization, and that the organisms thus produced
have all the attributes of what we call distinct species,
does not justify us in accepting the grotesque view—as it
appears to me—that all species have arisen by crossing,
or even the view that the organism is entirely built up
of separately transmissible ‘‘unit characters.’’
Bateson tells us that

Baur has for example crossed species so unlike as Antirrhinum majus
and molle, forms differing from each other in almost every feature
of organization,

Surely the latter part of this statement can not be
correct, for after all Antirrhinum majus and molle are
both snapdragons, and exhibit all the essential charac-
ters of snapdragons.

I think it is a most significant fact that the only char-
acters which appear to be inherited in Mendelian fashion
are comparatively trivial features of the organism which
must have arisen during the last stages of phylogeny.
This is necessarily the case, for any two organisms suffi-
ciently nearly related to be capable of crossing are iden-
tical as regards the vast majority of their characters.
It is only those few points in which they differ that re-
main to be experimented on. Moreover, the characters
in question appear to be all non-adaptive, having no ob-
vious relation to the environment and no particular value
in the struggle for existence. They are clearly what
Weismann calls blastogenic characters, originating in
the germ-plasm, and are probably identical with the mu-
tations of de Vries. These latter are apparently chro-

No. 579] PROGRESSIVE EVOLUTION 173

matin-determined characters, for, as Dr. Gates has re-
cently shown in the case of nothera, mutation may
result from abnormal distribution of the chromosomes in
the reduction division."

We have next to inquire whether or not the Mendelian
results are really in any way inconsistent with the gen-
eral theory of evolution outlined in the earlier part of
this address. Here we are obviously face to face with
the old dispute between epigenesis and preformation.
The theory of ontogeny which I first put forward is
clearly epigenetic in character, while the theory of unit
gis aeaie represented in the germ-cells by separate
‘“factors,’’ is scarcely less clearly a theory of preforma-
tion, and of course the conception of definite organ-form-
ing substances in the cytoplasm falls under the same
category. The point which I now wish to emphasize is
that the ideas of epigenesis and preformation are not
not inconsistent with one another, and that, as a matter
of fact, ontogenetic development is of a dual nature, an
epigenesis modified by what is essentially preformation.

We have already dealt briefly with the question of
organ-forming substances in the cytoplasm, and it must,
I think, be clear that the existence of these is in no way
incompatible with a fundamental epigenesis. We shall
find directly that the same is true of Mendelian ‘‘fac-
tors’? or Weismannian ‘‘determinants.’’

We have seen that it is possible to conceive of even
a complex organism as inheriting nothing from its parent
but a minute speck of protoplasm, endowed with poten-
tial energy, and a sequence of suitable environments,
the interaction between the two bringing about a similar
result in each suceeding generation, with a slow progres-
sive evolution due to the operation of the law of accu-
mulation of surplus energy. If any of the conditions of
development are changed the result, as manifested in
the organization of the adult, must undergo a corre-
sponding modification. Suppose that the chromatin sub-

7 Quarterly Journal of Microscopical Science, Vol. LIX, p. 557.

174 THE AMERICAN NATURALIST [ Von. XLIX

stance of the zygote is partially modified in molecular
constitution, perhaps by the direct action of the environ-
ment, as appears to happen in the case of Tower’s ex-
periments on mutation in the potato beetle, or by the in-
troduction of a different sample of chromatin from an-
other individual by hybridization. What is the germ-
plasm now going to do? When and how may the
changes that have taken place in its constitution be ex-
pected to manifest themselves in the developing or-
-ganism? !

Let us consider what would be likely to happen in the
first stages of ontogeny. If the germ-plasm had re-
mained unaltered the zygote would have divided into
blastomeres under the stimuli of the same conditions,
both internal and external, as those under which the
corresponding divisions took place in preceding genera-
tions. Is the presence of a number of new colloidal mul-
timolecules in the germ-plasm going to prevent this?
The answer to this question probably depends partly
upon the proportion that the new multimolecules bear
to the whole mass, and partly upon the nature of the
modification that has taken place. If the existence of
the new multimolecules is incompatible with the proper
functional activity of the germ-plasm as a whole there
is an end of the matter. The organism does not. de-
velop. If it is not incompatible we must suppose that
the zygote begins its development as before, but that
sooner or later the modification of the germ-plasm will
manifest itself in the developing organism, in the first
instance as a mutation. In cases of hybridization we
may get a mixture in varying degrees of the distinguish-
ing characters of the two parent forms, or we may get
complete dominance of one form over the other in the
hybrid generation, or we may even get some new form,
the result depending on the mutual reactions of the dif-
ferent constituents of the germ-plasm.

The organism into which any zygote develops must be

No. 579] PROGRESSIVE EVOLUTION 175

a composite body deriving its blastogenic characters
from different sources; but this cannot affect its funda-
mental structure, for the two parents must have been
alike in all essential respects or they could not have in-
terbred, and any important differences in the germ-
plasm must be confined to the ‘‘factors’’ for the differen-
tiating characters. The fundamental structure still de-
velops epigenetically on the basis of an essentially simi-
lar germ-plasm and under essentially similar conditions
as in the case of each of the two parents, and there is no
reason to suppose that special ‘‘factors’’ have anything
to do with it.

We thus see how new unit characters may be added
by mutation and interchanged by hybridization while the
fundamental constitution of the organism remains the
same and the epigenetic course of development is not
seriously affected. All characters that arise in this way
must be regarded, from the point of view of the or-
ganism, as chance characters due to chance modifications
of the germ-plasm, and they appear to have compara-
tively little influence upon the course of evolution.

One of the most remarkable features of organic evo-
lution is that it results in the adaptation of the organism
to its environment, and for this adaptation mutation and
hybridization utterly fail to account. Of course the ar-
gument of natural selection is called in to get over this
difficulty. Those organisms which happen to exhibit
favorable mutations will survive and hand on their ad-
vantages to the next generation, and so on. It has fre-
quently been pointed out that this is not sufficient. Mu-
tations occur in all directions, and the chances of a favor-
able one arising are extremely remote. Something more
is wanted, and this something, it appears to me, is to be
found in the direct response of the organism to environ-
mental stimuli at all stages of development, whereby in-
dividual adaptation is secured, and this individual adap-
tation must arise again and again in each succeeding

176 THE AMERICAN NATURALIST — [Vou. XLIX

generation. Moreover, the adaptation must, as I pointed
out before, tend to be progressive, for each successive
generation builds upon a foundation of accumulated ex-
perience and has a better start than its predecessors.

Of course natural selection plays its part, as it must
in all cases, even in the organic world, and I believe
that in many cases—as, for example, in protective re-
semblance and mimicry—that part has been an extremely
important one. But much more important than natural
selection appears to me what Baldwin® has termed
‘‘ Functional Selection,” selection by the organism itself,
out of a number of possible reactions, of just those that
are required to meet any emergency. As Baldwin puts
it, ‘‘It is the organism which secures from all its over-
produced movements those which are adaptive and bene-
ficial.’ Natural selection is here replaced by intelligent
selection, for I think we must agree with Jennings® that
we can not make a distinction between the higher and
the lower organisms in this respect, and that all purposive
reactions, or adjustments, are essentially intelligent.

Surely that much-abused philosopher, Lamarck, was
not far from the truth when he said, ‘‘The production
of a new organ in an animal body results from a new
requirement which continues to make itself felt, and
from a new movement which this requirement begets
` and maintains.’ Is not this merely another way of
saying that the individual makes adaptive responses to
environmental stimuli? Where so many people fall foul
of Lamarck is with regard to his belief in the inheri-
tance of acquired characters. But in speaking of ac-
quired characters Lamarck did not refer to such modifi-
cations as mutilations; he was obviously talking of the
gradual self-adjustment of the organism to its environ-
ment,

8** Development and Evolution °’? (New York, 1902), p. 87.

9‘* Behavior of the Lower Organisms °? (New York, 1906), pp. 334, 335.

10‘* Histoire naturelle des Animaux sans Vertébres,’’ Tom. I, 1815, P.
185.

No. 579] PROGRESSIVE EVOLUTION 177

We are told, of course, that such adjustments will only
be preserved so long as the environmental stimuli by
which they were originally called for continue to exer-
cise their influence. Those who raise this objection are
apt to forget that this is exactly what happens in evolu-
tion, and that the sine qua non of development is the
proper maintenance of the appropriate environment,
both internal and external. Natural selection sees to it
that the proper conditions are maintained within very
narrow limits.

A great deal of the confusion that has arisen with re-
gard to the question of the inheritance of acquired char-
acters is undoubtedly due to the quite unjustifiable limi-
tation of the idea of ‘‘inheritance’’ to which we have ac-
customed ourselves. The inheritance of the environ-
ment is, as I have already said, just as important as the
inheritance of the material foundation of the body, and
whether or not a newly acquired character will be in-
herited must depend, usually at any rate, upon whether
or not the conditions under which it arose are inherited.
It is the fashion nowadays to attach very little impor-
ance to somatogenic characters in discussing the problem
of evolution. The whole fundamental structure of the
body must, however, according to the epigenetic view,
be due to the gradual accumulation of characters that
arise as the result of the reactions of the organism to its
environment, and are therefore somatogenic, at any rate
in the first instance, though there is reason to believe
that some of them may find expression in the germ-cells
in the formation of organ-forming substances, and pos-
sibly in other ways. Blastogenic characters which ac-
tually originate in the germ-cells appear to be of quite
Secondary importance.

We still have to consider the question, How is it that
organic evolution has led to the formation of those more
or less well-marked groups of organisms which we call
species? We have to note in the first place that there

178 THE AMERICAN NATURALIST [Vou. XLIX

is no unanimity of opinion amongst biologists as to what
a species is. Lamarck insisted that nature recognizes
no such things as species, and a great many people at
the present day are, I think, still of the same opinion.
In practise, however, every naturalist knows that there
are natural groups to which the vast majority of indi-
viduals can be assigned without any serious difficulty.
Charles Darwin maintained that such groups arose,
under the influence of natural selection, through gradual
divergent evolution and the extinction of intermediate
forms. To-day we are told by de Vries that species
originate as mutations which propagate themselves with-
out alteration for a longer or shorter period, and by
Lotsy that species originate by crossing of more or less
distinct forms, though this latter theory leaves quite un-
solved the problem of where the original forms that
crossed with one another came from.

I think a little reflection will convince us that the origin
of species is a different problem from that of the cause
of progressive evolution. We can scarcely doubt, how-
ever, that Darwin was right in attributing prime im-
portance to divergent evolution and the disappearance
of connecting links. It is obvious that this process must
give rise to more or less sharply separated groups of in-
dividuals to which the term species may be applied, and
that the differences between these species must be at-
tributed ultimately to differences in the response of the
organism to differing conditions of the environment. It
may be urged that inasmuch as different species are
often found living side by side under identical conditions
the differences between them can not have arisen in this
way, but we may be quite certain that if we knew enough
of their past history we should find that their ancestors
had not always lived under identical conditions.

The case of flightless birds on oceanic islands is par-
ticularly instructive in this connection. The only satis-
factory way of explaining the existence of such birds is

No. 579] PROGRESSIVE EVOLUTION 179

by supposing that their ancestors had well-developed
wings, by the aid of which they made their way to the
islands from some continental area. The conditions of
the new environment led to the gradual disuse and con-
sequent degeneration of the wings until they either be-
came useless for flight or, in the case of the moas, com-
pletely disappeared. It would be absurd to maintain
that any of the existing flightless birds are specifically
identical with the ancestral flying forms from which
they are descended, and it would, it appears to me, be
equally absurd to suppose that the flightless species arose
by mutation or by crossing, the same result being pro-
duced over and over again on different islands and in
different groups of birds. This is clearly a case where
the environment has determined the direction of evo-
lution.

In such cases there is not the slightest ground for be- .
lieving that crossing has had anything whatever to do
with the origin of the different groups to which the term
species is applied; indeed, the study of island faunas in
general indicates very clearly that the prevention of
crossing, by isolation, has been one of the chief factors
in the divergence of lines of descent and the consequent
multiplication of species, and Romanes clearly showed
that even within the same geographical area an identical
result may be produced by mutual sterility, which is the
cause, rather than the result, of specific distinction.

Species, then, may clearly arise by divergent evolu-
tion under changing conditions of the environment, and
may become separated from one another by the extinc-
tion of intermediate forms. The environmental stimuli
(including, of course, the body as part of its own en-
vironment) may, however, act in two different ways:
(1) Upon the body itself, at any stage of its development,
tending to cause adaptation by individual selection of the
most appropriate response; and (2) upon the germ-
plasm, causing mutations or sudden changes, sports, in

180 THE AMERICAN NATURALIST [ Von. XLIX

fact, which appear to have no direct relation whatever
to the well-being of the organism in which they appear,
but to be purely accidental. Such mutations are, of
course, inherited, and, inasmuch as the great majority of
specific characters appear to have no adaptive signifi-
cance, it seems likely that mutation has had a great deal
to do with the origin of species, though it may have had
very little to do with progressive evolution.

Similarly with regard to hybridization, we know that
vast numbers of distinct forms, that breed true, may be
produced in this way, but they are simply due to recom-
binations of mutational characters in the process of am-
phimixis, and have very little bearing upon the problem
of evolution. If we like to call the new groups of indi-
viduals that originate thus ‘‘species,’’ well and good, but
it only means that we give that name, as a matter of
convenience, to any group of closely related individuals
which are distinguished by recognizable characters from
the individuals of all other groups, and which hand on
those characters to their descendants so long as the con-
ditions remain the same. This, perhaps, is what we
should do, and just as we have learned to regard indi-
viduals as the temporary offspring of a continuous
stream of germ-plasm, so we must regard species as the
somewhat more permanent but nevertheless temporary
offshoots of a continuous line of progressive evolution.
Individuals are to species what the germ-plasm is to in-
dividuals. One species does not arise from another
species, but from certain individuals in that species,
and when all the individuals become so specialized as to
lose their power of adaptation, then changes in the en-
vironment may result in the extinction of that line of
descent.

It is scarcely necessary to point out that no explana-
tion that we are able to give regarding the causes of
either phylogenetic or ontogenetic evolution can be com-
plete and exhaustive. Science can never hope to get to

No. 579] PROGRESSIVE EVOLUTION 181

the bottom of things in any department of knowledge;
there is always something remaining beyond our reach.
If we are asked why an organism chooses the most ap-
propriate response to any particular stimulus, we may
suggest that this is the response that relieves it from
further stimulation, but we cannot say how it learns to
choose that response at once in preference to all others.
If we are asked to account for some particular muta-
tion, we may say that it is due to some modification in
the constitution or distribution of the chromosomes in
the germ-cells, but even if we knew exactly what that
modification was, and could express it in chemical terms,
we could not really say why it produces its particular
result and no other, any more than the chemist can say
why the combination of two gases that he calls oxygen
and hydrogen gives rise to a liquid that he calls water.

There is one group of ontogenetic phenomena in par-
ticular that seems to defy all attempts at mechanistic
interpretation. I refer to the phenomena of restitution,
the power which an organism possesses of restoring
the normal condition of the body after it has been vio-
lently disturbed by some external agent. The fact that
a newt is able to regenerate its limbs over and over
again after they have been removed, or that an echino-
derm blastula may be cut in half and each half give rise
to a perfect larva, is one of the most surprising things
in the domain of biological science. We can not, at
present, at any rate, give any satisfactory mechanistic
explanation of these facts, and to attribute them to the
action of some hypothetical entelechy, after the manner
of Professor Hans Driesch, is simply an admission of
our inability to do so. We can only say that in the
course of its evolution each organism acquires an indi-
viduality or wholeness of its own, and that one of the
fundamental properties of living organisms is to main-
tain that individuality. They are able to do this in a
variety of ways, and can sometimes even replace a lost

182 THE AMERICAN NATURALIST [ Vou. XLIX

organ out of material quite different from that from
which the organ in question is normally developed, as
in the case of the regeneration of the lens of the eye
from the iris in the newt. That there must be some
mechanism involved in such cases is, of course, self-evi-
dent, and we know that that mechanism may sometimes
go wrong and produce monstrous and unworkable re-
sults; but it is, I think, equally evident that the organism
must possess some power of directing the course of
events, so as generally to secure the appropriate result;
and it is just this power of directing chemical and phys-
ical processes, and thus employing them in its own inter-
ests, that distinguishes a living organism from an inani-
mate object.

In conclusion I ought, perhaps, to apologize for the
somewhat dogmatic tone of my remarks. I must ask
you to believe, however, that this does not arise from
any desire on my part to dogmatize, but merely from the
necessity of compressing what I wished to say into a
totally inadequate space. Many years of patient work
are still needed before we can hope to solve, even ap-
proximately, the problem of organic evolution, but it
seemed to me permissible, on the present occasion, to
indulge in a general survey of the situation, and see how
far it might be possible to reconcile conflicting views
and bring together a number of ideas derived from many
sources in one consistent theory.

SHORTER ARTICLES AND DISCUSSION

THE ORIGIN. OF A NEW EYE-COLOR IN DROSOPHILA
REPLETA AND ITS BEHAVIOR IN HEREDITY

‘

In September, 1913, a new eye-color ‘‘scarlet,’’ appeared in
one of my cultures of Drosophila repleta Wollaston. The new
eye color is a bright scarlet when first hatched and darkens but
little with age. The eyes of the wild flies, on the other hand, are
a deep mahogany which darken soon after hatching until they are
almost black. This last statement is true of the stocks I have
found in New York City, Woods Hole, Mass., North Manchester,
Ind., Brazil, Ind., and Terre Haute, Indiana. The eye-color of
the newly emerged mutant corresponds to the color chart in
Ridgeway’s Color Guide, Plate VII, No. 11 (Boston, 1886). The
large scarlet eye in contrast to the dark body of the fly makes
the new repleta an object of great beauty as contrasted with the
wild species.

The new fly in all probability came from heterozygous stock, as
is shown by the following facts. The original stock was obtained
by exposing a fruit jar with banana in a fruit store in North
Manchester, Indiana, September 10, 1913. From this bottle!
there hatched 777 92 and 206 ¢¢ of Drosophila ampelophila.
On November 5 appeared repletas. November 15, I found one
scarlet female among 35 repletas. November 16, one scarlet
male among 20 flies. November 17, one scarlet female among
25 flies. Some of the virgin flies were isolated and four scarlets
appeared on January 24. My assistant, Mr. Powell, also isolated
some of the original stock and later found three scarlets. This

would seem to show that the stock had mutated some time before
‘being taken into captivity. During September, 1915, I set a
great many traps in the region where the above stock was taken,

1I should call attention to the aberrant sex ratio found here in Droso-
phila ampelophila. Culture from this stock later gave 491 99 and 45 gg.
I have data on the sex-ratio in this species for over three years and in
many different stocks. With this exception I have found it approaching
equality. I mated 25 pairs of virgin flies from this stock with the expecta-
tion of finding a sex-linked lethal but in each of the twenty-five bottles the
sex-ratio was practically one of equality. The subsequent history of the
stock was not followed, owing to an accident.

183

184 THE AMERICAN NATURALIST [ Vou. XLIX

with the hope of finding whether or not scarlet was common in
this region. I have bred many of the stocks since that time, but
so far no scarlets have appeared.?

BEHAVIOR OF SCARLET IN HEREDITY

One of the original virgin scarlet females was mated to a
scarlet male. The union was fruitful and a pure scarlet race
was produced which has bred true since that time. The sexes are
easily distinguished, the life cycle is about thirty days, and after
long experience I have found it comparatively easy to breed
this fly in captivity.

Scarlet was crossed to a wild stock which had been taken about
four months previously in Terre Haute. This stock bred true
to black eyes. The flies were studied in mass culture and virgin
flies were used in crossing (the sexes were separated every 18
hours). The offspring, which had eyes like the wild stock, were
mated in mass culture for the F, generation. The apse
tables give the results from the crosses.

TABLE I
SHOWING THE RESULT IN THE F, GENERATION OF CROSSING SCARLET
ox Wop g
N Scarlet | Scarlet Black Black Total Total Total Total
-~ Cea 29 ad 99 ad 99 Scarlet | Black
1 57 64 139 159 196 223 121 298
2 42 27 126 85 168 112 69 211
3 73 31 148 132 221 163 104
4 61 61 210 166 271 122 376
5 72 73 263 193 266 145
6 ss RE ai nee _ — 175 530
7 pts coe sis ase le —_ 52 182
Total, 305 256 886 735 1,191 991 788 2,333

These tables bring out the fact that the new eye color is a
simple Mendelian recessive character since it approximates the

2 It is only fair to state that I had made earlier attempts to find muta-
tions in this species. In the fall of 1911 a female of D. repleta was taken
in the Zoological laboratory at Columbia University and from this a stock
was obtained which was kept going on well-ripened bananas with more or
less difficulty for more than a year. It was comparatively easy to keep the
colony going in the same bottle by adding food from time to time but
difficulty was experienced in founding new colonies. During the period of
observation I examined many hundreds of repletas without finding a single

mutation. ms ;

No.579] SHORTER ARTICLES AND DISCUSSION 185

TABLE II
SHOWING THE RESULT IN THE F, GENERATION OF CROSSING SCARLET
XxX WILD 2
N Scarlet Scarlet Black Black Total Total Total | Total
2 ad 29 ley ge PF ge Scarlet | Black
8 34 37 111 118 145 155 71 | 229
9 31 56 134 165 165 221 87 299
10 38 46 121 137 159 183 | 258
11 ai aon are S 102 300
12 69 91 216 257 285 348 160 473
13 22 19 39 61 61 80 41 141
14 — — — — — — 80 264
15 oe bas wad ee 25 — 74 230
Totall 194 | 249 | 621 | 738 | 815 987 | 699 | 2,194

expected ratio of three to one. There appeared in the F, gen-
eration from the scarlet male a total of 699 scarlets and 2,194
blacks,—a ratio of 3.14 black to one scarlet. From the scarlet
female there appeared in the F, generation 788 scarlets and
2,333 blacks,—a ratio of 2.96 black to one scarlet. It is to be
noted that the sex ratio is practically one of equality.

Roscoe R. HYDE

A WING MUTATION IN A NEW SPECIES OF
DROSOPHILA

A NEW wing mutation which appeared in my cultures of
Drosophila confusa Auct. (not Staeg.) is characterized by the
fact that the wings curve upward at an angle of about 45 degrees
from the region of the tip of the abdomen. The new wing re-
sembles somewhat the shape of a petal of the rose and is easily
distinguished from the wild species since the wings of the wild
fly project horizontally over and beyond the abdomen, as is
characteristic of the diptera. I shall refer to the new fly as
: jaunty C.2

The wild stock from which jaunty C arose was taken in an
orchard on the Coss farm about seven miles south of North
Manchester, Indiana, in September, 1913. The original stock
was bred in a glass vial to which fresh banana was added from
time to time. Several stock bottles were made up from this

1 The wing is like that of jaunty in D. ampelophila and is here designated
jaunty C( = confusa) to call attention to this resemblance.

186 THE AMERICAN NATURALIST [Vou. XLIX

bottle. All the offspring were examined with a hand lens but no
unusual forms appeared until the fourth or fifth generation when
jaunty C was discovered. Subsequently three or four similar
mutants were found in the cultures, which would seem to indi-
cate that they arose from heterozygous stock. Pure stock was
obtained by crossing to the wild flies and ‘‘extracting.”’

When jaunty C is crossed to the wild type all of the flies of the
F, generation have long wings. No exact record was kept but
this statement is true of several hundred that were observed.
The sex ratio was practically one of equality. In the F, genera-
tion jaunty C reappeared, as shown in the following tables.

F, GENERATION FROM JAUNTY C Jf F, GENERATION FROM JAUNTY C 9

TABLE I . TABLE II
No. Jaunty C Long No. Jaunty C Long
1 40 176 3 37 124
2 38 150 4 24 145
5 66 308
Totalc?. 78 326 127 577

Among the grandchildren from the jaunty C male the ratio is
one jaunty C to 4.18 long, while among the grandchildren of the
reciprocal cross the ratio is one jaunty C to 4.54 long. The
sex ratios were near equality.

These ratios do not conform very closely to Mendelian ex-
pectations, but I have found this species very hard to breed, and
since the flies were bred in mass cultures it may be that jaunty C
was repped affected by crowding of the larve.

to carry out more elaborate experiments during
the summer of 1914 and had about twenty bottles of the new
stock in pure culture and also some wild stocks, when the flies
commenced to die during the hot days in the latter part of May
and June. Finally the last individual disappeared despite all
the care that I could exercise, and no larve were left in the
bottles to take their place. As the June temperature increased
other stocks failed to reproduce and died out. That the warm
weather was in all probability responsible is shown by the results
which were obtained by placing the stocks in a refrigerator. All
those stocks placed in the refrigerator remained very active and
continued to reproduce while all the stocks left on the outside
died out with the exception of the wild stocks of D. ampelophila.

No.579] SHORTER ARTICLES AND DISCUSSION 187

But even ampelophila does not thrive when the temperature
reaches 100°.

During September, 1914, I took several wild stocks of confusa
from the same region, and have examined many of the offspring
with the hopes of again finding this form but so far no unusual
forms have appeared.

Roscoe R. HYDE

MUTATIONS IN TWO SPECIES OF DROSOPHILA

In our cultures of Drosophila, mutations have appeared re-
cently in two species other than Drosophila ampelophila. Both
mutants are characterized by abnormalities in wing venation.
One of them has irregular extra veins in the axillary cell, and
hence may be called axillary. The other is distinguished most
clearly by the fusion of the distal end of the second vein to the
costa, producing a double vein for a considerable distance, for
which reason it is called confluent. In each of these cases other
abnormal characters are associated with those mentioned, but
they are relatively inconspicuous.

The mutant called axillary arose in normal stock of D. tri-
punctata Loew, which has been bred in the laboratory for about
six generations. This stock was kept in milk bottles and fed
on banana, but received no artificial treatment except anesthesia
with ether once per generation. Axillary behaves as a simple
Mendelian recessive when crossed with normal, and breeds true
in pure cultures.

The mutant called confluent appeared in a culture of an un-
described species of Drosophila, referred to as ‘‘species B” by
one of us in a paper describing its chromosomes.’ Confluent is
a dominant character (i. e., it appears in the heterozygous fly),
and so far as we have been able to ascertain it never occurs in
the homozygous condition. At least no flies homozygous for it
have as yet been found, although numerous matings have been
made which should have produced them. The original fly show-
ing the confluent character (a male) appeared in a stock culture,
all of his brothers and sisters being normal. He was hetero-
zygous, as shown by matings with normal females, which gave
15 normal and 13 confluent offspring. Seven of the latter, bred

1¢Chromosome Studies in the Diptera,’’ I, Jour. Exp. Zool, XVII. p.
45, 1914.

188 THE AMERICAN NATURALIST [ Von. XLIX

to normals in pairs, gave 778 normals and 691 confluents, show-
ing that they too were heterozygous.? The remaining six were
bred together in pairs and gave 261 normal and 431. confluent
progeny, or a ratio of approximately 1:2 instead of the expected
1:3. According to expectation one third of the 431 confluent
offspring in this generation should be homozygous, and random
matings in pairs (confluent by confluent), should give in five
cases out of nine only confluent progeny. Sixteen such matings
have been made, none of which gave this result; instead each gave
approximately one normal to two confluent, just as did the F,
heterozygotes. Normal brothers and sisters of confluent in both
generations bred en masse gave only normals, showing that none
of them was heterozygous for confluent. From these data we
conclude that the homozygous confluent flies are not viable, and
that the 1:2 ratio is due to the total absence of this class. To
our knowledge such a condition as this has been previously re-
corded in only three cases: the ‘‘aurea’’ Antirrhinum of Baur,
the yellow mouse of Cuenot, Castle, ete., and the dwarf wheat of
Vilmorin. Baur’s case differs somewhat from the others and
from ours in that the homozygous mutant class appears, but
soon dies (due to the absence of chlorophyll).

With regard to the origin of mutations the present cases are
instructive in showing that they may appear without the use
of artificial chemical or physical agents, and without hybridiza-
tion. No radium, X-rays or any chemicals whatever have been
applied to these cultures, except ether, and that only for anes-
thesia of the adult flies in each generation. The stock of D.
tripunctata from which axillary arose was obtained wild, and
had been inbred for six or seven generations; that of the other
species, from which confluent arose, is all descended from one
pair of wild flies, almost certainly brother and sister, and had —
been inbred for about twelve generations when the mutant ap-
peared. In neither case had flies from two localities been crossed ;
both stocks were pure and inbred. The only agent that could
possibly fall under suspicion as a causative one, then, is ether,
but this was used uniformly throughout the experiments, and
since only two mutations appeared among many thousands 0
flies, there is no reason for attributing them to the specific
-~ 2The offspring per pair were respectively: 242 : 194, 93 : 73, 97 : 106,

42 : 47, 133 : 125, 76 : 65, 110: 94.

No.579] SHORTER ARTICLES AND DISCUSSION 189

effect of ether ;? a conclusion made even more certain by the fact
that other species were bred during the same time, under iden- -
tical conditions, and with the same treatment, but without the
production of mutations. There is every reason to believe,
therefore, that the cause of the mutation in each case was piiely
fortuitous.

One of the aims of our work on the Drosophilas is to apply
the chromosome hypothesis to species having chromosomes dif-
ferent from those of D. ampelophila. The experimental work of
Morgan and others on D. ampelophila has pointed directly to the
conclusion that the four groups of linked factors which they have
studied are located, respectively, in the four pairs of chromo-
somes of this species. One of us has recently shown in the paper
above cited that several other species of Drosophila have chromo-
some groups differing from that of ampelophila in the number
and relative sizes of the chromosomes. Of the two species con-
sidered in the present paper, one, ‘‘species B,’’ has six pairs of
chromosomes, and should therefore, on the chromosome
hypothesis, give six series of linked characters. The other, D.
tripunctata, has four pairs of chromosomes, but of a type
essentially different from that of ampelophila, and consequently
should also give essentially different linkage series.

It is significant that both of the mutations which we have
found (axillary and confluent), are represented by similar muta-
tions in D. ampelophila. Judging from these it is not too much
to expect that among other mutations which may subsequently
arise in our species, some will likewise correspond to some of
those in ampelophila, and that upon this basis it may be possible
to homologize linkage groups, and thus more definitely homolo-
gize chromosomes in different species.

C. W. Merz anv B. S. Merz

CARNEGIE INSTITUTION,

STATION FOR EXPERIMENTAL EVOLUTION

A SEX-LINKED CHARACTER IN DROSOPHILA
REPLETA

` Drosophila repleta Wollaston (D. punctulata Loew) is a
cosmopolitan species, though only recently introduced into the

3 Professor grae 1c arrived at the same conclusion with regard to
the appearance of m in Drosophila ampelophila, Cf, AMER, NAT.,
1914, ‘‘The Failure of F his to Produce Mutations in Drosophila. ”?

190 THE AMERICAN NATURALIST [Vou. XLIX

greater part of this country. The color of the thorax (dorsal
side), in most specimens, is light gray, each hair having a dark
blackish brown spot at its, base. These spots are somewhat
irregular, and coalesce in certain regions,

In October, 1914, I collected a number of specimens of D.
repleta in the zoological laboratory at Columbia University.
About one sixth of these had a lighter color on the thorax than
that found in normal flies. The dark spots, while of about the
same number and color as usual, were much smaller and only
coalesced in a few small regions. Several females of both kinds
were isolated and their offspring observed. These females were,
in each case, mated with males of their own kind: but they were
of unknown age when captured, and several of them had prob-
ably already mated with other males. In the tables given here
‘*dark’’ refers to the normal type; ‘‘light,’’ to the new character.

TABLE I
WILD FEMALES
Offspring
Culture Mother
Dark 9 Dark g Light 9 Light #
PA Light 5 0 91 86
Q Light 12 0 50 53
ig Dark 62 76 0
U Dark 71 52 11 41
y Dark 96 50 0 41
wW Dark 36 30 0 0
X Dark 32 47 0 0

Light offspring from J and from Q, when mated together,
gave 166 lights in the next generation—no darks. Darks from
T, mated together, gave 180 darks—no lights.

On the basis of these results it is probable that the light char-
acter is a sex-linked recessive. The two light females, J and Q,
had paired with dark males before being captured, since they pro-
duced a total of 17 dark offspring: but these darks were all
females, showing either that the male-producing sperm of the
father carried no dark factor (i. e., that the factor is sex-linked),
or that the light character is dominant in the males and recessive
in the females,

Female V, since she produced light sons but no light daughters,
must, on either of the above views, have been mated only by a
dark male, and she must have been heterozygous for the light

No.579] SHORTER ARTICLES AND DISCUSSION 191

character. Female U must have had the same constitution, but
had probably mated with both kinds of males.

The crucial test between the two views was furnished by
mating a dark female from culture T to a light male from J.
The result was 25 dark females and 26 dark males. This is the
expectation if the character is sex-linked; but if light is recessive
in the females and dominant in the males, the mating should
have given only dark females and light males. The light char-
acter is, therefore, sex-linked and recessive.

A further test was made by mating heterozygous females (one
from Q and one from U) by their light brothers. Table II
shows that the result approximates to the expected 1:1:1:1
ratio.

TABLE II
Culture Dark 9 Dark of Light 9 Light #7
Q2 12 12 19 15
U1 15 17 17 13
27 29 36 28

In all the cultures it has been observed that the heterozygous
females average a little lighter in color than do the homozygous
darks. This difference, however, is not sufficient to allow an
accurate separation of the two classes. Dark males are of the
same color as the homozygous dark females.

In October, 1914, I received some banana collected by Mr. B.
Schwartz at Fayetteville, Ark. From it there hatched one
repleta male, which was of the light type. Bred to light females
from culture J, this male produced 133 offspring, all of which
were light.

An examination of the pinned material in my own collection
and that of the American Museum of Natural History has shown
the existence of a number of specimens which seem to belong to
the light type. The following table shows the distribution of the
specimens examined. Those marked ‘‘not workable,’’ are not
in good enough condition to be classified with certainty.

The table shows the light form to oceur in New York, Alabama,
Arkansas, California, and Cuba. The Cuban record is of in-
terest because the date, 1904, is the earliest of the seven cases.

192 THE AMERICAN NATURALIST [ Von. XLIX

TABLE III
Locality Date Dark Light Not Workable
Woods Hole, Mass........ June, 1913 3 0
New: York; N: Ya cctv: Feb., 1913 2 1
eee June, 1913 3 1
fe eeu EA ct., 19 83% 17%
Washington, D. C........ ct., 1912 3 0
N. Manchester, Ind....... Sept., 1913 2 0
LIRGLOBR, Fk. cs ap oe ees Mar., 1914 1 0
Gene. Als o Oi ewe ie June, 1914 3 1
Pavestevilie, FP Oct., 1914 0 1
Claremont, Calif......... May, 1914 4 0
ewport, Califs,:. 6.0062: Sept., 1913 2 6 1
Berkeley, Calif........... 191 2 0
ear ana, Cuba...... Nov., 1904 3 3 3
Guantanamo, Cuba....... Dec., 1913 5 0 1
u, Dominica........ June, 1911 5 0 1

At that time D. repleta seems to have been rather rare in the
United States.
A. H. STURTEVANT
COLUMBIA UNIVERSITY,
January 1915

%
VOL. XLIX, NO. 580 |

The American

Naturalist

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

MSS
sent to the beatae of THE AM
Short icles containin

MERICAN NATURALIST, Ga
summaries of

ew York,

roblems of organi evolution are especially welcome, and will be given preference

in publication

One hun reprints of contributions are supplied to authors free of charge.
Further reprints will be supplied at cost.

Subscri and eghtncmcagesce a beste be oer to the — The
rodeo eg price is oe dollars en F n postage is fifty cents and

five cents

ees nal. The charge for single eri is

n postage
forty sents. The vorili rates are pas Dollars for a page.

THE SCIENCE PRESS

Garrison, N. Y.

Lancaster, Pa.
NEW YORK: Sub-Station 84
Entered d-class matter, April 2, 1908, at the Post ee. at Lancaster, Pa., under the Act of
Congress of March 3, 18

: R SALE
ARCTIC, ICELAND and GREENLAND

RDS’ SKINS,
Well Prepared Low Prices
rs of

_ DINESEN, Bird Collector

ence solicited.

JAPAN NATURAL HISTORY SPECIMENS .
Perfect Condition and Lowest Prices.
Specialty: Bird Skins, Oology, Entomology, Marine
Animals and others. Catalogue free. Correspond-

T. FUKAI, Naturalist, ;
Konosu, Saitama, Japan

a : Husavik, North iceland, Via Leidle, Eapiang

Marine Biological Laboratory 7 |

Woods Hole, Mass. .
"INVESTIGATION pnas aa
Entire Year tae rete .

THE
AMERICAN NATURALIST

Vor. XLIX April, 1915 No. 580

ORIGIN OF SINGLE CHARACTERS AS OBSERVED
IN FOSSIL AND LIVING ANIMALS
AND PLANTS!

HENRY FAIRFIELD OSBORN
COLUMBIA UNIVERSITY
AMERICAN Museum or NATURAL HISTORY

In the last thirty years two biologies have been develop-
ing. The first is the biology of the garden, the seed pan,
the incubator, and the breeding pen. The second is the biol-
ogy of the field zoologist, of the field botanist, of the pale-
ontologist. Inasmuch as one regards unnatural processes
and the other regards natural processes it is small wonder
that these biologies have become as far apart as two re-
ligions and have developed their sects and their dog-
matists. Yet the actual facts assembled in these two
biologies as distinguished from the opinions based there-
upon can not be in the least discordant, for certainly
there is only one system of law operating in the living
world and there can be only one ultimate and final biology.
In my Harvey lecture of 1912? the search for some unity
between the observations in these two great fields of `
natural and experimental research met with some failure

1 Presidential address before The Paleontological Society of America,
delivered in the Academy of Natural Sciences of Philadelphia, Wednesday,
December 31, 1914.

2 The present address, as a comparison of zoological, paleontological, and
experimental results, is a sequel to the author’s Harvey Lecture of 1912,
entitled ‘‘The Continuous Origin of Certain Unit Characters as Observed
by a Paleontologist.’’ Harvey Soc. Vol, 7th ser., Nov., 1912, pp. 153-204.
It employs in part the same materials and illustrations.

193

194 THE AMERICAN NATURALIST [Vou. XLIX

and some success, and in the present address I am push-
ing inquiry along the same line, choosing the ‘‘single
character’’ as the point of investigation and comparison.

THE ORIGIN oF CHARACTERS

The old and ever vague problem of the origin of species
is being resolved into the newer and more definite prob-
lem of the origin of characters; in the dim future when
we know how and why new characters originate, and how
and why they transform and disappear, the problem of
natural and experimental research met with some failure
and some success, and in the present address I am push-
ing inquiry along the same line, choosing the ‘‘single
character’’ as the point of investigation and comparison.

THe ORIGIN OF CHARACTERS

The old and ever vague problem of the origin of species
is being resolved into the newer and more definite prob-
lem of the origin of characters; in the dim future when
we know how and why new characters originate, and how
and why they transform and disappear, the problem of
„Species will have long been solved and well-nigh for-
gotten. This is because a species is an assemblage or
colony of similar individuals, each individual is composed
of a vast number of somewhat similar new or old charac-
ters, each character has its independent and separate
history, each character is in a certain stage of evolution,
each character is correlated with the other characters of
the individual.

Thus in a sense the species, the subspecies, the variety,
even the individual is not a zoological unit, whereas the
‘‘character’’ when narrowed down to the last point of
divisibility seems to be a unit both among plants and ani-
mals, and a very stable one, with certain distinctive
powers, properties, and qualities of its own. We have
been approaching this new conception from many dif-
ferent lines of observation among fossil and living ani-
mals and plants, and a preliminary survey of results is _
opportune.

No. 580] ORIGIN OF SINGLE CHARACTERS 195

My chief purpose in this address is to show what one
of these ‘‘single’’ or ‘‘least characters’’? is and what pe-
culiar powers and properties it possesses which distin-
guish it from other ‘‘least characters’’ and give it a cer-
tain individuality and separateness.

If you read your Lamarck, your Darwin, your Cope
afresh with this general conception in mind you will find
that throughout biological literature the problem of
species has always been an incidental one, a sort of
by-problem and relic of the very ancient controversy as
to whether species were created suddenly or evolved
gradually. The real problem has always been, that of
the origin and development of characters. Since the
‘‘Origin of Species’? appeared the terms variation and
variability have always referred to single characters; if
a species is said to be variable we mean that a acne.
able number of the single characters or groups of char-
acters of which it is composed are variable. In botany the
long overlooked discovery of Gregor Mendel in 1865 had
as its most essential feature the separability of characters.
in heredity. In paleontology as long ago as 1869 Waagen
sharply focused our attention on single phyletic charac-
ters as of far greater significance and importance than
the matter of local races, varieties, and subspecies. The
modern observers in experimental zoology and heredity
are far less concerned with ‘‘species’’ than with the sep-
arate characters of which the individuals within a species
are composed :

Some naturalists incline to regard the ‘‘character’’ as
observable only by certain methods of their own, but it is
obvious that since all hereditary ‘‘characters’’ are germi-
nal there can be no royal or exclusive road by which we
may observe their origin and transformation, for the ger-
minal and somatic laws controlling the characters of

83 T. H. Morgan has pointed out that the term ‘‘unit character’’ was im-
properly used in my Harvey address. ‘‘Unit character’’ is a germinal
rather than a bodily term. I am treating here of single bodily or mew
characters which may be represented by one or more ‘‘unit characters’? i

Tm. i

196 THE AMERICAN NATURALIST [ Vou. XLIX

the bean,* the fly, the molluse,® the titanothere,” and
man’ are doubtless identical.

Accordingly my second purpose in this address is to show
that there is a certain harmony in the results obtained in
widely different fields of research although some of these
results may appear at first to be entirely unrelated and
even discordant.

In this attempt to discover an underlying harmony let
us first glance at the ‘‘character’’ conception in the older
natural sciences of animals and plants. CHARACTER is
the most frequently used term in the vocabulary of zool-
ogy and botany. It occurs far more often than any other
word. It has been used millions of times in systematic
definition since Linneus. Yet I do not know of any
attempt to clearly define or analyze the meaning of the
word character in its biologic sense while hundreds of
attempts have been made to define the word species.
Here again the greater is involved in the less and when-
ever we shall succeed in clearly defining ‘‘character’’ the
definition of species will follow as an incidental result.

The derivation of the word is from the Greek yapaxryp
properly an instrument for marking or graving; as ap-
plied to a person, an engraver; as commonly used, any
mark engraved or impressed, the impress or stamp on
coins and seals. It passes into the word characteristic,
which means a distinguishing feature.

The use of the word is not only universal among sys-
- tematists and experimentalists of our day, but it has be-
come one of the most elastic words in our language; the
“character”? may be as comprehensive as the general
habit of an entire organism, as where we speak of the
lethargic character of the sloth, or as restricted as a single
minute cuspule on a fossil tooth, or the barely visible
outgrowth on the surface of a fossil shell. The speed of
the race horse is a character, its tractability or viciousness
are characters, the position of the horse’s tail in running

4 Johannsen, 7 Osborn.

5 Morgan. 8 Galton.

6 Waagen, Neumayr, Hyatt, Jackson.

No. 580] ORIGIN OF SINGLE CHARACTERS 197

is a character, the color of the horse’s hair is a character,
the most minute cellular structure of the tissue of the
hoof is a character.

There is an underlying reason why this very elastic
use of the term is absolutely scientific: it is, that every
one of the above diverse applications of the term to animal
or plant life refers to some structure or some quality
which is heritable; heredity is the unifying principle.

The word is again elastic and often confusing in being
used both for germinal characters which are always herit-
able and for bodily modifications of character acquired
through habit or environment which may not be heritable.
When we speak of characters which are not known to be
hereditary we should qualify them as acquired, as modi-
fied, as due to nurture, to habit or ontogeny, to environ-
ment, as somatic rather than as germinal. Thus it is per-
fectly proper to speak of ‘‘ ontogenetic species’’ as Jordan
does, species the bodily characters of which are due to
certain habits; or of ‘‘environmental species’’ the bodily
characters of which are due to peculiarities of environ-
ment. While such modifications by habit and by en-
vironment make up a considerable part of the characters
which distinguish geographic species, subspecies and
races, it is not the origin and the transformation of these
characters which we are now considering, for that prob-
lem is comparatively simple, but rather of those under-
lying germinal and heritable characters the origin and
transformation of which is absolutely an impenetrable
mystery at the present time.

How do we know through zoology, botany, and paleon-
tology as well as through experiment that ‘‘characters’’
are real units of structure with some individual and dis-
tinct qualities and properties of their own which separate
them from all their fellows and at the same time with
certain properties of correlation which unite them with
all their fellows? i

First, we may observe in these living and extinct forms
evidences of two such antithetic principles, a principle

198 THE AMERICAN NATURALIST (Vou. XLIX

of hereditary separability whereby the body is a colony,
a mosaic of single individual and separable characters,
which is combined with a principle of hereditary correla-
tion whereby the body is a complex of minutely related
and interacting units so that functionally and structur-
ally many of these units are linked with others. Neither
principle is simple; on the contrary, both principles are
extraordinarily complex and go back to the very begin-
ning of things. Comparing more closely the observa-
tions on fossil vertebrates and invertebrates, we develop
laws of separability as well as laws of correlation, and
note that certain of these laws are far more clearly per-
‘ceived in some fields of observation than in others.

The biologic value of the field to which our Paleonto-
logical Society is especially devoted lies in the revelation
of certain of these laws and causes of the separability of
characters which are not revealed at all to the zoologist
or to the experimentalist. The paleontologist is in a
position to understand why certain characters fall apart
and become separable in cross breeding, the cause being
connected with their origin and antecedent history.

Of far broader biologic significance is the fact that all
principles which may be discovered through paleontology —
regarding the ‘‘origin of characters’’ in the hard parts,
govern alike characters of the soft parts as well as of
other structures and functions. For there can not be
one principle governing the ‘‘characters’’ of bones, an-
other those of the muscles, another those of nerves; one
principle for structures, another for functions. But
while these principles are unlimited, our comparisons
with zoology, for example, are limited to the origins of
characters which may be observed both in living and
fossil forms, namely, in the skeleton and in the teeth;
and at the outset a convenient and readily understood
distinction may be made between the origins of numerical
and of proportional characters, as follows:

No. 580] ORIGIN OF SINGLE CHARACTERS 199

Numerical Proportional
Presence and absence characters, Changes of form in the length,
e. g., numbers of teeth, of cusps breadth and height of parts.
on the teeth, of vertebra, of Quantitative changes in the hard

toes, of pads on the feet, of parts. Such characters as may
mamme. Meristic or segmental partly be expressed in indices
characters, such as may be and ratios.

partly expressed in formule.

Proportional characters may through prolonged reduc-
tion lead into numerical characters. Thus the reduction
in length of one of the toes may precede the loss of the
toe, which is a numerical change. Yet we shall see that
somewhat different principles prevail in the origins of
certain numerical characters as contrasted with the
origins of proportional characters.

1. Use of Numerical and Proportional ‘‘Characters’’ in

Classification of Mammals
In our attempt to analyze ‘‘characters’’ as they are re-
vealed to the systematic and field zoologist let us take as
two examples, first, ‘‘ The Catalogue of the Mammals of
Western Europe” by Gerrit S. Miller,’ and, second, the
‘Revision of the Mice of the genus Peromyscus,” by
Wilfred H. Osgood. It is of the utmost importance that
mammalogists, whether working among living or fossil
forms, should use similar methods of description and defi-
nition of characters, and we especially welcome in the
monumental work of Miller the fact that the definitions
and the keys are chiefly upon the hard parts which are
also available to the paleontologist. We select as typical
his treatment of the Order Carnivora and of the Family
and Genera of the wolves and foxes, which he distin-

guishes by the following enumeration of characters:

Miller’s Our Analysis of _ Kinds of

Charac

Diagnoses and Definitions
ORDER Carnivora. Characters. Chief habits, oie adaptations

—Terrestrial (rarely aquatic or of the teeth and limbs; chief char-

9 Miller, Gerrit S., ‘‘Catalogue of the Mammals of Western nhs
(Europe exclusive of Russia) in the Collection of the British Museum
London, 1912, 1019 pp.

200

semi-aquatic), non-volant, placen-
tal mammals with rather high de-
velopment of brain. The cerebral
hemispheres with distinct convolu-
unguiculate, never

tion of a modified tubereulo-sec-
torial type, the posterior upper
premolar and anterior lower molar
usually developed as special car-
nassial or flesh-eutting teeth

amily CANID#. Characters.—
Larger cheek-teeth of a combined
trenchant and crushing type, the
last upper premolar and first lower
molar strongly differentiated as
carnassials, the former 3-rooted,
its inner lobe in front of middle
of crown, its position, somewhat
posterior to level of anteorbital
foramen, at point of greatest me-
chanical efficiency; auditory bulla
moderately or considerably in-
flated, without septum; form
rather light, the legs long; size
moderate; feet digitigrade; toes,
5-4 or ;

Genus Canis. Characters.—
Skull heavy and deep (depth of
brain-ease more than one-third
condylobasal length); interorbital
region thickened and elevated, the
frontal sinuses rather large, the
postorbital processes thick, convex
above, their edges rounded off;
dorsal profile of forehead rising
rather abruptly and noticeably
above level of rostrum; dental for-

;
j 83 1-1 4-4 2-2
Sa Sa Pa ae
teeth heavy and large, the length
of carnassial and upper molars
together contained about 214 times

THE AMERICAN NATURALIST

[Vou. XLIX

acters of the brain. Inherited
from “ least characters ” which ac-
cumulated and evolved in Mesozoie
and early Tertiary time.

Characters of proportion and
changes of form; characters of
funetion or adaptation; presence
or absence of certain numerical
characters. i
the ancestors of the
verged as terrestrial and cursorial
Carnivora from other Carnivora.

Chiefly characters of propor-
tion; also numerical characters of
the teeth. Characters clearly mani-
fested in lower and middle Eocene
time and taking on their modern
aspect in early Oligocene time.

No. 580]

in palatal length; canines robust

scarcely beyond middle of mandi-
bular ramus when jaws are closed
(Fig. 65

[Species] CANIS LUPUS. Diag-
nosis.—Condylobasal £
225 mm.) ; cheek-teeth larger than
in the largest races of domestic

ORIGIN OF SINGLE CHARACTERS

201

Chiefly characters of propor-
tion; certain minor numerical
characters. Characters distinguish-
ing Canis lupus from Vulpes, first
apparent in Miocene and E hosie

dogs, the upper carnassial 25 to time.
27 mm. in length, but structure
not peculiar, the upper molars
with narrow, inconspicuous cingu-
lum on outer side (Fig. 61).
[Subspecies] CANIS LUPUS LU-
PUS. Characters.—Size maximum
for the species; general colour not
markedly tawny; white of throat
not extending to cheeks. The few
skulls examined agree with Asiat-
ic specimens in having the outer
cusps of m! moderately large, the
paracone with transverse diameter
of base about equal to width of
large flattened portion of crown.

Size characters ; color charac-

veloped in early Pleistocene time,
perhaps 500,000 years ago.

In the ascending order of Miller’s definitions we note
that ‘‘subspeciesare mainly distinguished hy characters of
proportion and of form and by the degrees and intensities
of color, but rarely if ever by numerical characters. ‘‘Spe-
cies’’ are mainly distinguished by the proportions of the
various hard parts and to a less extent by the presence
and absence of minor ‘‘numerical ’’ characters. ‘‘Genera’’
are distinguished by the proportions, by the presence or
absence of several numerical characters, also by fune-
tional characters such as dental succession. ‘‘Families’’
are distinguished by changes of proportion and of form,
by many numerical characters, such as the presence or
absence of certain parts, by structural adaptations in the
teeth and feet. ‘‘Orders’’ are distinguished by the funda-

202 THE AMERICAN NATURALIST [ Vou. XLIX

mental and very ancient chief habits, chief adaptations in
the hard parts, chief brain features.

Thus we see that two kinds of characters are employed
by Miller throughout, namely: first, characters of pro-
portion of form and of degree; second, numerical or

Fic. 1, Skulls and cheek teeth of the wolf (Lupus), arctic fox (Alopex), and
red fox (Vulpes), illustrating the differences in proportional characters of the
skulls and teeth and the resemblances in the numerical characters, or rectigra-
dations (R).

presence and absence characters. We are struck by the
fact that changes in proportion embrace by far the larger
part, perhaps nine tenths, of the ‘‘characters’’ enumer-
ated by Miller in his systematic descriptions; this is þe-
cause change of proportion is the chief and most universal
phenomenon in the adaptation of mammals to different
habits and habitats. Numerical change is hardly less im-
portant, but is less universal and less frequent.

Similar weight upon the value of characters of propor-
tion is seen in the contrast between Miller’s definitions of
the three genera of dogs, namely: Canis, Alopex, and
Vulpes. Here again the vast majority are characters of
proportion and of form.

No. 580]

I. Genus CANIS

Characters. — Skull
heavy and deep (depth
of brain-case more than
one-third ™ condylobasal
length) ; interorbital re-
gion thickened and ele-
vated, the frontal si-
nuses rather large, the
postorbital processes

ick, convex above,
their edges rounded off;

above level of rostrum;
dental formula:

gk 1 4-4
3-37 So m 4-4’
es

teeth heavy and large,
the length of earnas-
sial and upper mo-
lars together contained
about 2% times in pal-
atal length; canines ro-
bust and not specially
elongated, the point of
upper tooth extending
scarcely beyond middle
of mandibular ramus
when jaws are closed
(Fig. 65).

The fact that changes of proportion

ORIGIN OF SINGLE CHARACTERS

II. Genus ALOPEX

Characters. — Skull
intermediate in general
form between that of
Canis and Vulpes; oecip-
ital depth about one-
third eondylobasal
length; interorbital re-
gion more elevated than
in Vulpes owing to
greater inflation of the
frontal sinuses; postor-
bital processes thin, flat
or slightly concave
above, with bead-like,
overhanging edges; dor-
sal profile of forehead
rising abruptly above
rostrum as in Canis;
teeth moderately heavy
and large, the length of
earnassial and upper
molars together con-
tained about 234 times
in palatal length; ca-
nines and incisors inter-
mediate between those
of Canis and Vulpes
(see Fig. 65); external
form fox-like, but ear
short and rounded, not
conspicuously overtop-

the

surrounding ga

fur.

203

III. Genus VULPES

Characters. — Skull
slender and low
(depth of brain-case
less than one third
a ee A
interorbital regio
nearly flat, the Sotal
sinuses scarcely in-
flated, the postorbi-
tal processes thin,
slightly concave
above, their edges
overhanging and
bead-like; dorsal pro-
file of forehead ris-
ing very slightly and
gradually above level
of rostrum; dental
formula as in Canis;
teeth relatively light
and small, the length
of upper carnassial
and molars together
contained about 234
to 3 times in palatal

teeth somewhat more
trenchant than

Canis, the canines
slender an e
ed, point

margin of mandibu-
lar ramus when jaws
are closed (Fig. 65).

include the most

frequent characters while numerical changes include the
least frequent characters is again very strikingly brought
out in Miller’s remarks on the origin of the domestic
dogs from the wolf (Canis lupus):

204 THE AMERICAN NATURALIST [ Von. XLIX

The only known characters by which the skull of Canis lupus can
be distinguished from that of the larger domestic dogs is the greater
average general size and the relatively larger teeth. In a dog’s skull
with condylobasal length of 230 mm. the length of upper and lower
carnassials is, respectively, 21.6 and 25.0 mm. In ten skulls with con-
dylobasal length of more than 200 mm. the average and extremes for
these teeth are: upper, 20.5 (19-22); lower, 24.0 (22.8-26.0). In all
the dog skulls which I have examined, representing such different
breeds as the pug, fox-terrier, bloodhound, mastiff, ancient Egyptian,
ancient Peruvian, Eskimo (Greenland and Alaska) and American In-
dian, the teeth are strictly of the wolf type, never showing any ap-
proach to that of the jackal (Fig. 62).

This indicates that the profound differences of osteolog-
ical character which separate the larger breeds of domes-
tic dogs are chiefly in the proportions.

No numerical, or presence and absence characters are
used in Miller’s definition of the wolf, arctic fox, and red
fox although a number of minor numerical characters are
clearly described and figured in his text, especially the
cuspules on the incisor and premolar teeth, as shown in
Fig. 1. These numerical cusp characters would have re-
ceived more attention from a paleontologist partly be-
cause of the paucity of material which comes into his
hands, partly because he is in a position to observe the de-
velopment of these cuspules.

This contrast between proportional and numerical char-
acters brings out a fundamental law in the evolution of
the hard parts of mammals which is of great importance.
First, characters of form and proportion, without numer-
ical change, are constantly originating as a universal prin-
ciple and forming the chief distinctions between divisions
from the high rank of orders down to those of subspecies,
races, and even individuals. Second, numerical loss or
fusion of old characters of teeth, digits, or vertebre is
next in frequency, the loss always following diminution
in form and proportion. Third, numerical gain of new
characters is the least frequent process; it is relatively
rare in the endoskeleton, that is, in added teeth, added
vertebre and other segmental parts, added cranial bones,
added phalanges; it is more frequent in added cusps on

No. 580] ORIGIN OF SINGLE CHARACTERS 205

the teeth or added horns and appendages of the skull; it
is still more frequent in added exoskeletal characters, such
as dermal ossicles and armatures.

The contrast between the wolf (Canis lupus), the arctic

GEOGRAPHIC DISTRIBUTION
AT PRESENT TIME

True Fox Arctic Fox Species or Worf
Zoological VuLpes ALopex ANIS
Observation LUPUS l }
ae
7 —..o
ee
+ ©
ose
oo g 285
joS2
= o> |
ay <5
Se aoe dt E
ee ee
l i PESTER |
<— GEOGRAPHIC DISTRIBUTION ——>
IN PAST TIM

Lines and dots represent the pyle also the pot and existing
distribution of. geograp ic (onto netic and environmental)
sub- species, races, and intergrades A= ancestral type

zontal lines).

fox (Alopex lagopus), and the red fox (Vulpes vulpes)
may, moreover, be adduced for four purposes: first, to
direct attention to the nature of the numerical characters
which separate these three genera; second, to direct at-
tention to the fact that these numerical characters are
very inconspicuous and unimportant in contrast with
characters of proportion; third; to illustrate the extremely
slow development in time of new numerical characters;
fourth, to illustrate the difference between paleontological
and zoological observation, a difference which is graphic-
ally represented in the diagram (Fig. 2y.

As to time Vulpes has been separated from Canis for an

206 THE AMERICAN NATURALIST [Vou. XLIX

enormous period.’ It is clearly distinguishable in the
European Pliocene where three species of canids are
referred to the genus Vulpes by Schlosser and others.
Again, in the Upper Pliocene of India there occurs a spe-
cies of fox as well as in the Pliocene of Chine. In North
America the fox is first recorded from the Pleistocene
definitely, although an Upper Miocene species (Canis
vafer Leidy) is regarded by some as the forerunner of
Vulpes and by others as a pro-Vulpes genus. It is there-
fore probable that the phylum of the fox diverged from
that of the wolf as early as Miocene times, perhaps a
million years ago, although the generic distinctions of
proportion-characters were not fully acquired until Plio-
cene times. The ancient geologic separation of Canis and
` Vulpes is further indicated by the fact that they do not
interbreed. The marked divergence in proportions—the
fox small, slender, narrow-headed, a small-mammal and
bird catcher, the wolf: relatively large, massive, broad-
headed, a large-mammal catcher—is accompanied by the
gain or loss of several relatively obscure numerical char-
acters, such as the cuspules on the incisors and premolars
(Fig. 1. r,r, r), which are strong in the wolf, intermediate
in the arctic fox, and absent in the red fox. It would
appear that the wolf had developed these numerical cusp
characters more rapidly than its congeners. In a fossil
series the development of such cusp characters may be
followed stage by stage.

2. Observations by a Field Zoologist on the Modes of
Origin of Numerical and Proportional Characters

The special features of the field work developed under
C. Hart Merriam’s direction in the U. S. Biological Sur-
vey are: (1) the vast quantity of comparative material
brought under examination, (2) the exact geographic, cli-
matic and environmental records, (3) the assemblage o
numerous intergradations between species and sub-spe-
cies, (4) the precision of the measurements and observa-

10 I am indebted to the authority of Dr. W. D. Matthew for these remarks.

No. 580] ORIGIN OF SINGLE CHARACTERS 207

tions, but above all (5) that the facts are recorded entirely
without the influence of any biological theory, the mind
of the observer being absolutely fresh and unprejudiced.

The observations published in 1909 by Wilfred H. Os-
good on the mice of the genus Peromyscus therefore
constitute a notable and wholly unbiased research on
bodily ‘‘characters’’ as they appear to a zoologist col-
lecting and observing in the field, but examining and re-
viewing his material in the museum. The following ab-
stract is mainly in the author’s own language and has been
verified by him, although the order of treatment is rear-
ranged entirely and italics are added for emphasis in its
bearing upon the modes of origin of single characters in
living mammals.

As recorded there have been examined more than 27,000 specimens
of the American rodent genus Peromyscus (Gloger, 1841), ineluding
the so-called wood mice, deer mice, vesper mice, or white-footed mice,
having a total range from the Mexican province of Oaxaca on the
south to the Yukon, Alaska, on the north, and from Labrador to
Florida on the east to Alaska and southern California on the west.

The “genus” Peromyscus is for convenience divided into five “ sub-
genera,” which are distinguished mainly by the presence or ie
of three numerical characters, namely, tubercles on the soles of the
feet five or six, presence or absence of accessory tubercles on the first
and second molar teeth, presence. or absence of two or three pairs of
mammæ. The remaining subgenerie characters lie in differences of
proportion and in color relations. The “subgenera,” which are usually
defined by a combination of characters, may merely represent opposite
ends of an almost continuous series (e. g., oe E
Megadontomys). Intergradation is- observed also ‘cert. of the
numerical characters, as in the six to vestigial we. lake peo
of the P. maniculatus group.

The species of Peromyscus maniculatus (Wagner), alone including
forty-four subspecies, ranges from Vera Cruz to Labrador and has a
wider distribution and a larger number of intergrading forms than
any similar group of mammals known. From the typical P. manicu-
latus development may be traced step by step absolutely without
break through all the numerous subspecies.

Perfect integradation, in proportion and color intensity and dis-
tribution characters, is observed between the related forms e sub-

11 Osgood, Wilfred H., ‘‘Revision of the Mice of the American Genus
Peromyscus,’’ U. 8. Sask: t. of Agric., Bureau of Biol. Surv. No. Amer.
Fauna, No. 28. Apr. 17, 1909, 285 pp.

208 THE- AMERICAN NATURALIST [Vou. XLIX

rey

SUBSPEC IESO OF mowed YSCUS
es Bae MCPS m 1
petari

2
pE
by coolidgel ie
a z paes
A S —* 6 Jj
3 zi sonoriensis “ 7 y ie |
ee rufinus ag We
rubidus ee qi
g * nebrascens. 10 AR
nubiterrae t A
Penre bairdi wi) Å i
3 i 1s j mY j
2 S artemisiae “14 &
= gracilis "Is
2 = -g
s3 oreas “7 Aint
hy! “18
nf maniculatus “ 19
re macrorhinus “
bes me: hylacus ea

la

pecies are represented by continuous or

omplete intergradation or continuity be-

ted a r. eniai shading. Where there is no

spys uity which, as indicated in the diagram

t , is not real; this dia shows the various con-

tinuous chains of s subspecies kúk intergradations the terminal members of which
appear to be disconti

species”) of the many different faunal areas. Hundreds and even
thousands of specimens are intergrades almost equally resembling two
or more adjacent forms. Many specimens fall so near an imaginary

line between two or more “ ERS ” that it is practically a

No. 580] ORIGIN OF SINGLE CHARACTERS 2C9

to classify them other than as intergrades. Particularly difficult cases
are those in which the intergrades approximate the color of one “ sub-
species ” and the cranial characters of another, thus reducing the ques-
tion of definition to one of the relative importance of characters.
Classification becomes like the division of a spectrum and depends
largely upon the standards set, for theoretically at least the possibili-
ties of subdivisions are unlimite

Some of the principles of variation [and perhaps of hybridization,
H. F. O.] are as follows:

(1) Fortuitous individual variation is greatest in specimens from
localities lying just between the ranges of two well-established forms.

Where two genuine “subspecies” inhabit the same area and
maintain themselves distinct, each may in certain cases be traced by
a definite geographic route through every degree of intergradation to
one parent form. For example, P. arcticus lives side by side with P.
algidus in the upper Yukon, but both intergrade toward the south
with P. oreas (see Fig. 3). If from sudden or gradual natural
causes these intergrades between P. arcticus, P. oreas, and P. algidus
were to become extinct three entirely separate and distinct subspecies
would apparently be created.

(3) Sexual variation in proportional characters is so ‘slight as to
be practically unmeasurable.

4) The “species” are fairly well characterized in eranial propor-
tions, but the cranial proportions in “ subspecies” are seldom constant
throughout a series although they often afford average proportional
characters of considerable value. For example, among “ species” that
are normally brachycephalie a greater or less tendency to dolicho-
cephaly is sometimes found, and vice versa. The teeth vary chiefly in
‘proportions but seldom to great extent. Some subspecies are
dichromatie.

(5) In color there is a range of seasonal, polychromatie, and local
or geographic variation. A complete intergradation between two color
extremes may often be found in localities lying just between the ranges
of two well-established forms. Color intensities are often extremely
local and doubtless are produced immediately upon contact with cer-
tain environments. Thus if the range of a given subspecies includes
a few square miles of lava beds, specimens from that area show ap-
preciably darker color than the normal members of the subspecies oc-
cupying the surrounding region. Again, specimens from the bottom
of a dark, wooded cañon may be noticeably darker than those inhabit-
ing an open hillside only a few hundred yards away. One can hardly
avoid the suspicion, observes Osgood, that if the progeny of paler
individuals were transferred at an early age to the habitat of the darker
ones they would quite regardless of heredity develop darker color.
Such local “ geographie variations” are so great that most of the
species have developed- gece apie a by means of which

210 THE AMERICAN NATURALIST [Vowu. XLIX

they have been subdivided into -numerous “ geographic races” and
“subspecies.” Thus P. maniculatus, which ranges from the Arctic
Cirele to the Isthmus of Tehuantepec, remains constant only where
the environment is identical, hence it is represented by definable

subspecies” in almost every faunal area which it enters (see Fig. 3).

These observations of Osgood may be compared with
the taxonomic results of Miller on the one hand and with
the observations of paleontologists on the other:

First: we note that in the ‘‘species’’ of the subgenus
Peromyscus and in the ‘‘subspecies’’ of Peromyscus man-
iculatus among the vast number of characters enumer-
ated there is not a single distinction recorded in numerical
or presence and absence characters; every single charac-
ter recorded is either in the proportions of the skull, ears,
feet, and tail, or in the intensities and distribution of the
color areas—all characters of degree. The field and
museum work of Osgood thus independently accords with
the taxonomic work. of Miller, namely, ‘‘proportional
characters’’ are universal and abundant, ‘‘numerical
characters’’ are less frequent and of a higher or different
taxonomic order because much more gradual in evolution.

Second: the evidence for continuity in the origin of pro-
portional characters is absolute.

Third: in a broad way continuity is also the mode of
origin of the so-called numerical characters for it is posi-
tively observed except in the case of the mamma, and
there is no apparent reason, remarks Osgood, why the
mamme also may not have developed in the same way as
the more trivial characters. In other words, there is
almost complete continuity between groups which many
taxonomists would regard as different ‘‘genera.’’ The
numerical differences in the plantar tubercles on the soles
of the feet have not been sufficiently studied, but it is clear
that the change from 5 to 6, or vice versa, has come through
the gradual. reduction or growth of one tubercle and not
through any sudden change. Most interesting also is the
fact that the 5-tubercled Peromyscus shows decided simi-
larity to the genus Onychomys, which is 4-tubercled and
-closely allied to a

No. 580] ORIGIN OF SINGLE CHARACTERS 211

Fourth: while the numerical characters are solely ger-
minal, it is difficult or impossible to distinguish both in re-
spect to color intensities and to proportions, what is ger-
minal, permanent and hereditary from what is somatic or
due to environmental and ontogenetic influences.

These four chief conclusions drawn from the observa-
tions of Osgood may now be compared with those inde-
pendently obtained by paleontologists. i

3. Likeness and Unlikeness Between Paleontologic and
Zoologic Observation

The mammalian paleontologist observes exactly the
same kinds and degrees of characters as the zoologist,
namely, very numerous changes of proportion and form,
and relatively infrequent numerical changes. In both
respects, however, the paleontologist has the very great
advantage of observing the extremes and also many of the
intermediate stages.

The chief distinction between these observers is that
as the zoologist sees characters they are stationary, he
can only infer their separability in movement through his
inferences from the comparison of forms like Canis, Alo-
pex, and Vulpes, while the paleontologist observes several
new evolution properties in these same ‘‘characters,”’
namely their actual movement and their relative rate of
movement in various lines of descent, as well as their
origin and subsequent progression or retrogression, in
brief, their phyletic history. Thus the paleontologist is
in a position to observe more of the evolution properties
in characters of exactly the same kind. Whereas in a
series of living forms each character appears to the zool-
ogist-observer as dead or static, in a fossil series each
character appears to the paleontologist-observer as living
or dynamic, the life being displayed in what may be called
its two movements in a phyletic series. ;

The first property of the ontogenetic movement of char-
acters in fossils constituted the life work of our great
observer Alpheus Hyatt, who proposed the significant and
easily recalled terms acceleration and retardation for the

212 THE AMERICAN NATURALIST (Vor XLIX

two directions of movement seen in ontogeny and phy-
logeny. Accelerated characters are those which hurry
forward and appear in successive generations at earlier
and earlier stages in the development of the individual;
while retarded characters are those which hold back or
slow down and appear in later and still later stages in
the development of the individual of succeeding gener-
ations. We know that such ontogenetic movement is
shown both in embryonic and phylogenic development of
the individual; it causes characters to appear in ontogeny
out of the order in which they arise in phylogeny; it gives
rise to the heterochrony of Gegenbaur; its rate is meas-
ured by comparing one character with all the other char-
acters of an individual.

Psi

5 Eotitanops

~ Pa
a A d
`, 4
“Ny F

Fic. 4. Evolution = jane different proportional “ion a (B, C) from
stors (A) having similar proportio

Quite distinct is what we may call the phyletic move-
ment of a character; its rate is measured by comparing a
character in individuals of one phylum with the same
character in individuals of other phyla. It is illustrated
in the comparison of the secondary cusps of the incisor
and premolar teeth in Canis, Alopex and Vulpes; in each
phylum the same cusp has its distinctive rate of evolution
and thus may appear early in geologic time or late in
geologic time. Thus comparison of the phyletic move-
ment of the same ‘‘character’’ in various lines of descent,

No. 580] ORIGIN OF SINGLE CHARACTERS 213

which is a matter of phylogeny, is quite different from
comparison of the relative movement of a number of
different characters in single lines of descent, which is
the basis of Hyatt’s law.

To illustrate the distinction between ontogenetic and
phyletic movement: a rudiment of a horn may appear
upon the skull in one phylum of titanotheres during the
period of deposit of the base of the Bridger beds (Fig. 5),
which are 1,500 feet in thickness, and in another phylum
(Fig. 5) at the summit of these beds, many thousands
of years later; this is its relative phyletic movement.
Second, after the same horn-character has appeared long
subsequent to the birth of the individuals, in both phyla
it begins to be thrust forward in the ontogeny of individ-
uals, so that in Lower Oligocene time it begins to appear
long before birth; this is its acceleration or ontogenetic
movement,

Paleontology has also revealed the marked distinction
in the mode of origin of the two kinds of characters ob-
served in zoology, namely, between the almost universal
changes of proportion and the comparatively rare new
‘‘numerical characters. ’’!?

To the former I have applied the term allometrons, >
which signifies that differences of measurement express
all changes of proportion. From these differences indices
and ratios may be calculated. Such differences arising in
the head and in the feet are indicated in the familiar
terms dolichocephaly, brachycephaly, dolichopody, brachy-
pody, and many other convenient combinations of Greek
terms. That these changes of proportion become distinct
hereditary ‘‘characters’’ is proved in certain hybrids of
Mammals where they appear to be partly or completely
separable. Thus the cross of human broad-heads with
long-heads does not produce a blend between the two but
produces, for some generations at least, either pure doli-

12 Osborn, H. F., ‘‘Coincident Evolution Through Rectigradations (Third
Paper),’’ Science, N. S., Vol. XXVII, No. 697, May 8, 1908, pp. 749-752
(p. 752).

18 ‘í Biological Conclusions Drawn from the Study of the Titanotheres,’’
Science, N. S., No. 856, May 26, 1911, pp. 825-828 (p. 826).

214 THE AMERICAN NATURALIST [Vou. XLIX

chocephals or pure brachycephals. Characters of pro-
portion are thus ‘‘single characters’’ in the hereditary
sense.

In the comparison, for example, of certain broad-heads
with other broad-heads such characters are termed
analogous because due to similarity of structure arising
from similarity of function. Thus brachypody (abbrevia-
tion of the digits) is analogous in the rhinoceroses and
the titanotheres. The broadening of the shell of one mol-
luse is analogous to the broadening of the shell of another
molluse. The broadening is none the less the heritable
characteristic of the skull or of the shell.

Quite different are certain of the new numerical charac-
ters to which I have applied the term rectigradations,
such as new cuspules on the teeth and new rudiments of
horns, for these give rise to characters which are regarded
as homologous although not directly descended from each
other. Thus the horns in all the titanotheres are consid-
ered homologous, although they arise independently at
different times in different phyla. The larger number
of cusps in the teeth of mammals are termed homologous,
although they also have arisen quite independently of
each other. Itis obvious that unless all similar new char-
acters have originated in the offspring of a single pair,
which we know is not the case, that the vast majority of
similar new numerical characters both in vertebrates and
invertebrates are related through similarity of ancestry,
through the similarity of the tissues from which they
arise, and through the similarity of their relations, form-
ing a special kind of homology which Firbringer has
termed homomorphy.

While different in these respects of analogy and
homology there are many properties which allometrons
and rectigradations as heritable characters have in com-
mon, such as the laws of growth, correlation with sex,
mechanical correlation, differential ontogenetic move-
ment, differential phyletic movement, or differential rates
of evolution, continuity of origin, increasing intensity of

No. 580] ORIGIN OF SINGLE CHARACTERS 215

5
rd
“ihe v
atA mest
vs ý
ah mcesd
asd Il
vy i
3
Palæosyops >
P: d
‘Op l 5 I
Eotitanops J
te
$ I
A B D
A cti igradatio Allom
lew characters, ailt but ; New piperis iirin
si alee tly arising of S26 P hylum. Different
every phy sy Lerma Le all other phy la
Fig. 5. hiaai origin and evolution of similar numerical charac-
ters (A, P Garh of ns roportional characters. (0, D) all arising in
pendently ig same ancestors. Each of the five phyla (I-V)

exhibits Mati praes a (pad, mes, H) and dissimilar proportions both
in ">e. skulls and metapodials.

216 THE AMERICAN NATURALIST [Vou. XLIX

development in successive generations. For example,

rectigradation like the hypocone may become more dis-
tinct, a change of proportion like brachyecephaly may be-
come more pronounced in successive generations. Yet
there are a number of additional contrasts between the
proportional and certain numerical characters, a few of

which may be enumerated:

Proportional characters = Allo-
etrons

Allometrons give rise to analo-
gies, never to homologies; they
are quantitative and intensive and
not numerical; _ closely related
forms give rise to different al-
lometrons even within SAP ;
they may be induced experim
tally in pe ae they pena
afford indices and ratios; even spe-
cific affinity may not predispose to
the same allometrons; Sue
—both harmonic and rmonie
—frequently accompany eas
of environment; they give rise
both to convergence and to di-

Orthogenie numerical characters =
Rectigradations
Rectigradations give rise to
homologies, strictly speaking ho-
momorphie structures; they are
neomorphs, new outgrowths, nu-
merically new characters; similar
rectigradations may

parallelism or convergence be-
tween the members of related
phyla; they are comparatively in-
frequent phenomena; they are not
known to be produced experimen-
tally in ontogeny; they arise from

vergence. minute beginnings at different
points in the tissues; they adopt
the characters of proportion in

surrounding parts; no true recti-

merous the similar rectigradations.

4. Differences of Opinion as to the Origin of New
Numerical and Proportional Characters

In my opinion, which is not shared by all my co-workers,
rectigradations and allometrons are qualitatively differ-
ent characters and are attributable to different combina-
tions of causes. For example, the additional cuspules on
the teeth of Canis, Alopex and Vulpes are typical recti-
gradations; they are ‘‘characters’’ qualitatively different
from the dolichocephaly of Vulpes or the relative brachy-

No. 580] ORIGIN OF SINGLE CHARACTERS 217

cephaly of Canis. This opinion was formed in 1905 and
has in my mind been established by further research.

It is, moreover, my theoretical view that rectigrada-
tions arise from some kind of germinal predisposition or
perinad or potential homology. While the‘‘homology’’

r ‘‘homomorphy’’ uniting these new characters seem
2 be due to some internal hereditary kinship between the
descendants of similar ancestors, their appearance is not

Fic. 6. Distinction between sport or mutational characters (s), which have
no significance in the evolution of the teeth, and rectigradations (r), which
are very important.
spontaneous, but is invoked in some way connected with
similar bodily and environmental reactions which also we
do not at all comprehend. For certainly there is no evi-
dence that such ‘‘homologues’’ or ‘‘homomorphs’’ arise
from similar internal perfecting tendencies or teleologic
causes which operate independently of the reactions of en-
vironment and habit.

The fact that certain rectigradations appear to corre-
spond with antecedent mechanical reactions in certain
cases, such as in the cuspules of the teeth, has led to the
opinion of Cope that these bodily mechanical reactions are
causative, but this opinion is completely offset by the fact

218 THE AMERICAN NATURALIST [Vou. XLIX

that many rectigradations occur in both vertebrates and
invertebrates which are not preceded by mechanical reac-
tions in the bodily tissues, the ornamental characteristics
of the shells of molluses, for example.

In brief, the mechanical reaction hypothesis of Lamarck
and Cope fails both as to the origin of certain new recti-
gradations and of certain new allometrons. For example,
the extremely elongated limbs of certain young quad-
rupeds, such as young horses and young guanocos, are
proportional characters which are certainly not due to the
inheritance of mechanical reactions in the adults because
they are entirely different from the adult proportions.

For these various reasons I have reached the opinion
that, whatever the respective causes of rectigradations
and allometrons may be, they are different; that is, the
occasional origin of new numerical characters and the

.constant changes of proportion which are going on in all
organisms are due to a different series of direct causes.

This divergence into matters of opinion is, however,
parenthetical. Let us now return to the observation of
facts which throw light upon the properties and qualities
of these least characters.

5. Observed Differences in the Origin and Inheritance of
Proportional and Numerical Characters

Origin. Thefundamental distinction between the origin
of rectigradations and of allometrons is well illustrated
in the six phyla, I-V, of Eocene titanotheres (Fig. 5).

It is seen that similar horn rudiments and similar cusp
rudiments arise independently at different geologic times
in every phylum, giving rise to a great number of new
homomorphie characters. On the other hand, each phylum
has its peculiar and distinctive allometrons both in the
bones of the skull and of the feet. These changes of pro-
portion are so universal and so profound that by a vast
system of comparative measurements it has been ascer-
tained that every bone of the skull, of the limbs and of
the feet has its differential rate of increase and decrease.
Since these characters of form and proportion are real

No. 580] ORIGIN OF SINGLE CHARACTERS 219

characters and since they affect every bone in the skele-
ton we discover that characters of taxonomic import may
be found in every one of the small bones of the wrist and
ankle joints, which while less readily measurable are of
exactly the same kind of value in classification as the more
conspicuous changes of proportion in the skull and in the

KIANG

<5 :
MERYCHIPPUS MULE ZEBRA

m8
gypave
ORSE

Fig. 7. The pli eee (5) a ee pepanen in the grinding teeth
of different members of the family of horses. Present in the Miocene Mery-
chippus, in the jaren he kiang, zebra, owdi horse, common horse; absent
in the ass and the mule.

feet which Miller has used throughout in his definitions
of the Carnivora. In other words a ‘‘species’’ may as
consistently be defined by the proportion-characters of
one of its carpal bones as of one of its cranial bones;
such a definition would be strange and inconvenient, but
it would be quite as scientific.

The rectigradations are also used in systematic defini-
tion only as soon as they become sufficiently large and con-
spicuous to be computed numerically. Looking up the
ancient definitions of the Eocene horses by Marsh and

220 THE AMERICAN NATURALIST [Vou. XLIX

Cope we note in every instance that as soon as a cusp
passes beyond the rudimentary stage it is apt to be ob-
served and used in definition.

So far as we know both rectigradations and allometrons
arise continuously, definitely or determinately, and so
far as we have observed they arise adaptively or in an
adaptive direction from the very beginning.

Inheritance. The germinal separability of the ‘‘least-
characters’’ known as rectigradations is well illustrated
in the case of the ‘‘ pli caballin,’’ a delicate fold of enamel

A = Thag

E.frateynus E Aelius
m pt

E. sepli at E: complicatus
i pe

~-5

E.lauwrentous
pt

=
KSN

E. paci icus

E.niobravensis pt
7 a ~--5
pee heed "> eee E. giganteus

Fie. 8. The pli caballin (5) more or less distinctly developed in the superior
grinding teeth of twelve species of the Pleistocene horses of North America. it
is observed that these grinding teeth differ profoundly in the proportions of all
their parts. The pli caballin (5) is worn off in the aged grinding tooth.

which the French systematic writers a century ago selected
as a specific ‘‘character’’ by which the horse (E. cabal-
lus) could invariably be distinguished from the ass (L.
asinus). They little knew how very ancient and. stable
this minute character is. We see it strongly developed in
the Miocene Merychippus. We do not know whether it

No. 580] ORIGIN OF SINGLE CHARACTERS 221

developed gradually or suddenly within the highly varied
horses of this genus. It appears (Fig. 5) more or less
fully developed in all of the many known species of Pleis-
tocene horses of America as described by Leidy, Gidley
and Hay. It lies near the surface of the crown, and in
much-worn teeth it disappears because the fold is seldom
continued down into the lower half of the crown. It is
entirely absent in the grinding teeth of the domesticated
ass (E. asinus), yet it is present in the kiang (E. kiang).

The complete germinal separability of the ‘‘ pli cabal-
lin’’ as a hereditary character is demonstrated by its
absence in the grinders of the mule, the cross between
E. caballus Ẹ and E. asinus g; these grinders of the mule
hybrid also prove that rectigradations are distinct from
allometrons, because the rectigradations of the maternal
horse molar are not inherited in the hybrid while the
allometrons are inherited, namely, the elongated propor-
tions of the maternal horse molar. I am preparing to
investigate the grinding teeth of the hinny, the cross
between the male horse and the female ass, to ascertain
whether the same contrast in heredity prevails here. I
suspect not because the hinny appears to have the shorter
head of the ass rather than the very long horse-like head
of the mule.

The germinal separability of allometrons or propor-
tional characters of mammals is also observed, but it
appears to be less complete than that of rectigradations.
This is demonstrated not only in the grinding teeth but
in the skull of the mule hybrid, in which the majority of
the head proportions present the same indices as in the
horse, while the minority of the head proportions present
a blend between the indices of the horse and ass. Again
in Homo sapiens the allometrons are in the first genera-
tion completely separated; in intermarriage of dolicho-
cephalic and brachycephalic individuals the children
do not form a blend of their parents but inherit either
the pure dolichocephalic or pure brachycephalic head
form. Prolonged interbreeding and intermixture þe-
tween long-headed and broad-headed human races ap-

222 THE AMERICAN NATURALIST [Von XLIX

pears to break down these separable allometrons and ul-
timately results in blending. This may be partly due to
the fact that changes of cranial proportion occur not only
within the species, but within the races and sub-races of
Homo sapiens, as witnessed in the mongoloid Indiam
races of North and South America. In other words, the
allometrons in man are of more recent origin than in the
horse and ass, which probably separated from each other
as far back as the Lower Miocene. Further experiments
and observations are greatly needed as to the separable-
ness or blending of allometrons in hybrids.

As to the rapidity of evolution of proportional and
numerical characters it appears that in certain lines al-
lometrons may evolve more rapidly than rectigradations.
This is seen in the titanotheres (Figs. 4, 5, 10), in which
changes of proportion develop very rapidly, while the
rectigradations on the grinding teeth and the rudiments
of horns develop very slowly. On the other hand, in the
contemporary Eocene horses the rectigradations seen in
the addition of cusps develop very much more rapidly
than the changes of proportion in the skull. This con-
trast between horses and titanotheres, however, confirms
the universal law that every ‘‘character’’ has its differen-
tial phyletic movement as well as its differential onto-
genetic movement.

That these movements are not identical is further
shown by a familiar illustration. The median toes of
the feet of the desert-living Hipparion have a much more
rapid phyletic movement than the median toes in the
forest-living Hypohippus, yet we may be sure that the
limbs of the newly born foals of Hipparion and of
Hypohippus were alike relatively elongated to enable
these foals to accompany the mares in flight, this adapta-
tion being secured through ontogenetic movement, or ac-
celeration.

These differentials in the velocity of characters in their
phyletic and in ontogenetic movements may afford one
of several reasons why allometrons, or proportional
Sharhehirs are separable in hybrids, why some ‘‘unit

No. 580] ORIGIN OF SINGLE CHARACTERS 223

characters’’ are dominant and others recessive. This
raises the general problem of the various causes of sep-
arability of characters in the body and in the germ.
First, it will appear that continuity or discontinuity of
origin has little to do with separability in the germ.

6. Waagen’s Observations on the Continuous and Ortho-
genic Origin of New Characters

The first paleontologist to point out the separate origin
and phyletic movement of single new numerical charac-
ters as distinguished from contemporary proportional
changes was Waagen in his observations on Ammonites
subradiatus, published in 1869 (p. 23). His two great
PEIEE announced as follows:

I. [The Variety.] The characters observed in space by botanists
and zoologists to a “local varieties,” “ geographic varieties,”

“varieties in space” are of variable value and of small systematic
importance. They appear to be temporary. They do not reappear in
the next higher geologic stratum. For these characters the long-used
name “ variety ” will suffice.

II. [The Mutation.] In contrast to the variety I venture to pro-
pose a new term, “mutation,” for the early and later phases (formen)
of a species observed in time. These mutations are characters which
are highly constant, although minute they surely are recognizable
again, on which account far greater weight must be put on mutations.
They ought to be very precisely pointed out, for mutation characters
even when displayed in the most minute features are certain to re-

show a somewhat different appearance. Ordinarily the gradations
between the mutations are the more minute as the stratum from whieh
specimens come are the more closely connected.

An ascending series of mutations in successive geologic
horizons taken together constitute Waagen’s Collectivart,
‘which is equivalent to the Formenreihe of Beyrich; it is
also equivalent to the phylum of more modern termi-
nology. Each mutation stage includes a number of geo-
graphic ‘‘varieties.’’ In any given geologic stratum a

14 Waagen, W., ‘‘Die Formenreihe des Ammonites subradiatus. Versuch
einer Palkonitologivehsn Monographie,’’ Geognostisch-Paldontologische pei
padi a II, Heft II, Nov. 1869, pp. 179-256 ae pp. 1-78),

224 THE AMERICAN NATURALIST [Vou. XLIX

‘‘mutation of Waagen’’ would appear as a Linnean
‘<species’’ when compared by an observer with contem-
porary mutations in other phyla; that is, each phylum
may be separated from contemporaneous phyla by valid
Linnean characters.

The essence of Waagen’s discovery is that when we
observe the origin and evolution of single characters in
time we are able to detect the incipience of new charac-
ters and the profound hereditary phyletic movements

Fic. 9. “ Mutations of Waagen ” seen in Spirifer mucronatus. On horizonta
gees are the geographic “ varieties ” differing in proportions. ta vertical lines
are the successive “mutations ” differing in proportional and numerical char
acters. After Grabau.

which can not be observed by the zoologist at all. We
are able, moreover, to distinguish the minute and even
inconspicuous characters, which are evolving in definite’
directions and accumulating in successive generations,
from the indefinite and transitory characters of the geo-
graphic ‘‘variety.’? The underlying cause of this dis-
tinction in the light of our present knowledge is that mu-
tations denote germinal or phyletic evolution while vari-
eties may simply denote the bodily fluctuations caused by

No. 580] ORIGIN OF SINGLE CHARACTERS 225

habit and environment. This phyletic movement was
termed Mutationsrichtung by Neumayr.

The significance of the term mutation as defined by
Waagen is to be found only in his original definition;
it has been used in many different senses before and
since.!ë It is a taxonomic term for each of the minute
subdivisions of a specific phylum which may be defined
by certain degrees of advance in ‘‘mutation-characters”’
evolving continuously in definite directions. The verte-
brate paleontologist Depéret in 1907 pointed out that the
Waagen ‘‘mutation-characters’’ have this special charac-
teristic, they are always produced in the same direction
without oscillations or retarded steps. The lacune are
so infrequent as not to interrupt the general view of the
continuity. Each of the closely linked terms of any
series may be designated as ‘‘ascending mutations’’ in
rising strata. It becomes possible to recognize the in-
tensity of the action of time.

The mutations of Waagen demonstrate five very im-
portant principles in tlie evolution of certain ‘‘least char-
acters’’ as follows: (1) origin from inconspicuous begin-
nings; (2) continuity rather than discontinuity; (3) defi-
nite direction or Mutationsrichtung rather than indefinite
or variable evolution; -(4) a definite rate of phyletic
movement.

These principles appear to involve a hereditary or
germinal basis, a Mutationsrichtung, such as also appears
to underlie Osborn’s rectigradations.

The two kinds of characters observed by Waagen’? are
also exactly similar to those observed by field zoologists
and by vertebrate paleontologists, namely:

The mutation of Waagen is a stage of advance in the development
either of numerical or of proportional characters, definable when one or
more of these minute and originally unimportant characters become
visible and measurable. r

15 Scott, W. B., ‘‘On Variations and Mutations,’’ The Amer. Journ. of
Science, 3d Ser., Vol. XLVIII (Whole No. CXLVIII), Nos. 283-288, July-
Dec., 1894, pp. 355-374. In this paon wore continuous ‘‘ mutation’’ is
dietiniguighad from indefinite ‘‘ variation

16 Waagen, op. cit.

226 THE AMERICAN NATURALIST [VoL. XLIX

It has now been made clear why these stages are recog-
nizable and definable by the paleontologist and not by the
botanist, zoologist or experimentalist.

To sum up the observations of zoologists and pale-
ontologists with regard to the single property of the
movement of characters in progressive organisms the
following principles are observed:

All characters are in simultaneous movement; such
movements are differential, each character has its own
rate; some movements are ontogenic or with a velocity-
relation to other characters in the same individual; others
are phyletic or with a velocity-relation to similar charac-
ters in other species and genera; ‘‘least-character”’
movements are continuously progressive or retrogres-
sive; certain least-characters are stationary; ‘‘least-char-
acter’? movements may develop or manifest a certain
direction or trend, the Mutationsrichtung, and thus may
be cumulative to an extreme; all character movements
which are cumulative in successive generations have a
germinal or hereditary basis. Thus differential move-
ment is one of the most distinctive and important proper-
ties of the ‘‘character.’’

This principle of continuous germinal change which
appears to underlie the continuous development of visible
bodily characters may prove to be in harmony with
rather than in opposition to the law through which some
characters appear suddenly, or by saltation. As I
pointed out many years ago, there may be an apparent
but not a real eontrast between the ‘‘mutations of
Waagen” and the ‘‘mutations of De Vries.” If the
former rise through continuous germinal changes and
the latter rise through discontinuous germinal changes
the common element in both may be the Mutationsricht-
ung, or trend of development. In Waagen’s law this
trend of development appears to express itself in a con-
tinuous change of visible somatic development. of charac-
ters. In De Vries’s observations the change is believed
to be discontinuous, suddenly constituting the ‘‘mutant’’

No. 580] ORIGIN OF SINGLE CHARACTERS 227

or ‘‘elementary species’? which is a subspecific stage
comparable to the ‘‘mutation’’ of Waagen. l

T. De Vries’s Observations on the Discontinuous and
Indefinite Origin of Characters

The sudden origin of ‘‘characters,’’ new, germinal,
saltatory, in every direction but of sufficient value to
come under the influence of selection—these are the es-
sential features of the famous mutation hypothesis of
De Vries. While I do not accept this hypothesis as a
demonstrated natural principle like that of Waagen, the
opinions of its distinguished author may be clearly set
forth as compared with the observations of Waagen
which have been repeatedly confirmed and verified:

(1) De Vries’s ‘‘mutants’’ differ from the ‘‘muta-
tions’’ of Waagen in appearing as fully formed ‘‘charac-
ters’’ and attracting our immediate attention and obser-
vation, instead of passing through a long series of initial
and rudimentary stages in which they are barely dis-
cernible. (2) The ‘‘mutation-characters’’ observed by
De Vries differ from Waagen’s mutation-characters in
lacking any definite or determinate direction; on the con-
trary, it is of their essence that they appear in any or
-all directions. (3) They agree, however, with the ‘‘mu-
tation-characters’’ of Waagen and with the rectigrada-
tion-characters of Osborn in the fact that similar muta-
tions may arise independently at various times in
branches of the same stock, thus giving rise to homo-
morphic characters. (4) Let us note that the new sys-
tematic unit of De Vries, the ‘‘mutant”’ or ‘‘elementary
species,’’ is a space or geographic phenomenon; it may
be contemporary with many other mutants of a single
Linnean species. (5) The ‘‘mutation-character’’ of De
Vries is not a demonstrated equivalent to the ‘‘mutation-
character” of Waagen, which is a time or geologic phe-
nomenon character, observable only in a long series of
generations from the same ancestor.

Now as to the present state of evidence for the saltation

228 THE AMERICAN NATURALIST (VoL. XLIX

hypothesis of De Vries, let us be on our guard between
fact and opinion, between natural and unnatural phe-
nomena. That the saltation of characters occurs very
frequently in plants and in animals under artificial con--
ditions there can be no doubt; yet the mass of existing
evidence is from artificial rather than from natural
sources.

The recent opinion of Bateson,’* who by his advocacy
of discontinuity in the origin of all specific characters
would be predisposed to favor the saltation theory held
by De Vries, is partly negative.

The evidence for the appearance of mutations of higher order,
by which new species characterized by several distinct features are
ereated, is far less strong, and after the best study of records which I
have been able to make I find myself unconvinced. ... In so far as
mutations may consist in meristie [i. e. numerical]1* changes of many
kinds and in the loss of [germinal]?® factors it is unnecessary to
repeat that we have obtained evidence of their frequent occurrence.

Negative conclusions have also been reached by various
botanists, as, for example, Jeffrey :1°

1. The or amen are largely characterized by hybrid contamina-
tion in natur

2. This atest holds with particular force for Gnothera lamarck-
iana and other species of the genus @nothera, which have served as
the most important basis of the mutation hypothesis of De Vries.

7. The mutation hypothesis of De Vries, so far as it is supported
by the case of Œnothera lamarckiana, is invalidate

I do not know of a single instance where a field observer
in mammalogy or in paleontology. has recorded a new
saltation character which is known to be of any signifi-
cance in the evolution of the race. On the other hand,
certain field observers of birds (Beebe) and of molluscs
(Crampton) are of the opinion that they have discov-
ered proofs that certain characters arise by saltation

17 Bateson, William, í Problems of Geneties.’’ Oxford University Press,
1913, 250 pp.

18 These [] are insertions in Bateson’s text by the present author.

19 Jeffrey, Edw. C., ‘‘Some Fundamental Morphological Objections to
the Mutation Theory of De Vries,” Tue AMER. Naturalist, Vol. XLIX,
No. 577, Jan., 1915, pp. 5-21.

No. 580] ORIGIN OF SINGLE CHARACTERS 229

in a state of nature. The vast majority of observa-
tions on the evolution of mammals either in the field
(e. g., Osgood) or among fossil series, where the in-
tergradations have not been destroyed, points to con-
tinuity in the origin of changes of proportion. The evi-
dence as to this continuity both in proportional and in cer-
tain numerical characters in fossil vertebrates and in-
vertebrates is overwhelming. Saltation is, however, theo-
retically probable in certain numerical and meristic char-
acters, such as supernumerary teeth and vertebre.

8. The Separability of Characters in the Body

Apart from the question of the origin of new charac-
ters by saltation, the observations of De Vries and his
followers furnish additional evidence of the separability
of characters both in the body and in the germ. This
law of separability in its human aspect was well ex-
pressed by Galton in 1889.2?

. We seem to inherit bit by bit, this element from one progenitor, that
from another, under conditions that will be more pate expressed as we
proceed, while the several bits are themselves liable me small change
during the process of transmission. Inheritance may enon be described
as largely if not wholly ‘‘particulate,’’? and as such it will be treated in
these pages. Though this word is good English and accurately expresses
its own meaning, the application now made of it will be better understood
through an illustration. Thus, many of the modern buildings in Italy are
historically known to have been built out of the pillaged structures of older
days. Here we may observe a column or a lintel serving the same purpose
for a second time, and perhaps bearing an inscription that testifies to its
origin, while as to the other stones, though the mason may have chipped
them here and there, and Pape their shapes a little, few, if any, came
direct from the quarry. This simile gives a rude pi gh true idea of the
exact meaning of Particulate idle! sae namely, that each piece of the
new structure is derived from a corresponding piece of some older one, as
a lintel was derived from a lintel, a column from a column, a piece of wall
from a piece of wall

I will pursue this rough simile just one step further, which is as much as
it will bear. Suppose we were building a house with second-hand materials
carted from a dealer’s yard, we should often find considerable portions of
the same old houses to be still grouped together. Materials derived from
various structures might have been moved and much shuffled together in the

20 Galton, Francis, ‘‘ Natural Inheritance.’’ 8vo. The Macmillan Com-
pany, 1889.

230 THE AMERICAN NATURALIST [VoL. XLIX

yard, yet pieces from the same source would frequently remain in juxta-
position and it may be entangled. They would lie side by side ready to be
carted away at the same time and to be re-erected together anew. So in the
process of transmission by inheritance, elements derived from the same

ancestor are apt to appear in large groups, just as if they had clung
weet in the pre-embryonic stage, as perhaps they did. They form what
is well expressed by the word ‘‘traits,’’ traits of feature and character—
that is to say, continuous features and not isolated points.

The observations which we have been comparing on
the origin of new characters in vertebrates, inverte-
brates and plants certainly afford some insight into the
laws of germinal change wherever we can distinguish
between the true germinal expression of a character and
its visible modification by ontogeny or environment.
Recurring to the antithesis between the separability and
the correlation of characters we may again sum up some
of the many properties and qualities of single characters:

Laws of Fo Laws of Correlation

ndent germinal o Hereditary connection with the

ndepen

characters, either paei or
discontinuous; independent devel-
opment of similar new characters
in near or remotely related de-
scendants of the same ancestors;
development of dissimilar changes
of proportion in descendants of
nearly related ancestors; develop-
nape of similar “ rectigradations,”

“mutations of Waagen,” and
“mutations of De Vries” at dif-
ferent times in descendants of the
same ancestors; separate rate of
evolution in ontogenetic movement
and in phyletic movement; each
character with its own origin, in-
dividuality, and rate of movement,

origin of similar characters in
other lines of descent; strong sex
correlation in the size and pro-
portions of all characters; mechan-
ical correlation both in the al-

rons and rectigradations;
compensation correlations when
one character is developed by the
sacrifice of another; proportional

—
©

all the new characters and changes
of proportion share the general
form; a utility or selection cor- '
relation where one character is

adaptive, defensive and offensive
correlation where groups of char-
acters evolve to serve'similar pur-
poses.

At the present time we appear to understand the laws
of separability somewhat better than the laws of corre-
lation, for the latter are enveloped in the deepest mys-
tery. We have not even alluded to the chemico-physical

No. 580] ORIGIN OF SINGLE CHARACTERS 231

correlations, by means of the hormones, which are observ-
able in the field of physiology and medicine. Mechanical
correlation, as where an incipient rectigradation in an
upper tooth corresponds precisely with an incipient recti-
gradation in a lower tooth so that the upper and lower
teeth evolve. throughout in perfect mechanical harmony
by the addition of character after character, affords one
of the most vivid instances of our total ignorance of the
causes underlying the origin of new characters.

9. The Separability of ‘‘Characters’’ in the Germ

The pioneer discoverer of the separability of ‘‘charac-
ters’’ in the germ was Gregor Mendel, who in 1865 exam-
ined seven pairs of alternating characters in hybrids of
two varieties of the common pea (Pisum sativum). The
outstanding feature of this great discovery is that the
separability of the ‘‘determiners’’ of characters in the
germ is far more sharply defined than the separability
of the corresponding somatic ‘‘characters’’ in the visible
body; as regards the ‘‘determiners’’ of many characters
it is sharp and absolute.

A second feature of Mendel’s discovery is of special
significance in our present review of proportional and
numerical characters in the hard parts of vertebrates
beeause the alternative characters which Mendel experi-
mented with were of both the proportional and the nu-
merical kind.. For example, Mendel’s proportional char-
acters of ‘‘tallness’’ and ‘‘dwarfness’’ in the pea stalk
may not improperly be regarded as analogous with the
alternative characters of dolichocephaly and _ brachy-
cephaly of the skullofthe¢mammal. Again, Mendel’s form
-of the pea-seed, whether round or wrinkled, may not im-
properly be compared respectively with the folded or
smooth condition of the enamel in the molar teeth of the
ass and of the horse, which we observe is alternative in
heredity. On the other hand, Mendel’s definite color
characters, such as his flower color, whether purple, red
or white, or his seed color, whether yellow or green, have
been definitely compared by experiment and found to

232 THE AMERICAN NATURALIST (Vor. XLIX

correspond in heredity with the coat colors of mammals.

A third great feature of Mendel’s discovery is that in
such alternative pairs of characters as tallness and
dwarfness one may be dominant and the other recessive,
following in successive generations the well-established
Mendelian law.

It remains to be observed whether any of the propor-
tional characters of mammals follow this law.

It remains to be observed, also, whether rectigrada-
tonil characters like the ‘‘ pli baballin’” in the grinding
teeth of the horse exhibit in heredity the Mendelian laws
of dominance and recession. This could only be ascer-
tained among mammals by observations on hybrids in
such a family as the Bovide, which are fertile inter se,”
or on similar characters in the hybrids between wild
natural breeds.

The discovery through experiment by zoologists and `
botanists that many saltation characters, such as sports
and abnormalities, follow the Mendelian law of domi-
nance and recession has led to the entirely unwarrant-
able assumption that only characters which are discon-
tinuous, or of sudden origin, are sharply separable in
heredity. This we have seen to be not in accord with
the facts, for the principle of separability is quite as
sharply defined in certain characters which have evolved
slowly and continuously as in those which have evolved
suddenly.

The psem significance of recent Mandelo discov-

21 Bartle D ‘Wild en in Captivity.’’ 8vo. Chapman and
Hall, Ld. tee 1899, p. 219. ‘‘The two species of camel (Camelus
dromedarius and C. bactrianus) will breed together; the llama (Auchenia
glama) will breed with the alpaca (A. pacos), and the offspring are fertile.
Several species of deer, when crossed, produce fertile hybrids: for instance,
the Barbary deer (Cervus barbarus) with the red deer (C. elaphus), the
Mexican (C. mexicanus) with the Virginian deer (C. virginianus). Several
others are also recorded upon good authority. Several instances of hybrids
among the carnivora are well authenticated. The lion (Felis leo) has bred
with the tiger (F. tigris), the leopard (F. leopardus) with the jaguar (F.
geet = wild cat (F. catus) with the domestic cat (F. domestica).’’

‘‘ Wild Beasts in the ‘Zoo.’ ’? 8vo. Chapman and Hall, Ld. London,
1900, er 71-72. ‘‘Hybrid Bovine Animals. . . . In the first place, the bull

No. 580] ORIGIN OF SINGLE CHARACTERS 233

eries to the zoologist and paleontologist is that what
we zoologists describe as a ‘‘single character,’’ the horn
of a sheep or of a titanothere, for example, is probably
the expression of a very large number of germinal ‘‘de-
terminers’’ all of which are necessary in the germ to
produce the horn. If only one of these ‘‘determiners’’
should change the character of the horn would change,
or the horns might be lost altogether. Thus the germi-
nal correlation of ‘‘determiners’’ to produce what ap-
pears as a single character to the zoologist and paleon-
tologist is visibly represented in the bodily correlation
of the horn with a very elaborate offensive and defensive
mechanism including all the bony and muscular charac-
ters necessary to operate the horn effectively, as well as
the psychic desires and impulses to use the horn. We
thus have a vast array of internal and external, of struc-
tural and functional correlations with this single charac-
ter, the horn.

10. Conception of the ‘‘ Least Character”?

In the beginning of this address I noted that no one
has ever undertaken to define the ‘‘character.’’

It is very difficult if not impossible to define one of
these least characters as observed in living and fossil
series even when reinforced by the experimental evi-
dence of heredity. A definition may be essayed through

Zebu (Bos indicus) was introduced to the cow Gayal (Bibos frontalis), and
a female hybrid was born October 29, 1868 (A of pedigree).

(A) produced her first calf June 17, 1872, a second one October 16, 1873,
a third one January 5, 1875, a fourth March 11, 1876, a fifth November 2,
1878; these five calves were the produce of this female hybrid Gayal with
the Zebu bull. She was now introduced to the male American Bison pipt
earar and on May 21, 1881, she produced a fone. No. 2 (B
pedigree), ?

‘‘It will be seen that this animal (B) is the product not only of the
intermixture of three well-marked species, but, according to our present
definition, of three distinct genera.

‘‘ This remarkable animal, the result of the triple alliance, was last year
introduced to the bull Bison, and on March 12, 1884, she produced a female
(C of pedigree). This last individual, now eleven weeks old, is undistin-
guishable from a pure-bred Bison of the same age.’’

2AT: ee THE AMERICAN NATURALIST [Vou. XLIX

an enumeration of the known properties of a ‘‘least
character.”’ :

` As distinguished from a group of characters the prop-
erties of a ‘‘character’’ are its separability, its inde-
pendence, its individuality, its own rate of movement
ontogenetic and phyletic, its differentiation by these
properties from other least characters. Its separability
in heredity is shown where it can be hybridized.

From the structural or anatomic standpoint a least
character is a group of cells and tissues constituting a
diminutive organ or part of an organ subserving a dis-
tinct though subsidiary function. For example, the ‘‘pli
caballin’’ in the enamel of the horse’s tooth or the rudi-
mentary cuspule may be cited as least characters, for
each is composed of a vast number of cells and more than
one tissue, but seems to have the property of rising or
falling and behaving like a unit. -

11. “Least Characters” in Classification and Systematic
Work

This ‘‘least-character’’ conception is of great value to
the zoologist and botanist in systematic work, this con-
ception of an individual as a colony of ‘‘characters’’ each
with its principles of independence and its principles of
correlation, germinal in origin but subject to somatic
modification by environment and habit.

First, among these single characters are those ob-
served by Waagen which accumulate until they build up
into one of his ‘‘mutations.’’ One or more such single
characters compose the ‘‘mutants’’ of De Vries.

Second, the old but oft-confusing term ‘‘races’’ be-
comes clear; we may now understand the significance of
the races of the horse, for example, the Arab, the Forest,
the Steppe horse. These races are all fertile inter se
and thus have never been defined as species although fer-
tility and non-fertility are no more important in the dis-
tinction of species than any other ‘‘character.’’ The
characters which distinguish races are, nevertheless, often
of specific value; they are either proportional or numer-

No. 580] ORIGIN OF SINGLE CHARACTERS 235

ical, because in the production of these modern races
the pure ancestral forms had in their natural state
evolved a very large number of allometrons as well as
rectigradations and other numerical characters which
have to a slight degree blended in intercourse and to
a larger degree have maintained their purity and dis-
tinctness. Thus you will observe among the modern
races of horses the most incongruous mixtures; an old
eart horse with the head and quarters of the Forest
type will gallop across a field and raise the bones of the
tail perfectly erect exactly like a pure bred Arab.

Similarly a ‘‘species’’ is a mosaic of an infinite num-
ber of least characters in a state of movement only a few
of which may be so definite and measurable as to be
employed in systematic definition.

12. Theoretical Conclusions as to ‘‘Characters’’ and
the ‘‘Organism’”’

These least characters when assembled in an organism
and dominated by the principles of separability and cor-
relation present to our fancy the picture of a vast regi-
ment of soldiers walking in single line; each soldier pos-
sesses his own individuality and separableness from his
comrades, each advances or lags behind according to his
individual velocity, but each subserves the general pur-
poses of the entire regimental line through the uniting
force of training and the unseen spirit of the regiment,
which represents the law of correlation.

It appears that we paleontologists have already learned
much and that we have still far more to learn by the clos-
est observation of ‘‘characters’’ in a state of natural evo-
lution. We are on somewhat safer ground than the ob-
server of the unnatural or hotbed evolution of characters
in the artificial breeds, hybrids of animals and plants
under domestication. The contrast between the excess-
ively slow natural evolution during the past million years
of the wolf, the arctic fox and the red fox, and the feverish
unnatural evolution of the domestic breeds of dogs dur-

236 THE AMERICAN NATURALIST [Vou. XLIX

ing the last ten or twelve thousand years is extremely
significant.
If the student of genetics abandons the natural and the

Illustrating the Continuity Theory

The young
Titanotheres
of every
stage in-
herit ther
Characters,

new ana Fa ower OTR

e Brontothérium
from the 12

Corm Calls an HL Titanotheres

weer d Riv oe Eoc)
of the wie rinses a

Germ
parent cell
Titanotheres. B /
The body 2G
of the
Titanothere Ge .
tS an af Y J a
offshoot 1*Gen Lower Eocene
eee Germ, Eotitanops
Celty ate |
its parent Comtenserty

Germ Plasm

(Heredity)

1c. 10. Evolution by the constant addition of koe illustrated in the
descent of Brontotherium (B) from Eotitanops (A). The nimals are repre-
sented to the same scale. They flnstrate the constant haaie on of new char-
acters in every part of the organism. Evolution in this family of quadrupeds is
almost entirely by the addition of characters. Comparatively few characters
degenerate or disappear.

normal for the unnatural and the abnormal and sticks
solely to his seed pan and his incubator he is in danger
of observing modes of origin and behavior of characters

No. 580] ORIGIN OF SINGLE CHARACTERS 237

which never have and never will occur in Nature. He
may, moreover, never observe at all certain modes of
origin and behavior as well as certain properties and
qualities of characters which are of the most fundamental
importance in relation to his particular field of heredity
and hybridizing.

While twenty years of observation of the normal and
the natural aspects of nature have brought the zoologist
and paleontologist somewhat nearer to a conception of the
modes of evolution, twenty years of continuous observa-
tion of the abnormal and unnatural have landed one of
the leading experimentalists, William Bateson, in the
state of skepticism and agnosticism expressed in his recent
work (p. 248, italics our own) :??

The many converging lines of evidence point so clearly to the central
fact of the origin of the forms of life by an evolutionary process
that we are compelled to accept this deduction, but as to almost all
the essential features, whether of cause or mode, by which specific
diversity has become what we perceive it to be, we have to confess an
ignorance nearly total. The transformation of masses of population
by imperceptible steps guided by selection, is, as most of us now see,
so inapplicable to the facts, whether of variation or of specificity,
that we can only marvel both at the want of penetration displayed by
the advocates of such a proposition, and at the forensic skill by which
it was made to appear acceptable even for a time.

If the principle of the continuous and independent
movement of each member of a vast colony of single
characters is firmly established, as it appears to be
through vertebrate and invertebrate paleontology, we
must abandon entirely one tradition left by the master
mind of Darwin which has permeated the work of all the
original Darwinians and Neo-Darwinians, and which is
equally strong in the mind of De Vries. Bateson has re-
cently maintained this tradition of the origin of ‘‘species”’
from fortuitous saltatory characters in the following lan-
guage.?3

22 Bateson, William, ‘‘ Problems of Genetics.’’ 8vo. Oxford University
Press, 1913, 250 pp.

23 Op, cit., p. 248. Italics our own.

238 THE AMERICAN NATURALIST [Vou. XLIX

In place of this doctrine we have little teaching of a positive kind
to offer. We have direct perception that new forms of life may arise
sporadically, and that they differ from their progenitors quite suffi-
ciently to pass for species. By the success and maintenance of such
sporadically arising forms, moreover, there is no reasonable doubt that
innumerable strains, whether in isolation or in community with their
co-derivatives, have as a fact arisen, which now pass in the lists of
systematists as species.

Broadly stated, this tradition is that evolution mani-
fests itself suddenly in one character or group of char
acters; that either through individual variation such a
character or group of characters is preserved and accu-
mulated by selection, or, through saltation that such a
character or group of characters suddenly arises and is
imperishably fixed in the race by selection.

This is the essential feature of the Darwinian concep-
tion of evolution, namely, that an organism advances now
here, now there. Such a conception is one which would
naturally be fosterd by observers of single living plants `
or animals living under unnatural conditions, or by ex-
perimentalists who observe a brief contemporary chain
of organisms.

The observation of ‘‘characters’’ in phyla or groups of
organisms advancing on a grand scale in space or in time
shows that this Darwinian tradition is so partial and inad-
equate as to be practically false. It has been observed
that every organism consists of an almost infinite number
of characters, it has also been observed that the evolu-
tion of some of these characters may be so conspicuous as
for a time to attract the attention of the observer or as
to constitute the chief magnet for the power of selection.
It has not been observed that the entire organism waits
on any one of these characters. On the contrary, in all
progressive organisms in which a very large number of
characters are simultaneously observed it proves that
every character in every part of the body is in a contin-
uous state of movement. This is the actual result of
observation and measurement.

As regards natural selection in relation to the “oe of

No. 580] ORIGIN OF SINGLE CHARACTERS 239

characters we know nothing, we stand by the theoretic
opinion that: Selection is operating always upon the sum
of all the movements, actions and reactions of characters
known as the Orcanism and upon all single characters
of survival or elimination value.

Very recent is Bateson’s enunciation** of the novel
hypothesis that we may have to forego the theory of
addition of germinal factors or determiners and substitute
a theory of variation by loss of ‘‘factors’’:

- Paleontology affords only indirect evidence as to ger-
minal ‘‘factors’’ but it offers the most positive testimony
both as to evolution largely by the loss of characters, as in
the case of the family of horses, and evolution largely by
the addition of characters, as in the family of titanotheres -
displayed in Fig. 10. It is the constant addition of new
somatic characters in the evolution of members of the
latter family which forms the background of the present
address. Whether the incessant and most impressive ad-
dition of the new somatic characters which transform Eoti-
tanops into Brontotherium are the visible result of a sub-
traction of germinal ‘‘factors’’? may be a subject for
metaphysical discussion, but is certainly without the
bounds of all natural evidence. A natural view is that the
invisible germ is being continuously enriched with the vis-
ible body by processes of which we can form no concep-
tion whatever.

24 Bateson, Wm., ‘‘ Heredity.’’ Inaugural Address of President to The
Australian Meatiag of the British Association. Nature, Vol. 93, No. 2338,
Aug. 20, 1914, pp. 635-642. (Italies our own.) ‘ʻI feel no reasonable
doubt that though we may have to forego a claim to variations b
of factors, yet variation both by loss of factors and by fractionation of
factors is a genuine phenomenon of contemporary nature. If, then, we
have to dispense, as seems likely, with any addition from without we must
begin seriously to consider whether the course of evolution can at all reason-
ably be represented as an unpacking of an original complex which contained
within itself the whole range of diversity which living things present. I do
not suggest that we should come to a judgment as to what is or is not
probable in these respects’’ (p. 640).

THE INFERTILITY OF RUDIMENTARY WINGED
FEMALES OF DROSOPHILA AMPELOPHILA

PROFESSOR T. H. MORGAN

COLUMBIA UNIVERSITY

Wane the infertility of the females of the mutant stock
of Drosophila called rudimentary, was apparent from the
beginning, the cause of the infertility was uncertain. Many
rudimentary females bred to males of their own kind gave

no offspring. The males with rudimentary wings, on the

other hand, were perfectly fertile with wild females and
with females of other stocks. The results might seem to
show that sperm bearing the factor for rudimentary could
not fertilize the eggs carrying the same factor. But that
this was not the entire explanation was evident: for, he-
terozygous females fertilized by rudimentary males gave
rudimentary females and males as well as long winged
flies. In the heterozygous females, however, the egg, up
to its maturity, has developed under the influence of the
normal allelomorph of rudimentary, as well as of the rudi-
mentary factor. I suggested,' therefore, that, due to this
difference during the ripening period, the rudimentary
bearing egg of the heterozygote could be fertilized by the
rudimentary sperm, although the egg of the rudimentary
female itself could not sueceed in this combination, but I
have never felt satisfied with this tentative explanation;
for, there were other possibilities, not sufficiently studied,
that might affect the result. For instance, it was not
actually observed that the rudimentary males copulate
successfully with females of their own kind, although it
was known that they could mate with any other females.
This question had first to be settled by direct observation.

Rudimentary winged females were isolated for three

1 Morgan, Zeit. f. indukt. Abs. und Vererb., VII, 1912.

240

No. 580] INFERTILITY OF DROSOPHILA 241

days after hatching, and. then each was mated to a rudi-
mentary winged male that had similarly been isolated.
In about twenty minutes, on an average, mating occurred
in an entirely normal manner. The females that had
mated were kept each in a separate bottle and given the
best food. Examination showed that hardly one of the
females had laid eggs; but in the rare cases where a
few eggs were observed, some flies developed. In the first
experiment seventeen females were seen to mate, and were
then kept alone, or with their mates. One female pro-
duced one rudimentary winged son; another gave one
rudimentary winged daughter and one such son. These
three flies were the total output of seventeen females.

The next point was to determine whether these seven-
teen females were infertile only with their own kind of
males. Each was again paired, this time to a male with
bar eyes. The character bar eye is dominant. If sperm
of these males should be successful, the female offspring
from this cross would have bar eyes, and could be distin-
guished from any others that might come from the first
‘mating. One female gave one bar daughter; another fe-
male also had one bar daughter. ` These results show that
the rudimentary females were no more successful with bar
males than with their own kind.

In a second experiment eleven rudimentary-winged fe-
males were tested with rudimentary males. One gave one
rudimentary daughter and one rudimentary son, but also
two long-winged daughters. Since I had not taken the
Same care here (using twenty-four-hour flies) as before to
be certain that the females were virgin, these two long-
winged daughters are supposedly due to fertilization be-
fore isolation, since long-winged males were hatching at
the time in the parent stock. I tested this supposition by
mating the two females to a rudimentary male, and ob-
tained the following kinds of offspring:

Long ? Long ¢ Rudimentary 9? Rudimentary ¢
56 64 15 23

242 THE AMERICAN NATURALIST [ Vou. XLIX

Evidently then the rudimentary females had been fertil-
ized by a long-winged brother before isolation, as well as
by the rudimentary male later. No offspring were pro-
duced by a third mating, to bar males.

A third and similar experiment was made. Of twelve
females, none gave any rudimentary offspring, two gave
three bar daughters apiece, but no rudimentary sons.

In a fourth experiment thirteen rudimentary-winged
females were isolated from stock. They were not neces-
sarily virgins. One gave a rudimentary-winged son; an-
other gave two rudimentary daughters and six such sons.

If only a single pair of flies is present in a culture and
few or no larve are produced, the banana generally decays
instead of fermenting. It might happen under these con-
ditions that the few larve from rudimentary females
might fail to develop, and the rudimentary sons might
also suffer, even when some of their long-winged sisters
(the father being a normal male) succeed in developing.
To make conditions favorable in this respect I proceeded
as follows: A red-eyed rudimentary female was kept at
first with a red-eyed rudimentary male for three or four
days. Then a few old white bar females and males were
added and fresh food given. The presence of these flies
and their progeny would serve to keep the food in good
condition. Moreover if the rudimentary female had been
fertilized by the rudimentary male she would produce ru-
dimentary daughters and sons. If she were subsequently
fertilized by the white bar males she would also give red
(heretozygous) barred daughters, but these could not be
distinguished from the daughters that the white-bar fe-
male would give were she fertilized by the red rudimentary
males. Nevertheless all of the sons of a rudimentary fe-
male would be rudimentary round-eyed males, regardless
as to which male was their father, and their presence
would show to what extent the rudimentary females were
fertile. The experiment was varied and simplified by
removing the rudimentary males, when the white bar
males and females were added.

No. 580] INFERTILITY OF DROSOPHILA 243

One hundred and two rudimentary females were tested
in these ways. In only three cases were rudimentary
males produced. One female produced two; one female
produced four, and another female produced one male.
It is evident, therefore, that the scarcity of rudimentary
sons can not, in general, be ascribed entirely to the condi-
tions of the food.

As pointed out above, red bar daughters might appear
in the foregoing tests and such females might have either
of the two parentages specified. Most of the females of
this kind would be expected to come from white bar fe-
males by rudimentary males, since the converse case would
rarely be realized. Seven females, that appeared, were
tested by breeding to white bar males and gave the results
in the first of the two following tables. Four others were
tested by breeding to rudimentary males with the result
shown in Table Ia. The results confirm the expectation

TABLE I
F., Rep, Lone Bar (HETEROZYGOUS) 9 BY WHITE Bar ĝ
Red White Red White Red | White| Red | Whit ed | White
Long Long Rud. Long Long Rud. | Rud. | Long | Long | Rud.
Bar Round Bar Bar Round | Bar | Round | Round Bar
; g g F a Sor g
55 44 13 24 21 t
30 30 3 14 ) 2
6 61 s 27 10 5 2
85 71 17 34 28 19
48 53 15 27 21 17
37 39 5 25 13 9
12 10 1 5 5 0 ] 1
342 304 57 156 113 59 1 2 1
TABLE I®*
F., Rep, Lone BAR (HETEROZYGOUS) 9 BY RUDIMENATRY ¢
Red | Red | Red | Red | White| Red | White| Red | White, Red | White
Long | Rud. | Long | Rud. | Long | Long | Rud. | Rud. | Long Long | Rud.
Bar | Round! Round Round| Bar Bar | Round! Bar | Round Round, Bar
8 9 9 f a E ot: dae e cde
22 15 14 13 14 6 1
37 13 1 17 26 16 9
65 10 9 26 25 24
56 25 25 12 14 12 a ea
180 63 1 65 77 6 51 1

244 THE AMERICAN NATURALIST’ [Vou. XLIX

in regard to the character of these females. Three pairs
of factors are involved which should give the following
classes of males:

Non-cross-overs Single Cross-overs Double Cross-overs
Red, rud., round. Red, long, bar. Red, long, round.
White, long, bar. White, rud., round. White, rud., bar.

Red, rud., bar.

White, long, round.
It will be observed that the rudimentary males run far
behind their schedules, due beyond doubt to their poor
viability.

White

Red ong Round

ō
Fig. 1.

Double crossing-over took place four times in the ex-
periment. If the two X chromosomes that carry respec-
tively the factors for red rudimentary round and white
long bar are represented as twisted once around each
other, as in text-fig. 1, a, the result of fusion and recom-
bination at the crossing points would give the two chromo-
somes shown in 1, b. One chromosome now carries the
factors for red, long, round, and the other the factors for
white, rudimentary, bar. As the tables show there are

No. 580] INFERTILITY OF DROSOPHILA 245

four males that fall into these two classes. There is one
female in the second table, that is red, long, round. She
must have resulted from a cross-over gamete, a long,
round egg being fertilized by a female producing sperm
of the rudimentary male.

Still another experiment like the last one was made, but
vermilion-eyed flies instead of bar-eyed flies were added.
The virgin rudimentary females that were used were not
allowed to mate first (as before) with rudimentary males,
in order to meet a possible objection to the preceding ex-
periment, namely that the spermatheca, if first filled with
sperm from the rudimentary males, might be incapable of
filling again with sperm when another male of a different
stock is added. The vermilion (long-winged males) would
give, with their own females, flies with vermilion eyes,
while the vermilion males that mated with the rudimen-
tary females should give red-eyed, long-winged females
and rudimentary red-eyed males. In the following table
the number of rudimentary females tested in each culture
is given in the top line; in the second, central, line the
number of red, long offspring given by the flies in the
square above; and in the lowest line a record of the num-
ber of vermilion offspring given by the flies in the culture
under observation. |

T Í 1 |
Number of rudimentary — | | |
2 tested Lory Uy o To oe 8 156 4 | 4
| | | bee

Fia] | | 159)
Red long daughters... | 7| H | taa

Vermilion @ and 9..... 118463 76 119 105 3 8 654/179 0) 1 167 7 46

Out of a total of ge ay rudimentary females
only nine red-eyed daughters were produced (if we ex-
clude one culture in which five red long females and three
red-eyed males appeared which must be due toerror or to
contamination). Of the nine offspring seven came from
one culture, and possibly from one female in that culture
that laid an exceptionally high number of eggs. The
complete absence of rudimentary males may be explained

246 THE AMERICAN NATURALIST [Vou. XLIX

as a result of competition, for, as Morgan and Tice have
shown, such males tend to disappear if too many other
larvee are present.

Lastly sixty-eight more rudimentary females from stock
were tested with bar males. They gave twenty-four long
bar (heterozygous) females, two rudimentary round males
and one mosaic that will be described below. In one bottle
there had been twenty-seven rudimentary females and an
examination of their food showed over forty eggs present.
Since the eggs are not easily found I estimate that prob-
ably a hundred eggs were present. Out of these eggs
sixteen females and one male developed (included in the
total given above). It appears then that many of the
eggs laid by the rudimentary females do not develop.

‘The condition of the ovaries of the eight surviving rudi-
mentary females showed that seven were full sized and
contained mature eggs. The mosaic that appeared in one
of the last crosses (Fig. 2) is interesting in several ways.
Genetically it is a female, externally it is a male in ap-
pearance, in reality it is a male in part and a female in
part although the egg must have been fertilized by a
female-producing sperm. On the right side of the body
the eye is heterozygous for bar, there is no sex comb on
the fore leg, the spines on the thorax are long, and the
wing is large. On the left side the eye is pure bar, there
is a sex comb on the foreleg, the spines on the thorax
are short, and the wing is small. The difference in size
of the two wings, and of the spines, is a characteristic
difference between the male and female, connected with a
difference in body size. The abdomen is pigmented above
as in the male, and below there is a normal penis.

Despite the apparently normal male copulatory organs,
the mosaic, when placed with mature, unmated females,
paid not the slightest attention to them, although it was
quite active. Of course its organs of perception were
female on one side of the anterior end, although male on
the other side. What physiological complex this might
give, is, of course, problematical. The mosaic died by

No. 580] INFERTILITY OF DROSOPHILA 247

becoming stuck to the glass before its behavior towards
males could be studied.
There are two ways in which this mosaic can be ac-

Fic. 2.

counted for. If an egg of the round-eyed, rudimentary
female was fertilized by a female producing spermato-
zoon of the bar-eyed, long-winged male the result should

248 THE AMERICAN NATURALIST [Vou. XLIX

be a bar (heterozygous) eyed, long-winged female, since
these are the dominants. If, then, after fertilization, dis-
location of the X chromosomes occurred at some early
division, so that while the X carrying the factor for
bar and long divided normally (each daughter nucleus
getting its proper half), the other X chromosome carry-
ing the factors for round eye and long wings failing to
divide (or else one half failed to go to one daughter
nucleus) the characteristics of the mosaic can be ac-
counted for. On the male side, the left, there would be one
X chromosome in each cell, that carries the factors for
bar eye and long wings. ‘The size of the wing and of the
spines, and the presence of the sex comb are a consequence
of the ‘‘maleness,’’ resulting from the presence of only
one X. On the female side, the right, there would, be two
X chromosomes in all of the cells, hence the heterozygous
nature of the bar eye. The length of the wing (the female
being larger than the male) and the absence of the sex
comb are a consequence of the ‘‘femaleness,’’ due to the
two X combination. The fact that the posterior end of
the abdomen is purely male is owing to this region coming
from the male contingent of nuclei, that must have over-
. lapped to the right side in this region.
The other explanation of the mosaic is that two female
producing nuclei entered, one alone giving rise to the
male side, the other one uniting with the egg nucleus giving
rise to the female side. Boveri’s explanation of gynan-
dromorphs will not apply to this case. There is no way
to decide between the first two hypotheses, but, as I
have shown elsewhere,? the hypothesis of chromosomal —
dislocation will cover all cases of gynandromorphs in
Drosophila, while that of double fertilization will not ap-
ply to one ease that gives, for itself at least, a crucial test
of the alternative hypotheses. The hypothesis of chro-
mosomal dislocation is, therefore, to be generally pre-
ferred, unless in some special case it can be shown that
2 ‘í Mosaics and Gynandromorphs in Drosophila,’? Proc. Soe. Exp. Biol.
and Med., XI, 1914.

No. 580] INFERTILITY OF DROSOPHILA 249

double fertilization has actually brought about the par-
ticular results that that case shows.

Incidentally this sex mosaic (gynandromorph) and others
of its kind confirm the conclusions drawn from grafting
experiments in insects, namely, those in which the testes
were grafted into the female and the ovary into the male
without influence on the secondary sexual characters that
developed later. These characters in the insects must be
determined by the chromosomal composition of the cells,
and not be affected by the sex ‘‘glands’’ as such. In
contrast to this situation in the insects we find in birds
that the sex ‘‘glands’’ of the female play an important
role in the suppressing in the female of some of the sec-
ondary sexual characters—characters that appear only in
the males or.in castrated females. Gynandromorphs are
exceedingly rare in birds, but there are a few well-authen-
ticated cases. It is difficult to explain their occurrence
under the conditions named above. It is just possible, how-
ever, that their occurrence may be accounted for in the fol-
lowing way. -If a mosaic condition of the chromosomal
complex shouldarise the secondary sexual characters would
still all be like those of the female, owing to the presence
of the ovarian secretion, but, if, in such a case, the ovary
should become infected, or degenerate through senile
changes, the true male parts might sooner develop the
characteristics of the male than do the true female parts,
i. e., those parts of the body that have the female sex com-
plex. This suggestion has no value unless it may lead
some one to examine the condition of the ovary, when such
a Sex mosaic again appears.

An examination of the ovaries of many rudimentary
females was made. In the majority of cases the ovaries
become nearly as large as those in the normal female, and
while they may contain full-sized eggs most of the eggs
remain immature. Examination of the food shows that
very few eggs are laid; in fact, most females lay no eggs.
Of those laid some at least hatch. From these observa-
tions, and from the experiments, it seems clear that the

250 THE AMERICAN NATURALIST [Vou XLIX

infertility of the rudimentary females is due, largely at
any rate, to their retention of their eggs, even after copu-
lation; and since in a few cases rudimentary females and
males bred together have produced daughters as well as
sons, the hypothesis of prematuration that I suggested in
1912 is not the correct explanation of the sterility of the
females of the rudimentary winged stock mated to rudi-
mentary males. Moreover, since many of the females
tested, especially in the later experiments, were F,’s ex-
tracted through other fertile stocks, the sterility can not
be supposed to be due to any additional peculiarity that
has appeared in the rudimentary stock, but must be one of
the attributes of the factor for rudimentary itself.

NOTES AND LITERATURE

DIPTERA FROM THE SEYCHELLES.—An important work has just
come to hand* in which Mr. ©. G. Lamb describes the muscoid
flies of several groups, collected in 1905 in the Seychelles and
other islands of the Indian Ocean by the expedition under the
leadership of Mr. J. Stanley Gardiner. Considering the remote-
ness of the regions explored, and the total lack of general inter-
est in the taxonomy of these small Diptera, one would not at first
see any reason for calling special attention to this paper. When,
however, we think of the wonderful results attained by Morgan
and those working with him through the intensive study of
Drosophila, Mr. Lamb’s work, dealing in part with this very
group, gains new significance and suggests many strains of
thought. Some years ago, after looking at a number of the strange
Drosophila mutants in Professor Morgan’s laboratory, I raised
the obvious question: ‘‘How do you know that all this has any-
thing to do with the evolution of species?’’, and Professor Mor-
gan replied that he did not know it. The real answer to my
question must be found by investigating the allied species, to see
whether they do in fact differ in ways at all paralleled by Mor-
gan’s mutants, or likely to have arisen in similar fashion. The
important cytological paper by Metz, just published, represents
one way of attacking this problem; the taxonomic results of
Lamb, based on twenty species, afford us another.

At the outset, we are struck by the fact that the Seychelles
Drosophila species have been modified in a great variety of dif-
ferent ways, in several cases so remarkably that Lamb hesitates
whether to base new generic or subgeneric names upon them.
There is nothing like orthogenesis, apparently. Here is a list of
some of the more noticeable modifications:

1. Costa extending to third vein.

2. Remarkably constricted waist and sian wings. (Compare
Morgan’s short winged forms

3. A remarkable slit on the eet the end provided with spines
and bristles.

4. Remarkable transverse or oblique eyes.

5. Remarkable spines on front tarsus.

1 Trans. Linnean Sec. London, Zoology, Vol. XVI, Part 4, July, 1914.

251

252 THE AMERICAN NATURALIST [Voin XLIX

6. Curious curled hairs on front legs.

T. Marmorated thorax.

8. Entirely black, except brownish antennæ and lighter face.

All the species are new except two, one of these being the D.
ampelophila of Morgan’s experiments, the prior name for which
is D. melanogaster Meigen. One of the new species, D. similis, is
based on males differing from D. melanogaster in lacking the
large combs on the front legs, having instead only minute combs
which require a high magnification to be seen. In other respects
the flies are almost exactly as in melanogaster, and in the female
sex it is practically impossible to distinguish between the species.
There are in addition five females resembling melanogaster and
similis, but differing in a detail of the venation. May we not
suppose that D. melanogaster was introduced into the Seychelles
by man and that D. similis and the females (left unnamed)
with peculiar venation have arisen from it e mutation since
that time? T. D. A. COCKERELL

STERILITY IN A SPECIES CROSS

PROFESSOR J. A. DETLEFSEN, of the University of Illinois, has
recently published an interesting paper entitled ‘‘Genetie Stud-
ies on a Cavy Species Cross.’?! Several wild cavies from Brazil
(Cavia rufescens) were crossed with domestic guinea-pigs (C.
porcellus) and a study of the hybrid offspring was continued
for seven generations. The experiments were begun in 1903 by
Professor Castle who turned them over in 1909 to Professor
Detlefsen. They were carried on at Harvard and at the Bussey
Institution. The paper is divided into three parts; the first two
treat respectively of the genetics of color and coat factors, and
growth and morphological characters; the third and most im-
portant deals with a study of the sterility of hybrids.

Cavia rufescens differs from the guinea-pig in several charac-
ters. In size it is about half as large as the guinea-pig. It has
an agouti (gray) coat, but the animal has a darker appearance
than tame agouti guinea-pigs, owing to the yellow bands in the
ticked hair being much reduced. The belly of the wild species
varies from a light yellow to a slightly ticked condition. In
agouti guinea-pigs the belly hair is usually yellow with a dark
base, but never ticked.

Of the wild stock, only males were used in crossing, on account
of the difference in size. All F, males from such crosses were
1 Carnegie Inst. Publication No. 205, 1914.

No. 580] NOTES AND LITERATURE 253

found to be sterile but the F, females were fertile, and the line
was continued only by crossing these females with guinea-pig
males, Thus with each succeeding generation there was a reduc-
tion in the amount of wild blood and the author refers to his
hybrids as one half wild, or F,, one fourth wild or F,, ete.

As to the inheritance of coat color, the tame agouti coat is
dominant over the wild agouti. These two types segregate and
are allelomorphie to each other. Each is also allelomorphie to
its absence. Detlefsen finds that there is a constant relation be-
tween back color and belly color, but this condition is not due to
separate factors because they can not be transmitted independ-
ently. The two types of tame and wild agouti, he thinks, are
perhaps comparable with the types of gray mice described by
Cuénot and Morgan, viz., gray-bellied and light- (or white-)
bellied agouti. In crosses with non-agouti, the wild agouti type
is dominant over black and over red, as in domestic guinea-pigs.
After back-crossing the wild agouti colored hybrids with non-
agouti for several generations, it is found that the agouti factor
is modified, producing a darker coat, in some cases almost black,
the ticking being faintly seen only on the belly. Roughness of
coat is imperfectly dominant over smooth coat. In later genera-
tions it regains its dominance.

In respect to growth and vigor, the F, hybrids were heavier at
all ages than the guinea-pig parents, and were more vigorous
than either parent. These F, individuals were crossed with
guinea-pigs and gave young which were smaller than the F, ani-
mals in every way, and in size resembled the guinea-pig parent.
The variability of the wild stock in weight and vigor is unknown,
but guinea-pigs are remarkably uniform in both respects. In
morphological characters, the M-shaped nasal-frontal suture of
the wild species is dominant over the truncated nasal suture of,
the domestic form. -The truncated suture reappears in the sec-
ond generation but does not breed true. As to skull shape, the
wild has a pointed head and the tame species a round one. In
F, a blending oceurs, and in later generations the wild pointed
head disappears. In the wild cavy a narrow indenture is pres-
ent on the outer surface of the last upper molar—a character
held to be of much importance by systematists. In F, this in-
denture showed to a slight degree, and in subsequent genera-
tions was lost.

The section of the paper dealing with sterility is of great in-
terest. No previous investigations on sterility in animals have

254 THE AMERICAN NATURALIST [Vou. XLIX

been made on such a scale as the experiments reported by Det-
lefsen. The causes of sterility are very obscure and but little
understood. A change of environment and consequent lack of
exercise or difference of diet may be a contributing cause of ster-
ility in birds and wild animals in captivity, but none of these
influences were operative in Detlefsen’s experiments, because the
wild cavies from Brazil bred inter se under laboratory condi-
tions. Sterility is frequent in hybrids of species not closely re-
lated, and it is an axiom among biologists that crosses between
different species or genera produce sterile hybrids in one or
both sexes. As stated above, all F, hybrid males from a cross
between a wild cavy male and a guinea-pig female were sterile.
However, the F, females were fertile and were crossed back to
guinea-pig males. These likewise gave in F, sterile males and
fertile females. Repeated back crosses of fertile females with
guinea-pig males produced fertile males in increasing numbers
with each generation.

In order to test the fertility of the hybrid iia two methods
were used: (a) breeding tests; (b) microscopic examination of
spermatozoa obtained by transecting several tubules from the
epididymis on one side of the animal. Such an individual could
be bred subsequently. In all, 483 males were tested by one or
both methods; 50 by breeding alone; 331 by microscopie exami-
nation alone; and 102 by both methods. The following table,
giving the results of combined microscopic and breeding tests,
indicates the value of microscopic examinations in determining
the fertility of males:

; | Breeding Test
Microscopic Test No. |
| Sterile Fertile
* Without spermatozoa............... 23 23
With immotile sperm. 22 6 oo -isi 11 11
With few motile sperm. ............. 10 9 1
With many motile sperm............ 58 14 44
ee SES agen. Gear

It will be noted that of the 58 males with many motile sperma-
tozoa, 14 proved to be sterile upon breeding. Among these
males, Detlefsen attributes the sterility of 9 to external causes,
without specifying them, but he can assign no reason for the im-
potency of the remaining 5. He therefore concludes that

the number and motility of the sperm are not the only essentials for a
real fertility, inasmuch as real fertility in the last analysis may mean

No. 580] NOTES AND LITERATURE 255

the capacity to fertilize eggs and sire young. There are further reasons
for concluding that the motile sperm of the hybrid males may be physio-
logically different from those of the normal guinea-pig, for it often re-
quired much more time to obtain young from the hybrid males and the
litters were unexpectedly small.

It may be added that sterility was not due to the absence of sec-
ondary sexual characters, for all the males were normal in this
respect.

The percentage of fertile males in each generation from the
back crosses above described, was as follows:

F, F, F; F, 5 F, F, -

1/2 Wild 1/4 Wild 1/8 Wild 1/16 Wild 1/32 Wild- 1/64 Wild 1/128 Wild
14,29 33.32 60.67 69.39

In the generations having the more dilute wild blood the percent-
age of fertile males increased. The author holds that some dis-
turbance occurred in the gametogenesis of the males, subsequent
to hybridization. The females were normal, but they transmitted
this disturbing element to their sons. However, by back-cross-
ing with guinea-pigs this peculiar quality was segregated out.
It is evident that if the heredity of fertility and sterility in this
case is Mendelian, it is due not to one or two allelomorphic pairs
of factors, but to multiple factors. A table is given of the per-
centages of ultimate recessives expected in back crosses on the
basis of various numbers of factors involved. The series for 8
factors, given below, approaches most nearly the percentage of
fertile males obtained (see above) :

F F

1 i i oA F: P, F; 6 F.
0.0 0.39 10.1 34.36 59.67 77.58

7
88.16
The application is somewhat misleading, as the author states,
since the probable errors are not given. In each generation the ©
probable error would have to be calculated on the supposition
that the females of the preceding generation were normally dis-
. tributed, otherwise one would have to take into account the error
of all the preceding generations. It is improbable that the fe-
males of any generation, except F,, were normally distributed.

He concludes that fertility acts as a very complex recessive
character, the results being in accord with the expectations if a
number of dominant factors for sterility were present. After
these dominant factors were eliminated, there would be produced
a fertile recessive type.

Byron B. Horton

256 THE AMERICAN NATURALIST [Vou. XLIX

FLOWER PIGMENTS

RECENT researches by Wheldale and Bassett! have shown that
there are four flower pigments, i. e., ivory, yellow, red and
magenta, in Antirrhinum majus, and that these, in various com-
binations and in different states of concentration and dilution,
are responsible for all the color varieties. The ivory and yellow
pigments have been identified with apigenin and luteolin respec-
tively, i. e., members of the class of soluble yellow plant pigments
containing carbon, hydrogen and oxygen; the red and magenta
pigments are anthocyanins. The yellow, red and magenta pig-
ments occur only in the epidermis of the corolla, but ivory is
present in the inner tissues. The pigments are present in the
plant as glucosides, that is combined with sugar. For prepara-
tion, the flowers are boiled with water, the pigments precipitated
as insoluble lead salts from the filtered solution by adding lead
acetate. The lead salts are filtered off and decomposed with di-
lute sulphuric acid which forms insoluble lead sulphate and sets
free the pigment again in dilute acid solution. These solutions
are then boiled for several hours, whereby the sugar is split off
from the glucoside and the free pigment, which is less soluble,
separates out and is filtered off. The anthocyanins are separated
from the yellow (flavone) pigments by extracting the latter with
ether in which the anthocyanins are insoluble. The red and
magenta pigments have been purified and analyzed and shown to
contain carbon, hydrogen and oxygen only but a higher percent-
age of oxygen than the flavones. Determination of the molec-
ular weights of the anthocyanins also indicates that their
molecules are larger than those of the flavone pigments. Hence
if the anthocyanins are derived from the flavones, it seems likely
that the process is one of oxidation accompanied by condensation
of two or more flavone molecules or the union of flavone mole-
cules with other allied compounds in the plant. It is possible that
the factors for red and magenta color will come to be expressed in
terms of chemical substances which condense with the flavones to
from the larger molecules of the anthocyanins.

M. W.

1 Wheldale, M., ‘‘The Flower Pigments of Antirrhinum majus. 1. Method
of Preparation,’’ Biochem. Jour., 1913, 7, 87. Wheldale, M., and Bassett,
H. Ll., ‘‘The Flower Pigments of Antirrhinum majus. 2. The Pale Yellow
ór Ivory Pigment,’’ Biochem. Jour., 1913, 7, 441; ‘‘The Chemical Interpre-
tation of Some Mendelian Factors for Piwutesten? Proc. Roy, Soe., 1914,
B, 87, 300; ‘‘The Flower Pigments of Antirrhinum majus. 3. The Red and
_ Magenta Pigments,’’ Biochem. Jour., 1914, 8, 204.

VOL. XLIX, NO, 58 MAY, 1915

THE

AMERICAN _ ee

The American Naturalist

S intended for publication and books, etc., intended for review should be
k.

MS
sent to the Editor of THE AMER
Short articles containing sum

RICAN NATURA
maries of research work bearing on the

LIST, Garrison- -on-Hudson, New Yor

problems of aah evolution are especially welcome, and will be given preference

in stag

undrea reprints of ce pie are supplied to authors free of charge.

ionur h ‘reprints will be supplied a

tions and advertisements shouid be sent to the publishers.
ear. Foreign postage is fifty cents and

subscription price is four

Canadian postage twenty- ae cents additional.

The

The ch ra for single copies is

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

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of
Congress of March 3, 1879.

F ALE
ARCTIC, TER and GREENLAND
DS’ SKINS,
Well GS Low Prices
Particulars of
G. DINESEN, Bird Collector
Husavik, North Iceland, Via Leidle, England

a Arone HISTORY SPECIMENS
eot Condition and Lowest Prices
PRE Sir Skins, Oology, Ey Marine
Animals and others. Catalogue free. Correspond-
ence solicited
T. FUKAI, Naturalist,
Konosu, Saitama, Japan

WANTED

BIRDS OF AMERICA by J. J. Audubon, 7 volumes,
please report cash price, stating condition, binding and dates
of volumes.
F. C. HARRIS,
Box 2244 Boston, Massachusetts

The University of Chicago

Marine Biological Laboratory
Woods Hole, Mass.

a o0l08f»
geara Enbrsbey, research Jey ee Bee
tee

E
: ?
yE

INSTRUCTION
July—August

Bhios
and

SUPPLY Animals and plants,
DEPARTMENT iriam agd in, 207 yas
Open the Entire Year snimals and of Algae, Fungi,
worts Mosses

ed TO er Gu

1

sent on keaton State which

e aai For price 1ste an and
regarding material,

GEO, M. GRAY, Curator, Woods Hole, Mass

THE
AMERICAN NATURALIST

Vou. XLIX May, 1915 No. 581

ON THE NATURE OF THE CONDITIONS WHICH
DETERMINE OR PREVENT THE ENTRANCE
OF THE SPERMATOZOON INTO THE EGG

JACQUES LOEB
Tue ROCKEFELLER INSTITUTE ror MEDICAL RESEARCH, New YORK

I

Tae well-known fact that a spermatozoon can no longer
enter an egg after it is once fertilized raises the question
whether this is due to the changes necessarily connected
with development; or whether development of an egg can
take place without the existence of such a block. We are
in possession of facts speaking in favor of the second
view. Thus the writer has shown that if the eggs of
Strongylocentrotus purpuratus or Arbacia are induced to
develop by the methods of artificial parthenogenesis a
spermatozoon can enter the egg or an individual blasto-
mere of a segmenting egg, while the latter is in the full
process of development. This leaves no doubt that the
block caused by the entrance of a hale into an
egg for the entrance of further spermatozoa must be due
to a change not necessarily identical with that inducing
the development of the egg. \

A second group of observations made by the author
deals with the phenomena of specificity and these prove
that the block which an egg offers to heterogeneous spe
is rapidly reversible and confined to the surface of the e
or the spermatozoon or both. In the case of the egg o
purpuratus and sperm of Asterias (and many similar in- \

257 `

258 THE AMERICAN NATURALIST [ Vou. XLIX

stances) the specific block can be overcome if we slightly
increase the alkalinity of the sea water. The spermato-
zoon can only enter the foreign egg while both sperm and
egg are in the hyperalkaline sea water, whereas if the egg
and sperm are treated separately with hyperalkaline so-
lution (no matter how long) and put together in a suffi-
ciently large quantity of normal sea water no egg can be
fertilized’ while fertilization will take place as soon as
the hyperalkalinity is restored. This shows that the
change (brought about by the hyperalkaline sea water)
which makes the fertilization possible is rapidly reversible,
as we should expect it to be if it consisted merely in a
physical change at the surface of the egg. To this series
of facts, others might be added which point in the same
direction. In this paper we intend to discuss a little
more fully the various conditions which block or favor
the entrance of a spermatozoon into an egg, in order to
form an idea of the nature of the forces which control
these phenomena.
EL

1. When the unfertilized eggs of S. purpuratus are
treated for two hours with hypertonic sea water (50 c.c.
sea water +8 c.c. 24 m NaCl or Ringer solution) the
eggs of certain females will develop into blastulæ, gas-
trulæ and plutei, while the eggs of other females can not
be caused to develop in this way. These individual dif-
ferences coincide possibly with those observed by the
writer in regard to spontaneous membrane formation in
the eggs of different females? and it is possible that only
the eggs of such females of purpuratus can be induced to
form larvæ through a mere treatment with hypertonic sea
water in which the latter can induce the cortical changes
underlying the membrane formation. Whatever the na-
ture of the individual difference may be, purpuratus eggs

1 The large quantity of sea water is necessary so that the hyperalkaline
sea water at the surface of the egg and sperm can diffuse away before both
gametes ger in contact.

2 Loeb, Arch. f. Entweklngsmech., XXXVI, 626, 1913; ‘‘ Artificial Par-
thenogenesis and Fertilization,’’ Chicago, 1913, p. 219.

No. 581] ENTRANCE OF THE SPERMATOZOON 259

which have been induced to develop into larve by a hyper-
tonic solution can be fertilized with sperm while they are
in the process of segmentation. When such eggs are in
the two-, four-, eight-, or sixteen-cell (and possibly also
later) stages the sperm can enter into one or more blasto-
meres of such an egg and this entrance betrays itself by
a distinct and clear membrane formation around each
blastomere.? While the segmenting eggs which were not
fertilized with sperm develop into larvæ, those into which
Sperm enters perish very rapidly. This simple and
rather striking experiment which can easily be performed
in the eggs of Strongylocentrotus, where the membrane
formation around a single blastomere can be clearly recog-
nized, shows that the process of development in a fer-
tilized egg in itself can not be responsible for the block
caused by fertilization. It looks as if the entrance of a
spermatozoon into the mature egg, independently of the
developmental changes it induces in the egg, causes some
physical or physico-chemical change (of the surface of
the egg?) which renders the subsequent entrance of a
Spermatozoon impossible.

2. With the eggs of most females of purpuratus the
treatment with a hypertonic solution does not lead to a
development into larve, but only to the first segmentation
Stages in a limited number of eggs (provided that the
eggs have been exposed to the solution the proper period
of time). Such blastomeres afterwards go into a resting
Stage. If one waits long enough, until there is no doubt
left that the blastomeres have reached a resting condition
and will divide no further, and if one then adds sperm,
the individual blastomeres can again be fertilized, which
is indicated by a membrane formation around each indi-
vidual blastomere and the sul t of such
blastomeres into swimming larve. The fact that each
individual blastomere in this ease is fertilized independ-

3 Loeb, ‘‘ Artificial Parthenogenesis and Fertilization,’’ p. 240; Arch. f.
Bniwekingamech, XXIII, 479, 1907. aS

4Loeb, Arch. f. Entweklngsmech., XXIII, 479, 1907; ‘‘ Artificial Par-
Decree and Fertilization,’’ p. 237.

260 THE AMERICAN NATURALIST [Vou. XLIX

ently of its neighbors suggests that there is no protoplas-
mic connection between the neighboring blastomeres;
otherwise the entrance of a spermatozoon into one should
cause its neighbors also to form a fertilization membrane,
which does not happen.

All these facts show that the changes underlying devel-
opment do not necessarily prevent the entrance of a sper-
matozoon into an egg fertilized by sperm.

3. Development can be initiated in an unfertilized egg
by causing a membrane formation by a fatty acid. Eggs
after such an artificial membrane formation perish as a
rule rapidly at room temperature (if no second treatment
is given them) but they may segment if kept at a low tem-
perature. The eggs are usually put after treatment with
the butyric acid into normal sea water in which they form
a membrane. This membrane is different in the eggs of
different species of sea urchins. In the egg of S. pur-
puratus the membrane is tough and entirely impermeable
to the spermatozoon. When we add sperm to such eggs
with a butyric acid membrane they behave exactly as if no
sperm had been added, they all perish rapidly (at room
temperature). The question arose, if a spermatozoon
could still enter the egg of purpuratus after membrane
formation, provided the membrane could be destroyed.
This can be done in a certain percentage of the eggs of
purpuratus by shaking them after artificial membrane
formation; the number of eggs whose membrane is torn
varies in different experiments owing probably to dif-
ferences in the thickness and toughness of the membrane.
Even if the membrane is torn the edges may come close
together again so that the opening often is closed
again and no spermatozoon can go through. Kupelwieser
and the writer performed this experiment on the eggs of
purpuratus and it was found that such eggs with torn
membranes were fertilized upon the addition of sperm
and developed normally; while the eggs whose mem-
branes were intact all perished.®

5 Loeb, ‘‘ Artificial Parthenogenesis and Fertilization,’ p. 234.

No.581] ENTRANCE OF THE SPERMATOZOON 261

The writer repeated this experiment last winter with
the same result. He found that eggs with torn mem-
branes when subsequently fertilized with sperm did not
form any new membranes as he had stated before. It is
possible that he mistook at that time the new hyaline
membrane which forms around the egg after membrane
formation and fertilization for a new fertilization mem-
brane.

It is not necessary that these eggs be fertilized im-
mediately after the artificial membrane formation, the ex-
periment succeeds also after some time (one hour or
more) ; only with this difference that the eggs perish very
rapidly after the membrane formation if they receive no
second treatment. In order to avoid this difficulty the
writer last winter proceeded as follows: Artificial mem-
brane formation was produced in the eggs of a purpuratus
and all eggs had formed perfect membranes. One con-
trol was kept and the rest were shaken. These were di-
vided into three lots, one served as a control; the eggs of
the latter all perished as fast as the eggs of the first control
(which were not shaken). The second lot were fertilized
after about one half hour after membrane formation.
Twenty per cent. of these eggs developed into normal
larve, the rest perished. The percentage of developing
eggs corresponded roughly with the percentage of eggs
whose membrane was torn. The third lot of the shaken
eggs was put overnight into 50 c.c. sea water +7 drops
of 1/10 per cent. KCN, to prevent the disintegration of
these eggs. The next morning (sixteen hours after the
membrane formation) the eggs of Lot 3 were transferred
into normal sea water and divided into two lots, one was
fertilized with sperm, the other was kept as a control.
About twenty per cent. of the eggs which were fertilized
began to segment, but many in an abnormal way and
none developed into larve. Of the second lot to which
no sperm was added also a few began to segment. As
the writer has shown in former experiments, the eggs of
Strongylocentrotus can be caused to develop after artifi-

262 THE AMERICAN NATURALIST (Vou. XLIX

cial membrane formation if they are either treated for a
short time with a hypertonic solution or if for a longer
period the oxidations are suppressed in them by lack of
oxygen or the addition of cyanide. There is therefore no
doubt that the eggs of purpuratus in which the artificial
membrane formation has been induced by butyric acid
ean be fertilized subsequently with sperm.

4. The treatment of the eggs of Arbacia with butyric
acid leads to the formation of a membrane which varies
considerably in the eggs of the same female. Some eggs
have a thin membrane which is permeable to the sper-
matozoon, others have a tough fertilization membrane
which is as impervious to the spermatozoon as the regular
fertilization membrane. The percentage of the eggs with
membranes permeable for sperm varies very much in dif-
ferent experiments, according to the material and accord-
ing to the external conditions. If this is kept in mind it
is easily understood that the number of Arbacia eggs
which can be fertilized after they have been treated with
butyric acid differs in different experiments. Since the
membrane called forth by butyric acid is not always
plainly visible, it is a prerequisite that always one set of
such eggs should be set aside as controls to ascertain
whether or not all the eggs disintegrate rapidly (if no
second treatment is given to them). Only if they all dis-
integrate rapidly have we any guarantee that in all of
them the membrane formation has been effective. The
former experiments of the writer show that such eggs
can be fertilized by sperm; in fact they show that while
the unfertilized eggs disintegrate rapidly after the in-
ducement of the membrane formation with butyric acid,
the subsequent fertilization of such eggs by sperm saves
their lives and makes them develop.®

Il
1. It is a well-known fact that most eggs can only be
fertilized by sperm of their own or a closely related
species. The writer thought that in order to obtain light
€ Loeb, Arch. f. Entweklngsmech., XXXVIII, 416, 1914.

No. 581] ENTRANCE OF THE SPERMATOZOON 263

on the nature of the block to the entrance of hetero-
geneous sperm it was necessary first to find the means by
which this block could be overcome. He succeeded in
showing that the egg of the sea urchin S. purpuratus can
be fertilized by the sperm of starfish, brittle stars, and
holothurians in sea water (or other balanced solutions)
if their alkalinity was a trifle higher than that of ordinary
sea water (e. g., in a solution of 50 c.c. sea water + 0.6 c.c.
N/10 NaOH). Godlewski’ succeeded by the same
method in the fertilization of the egg of the sea urchin
with the sperm of crinoids

The most important fact found out in this connection
was the following, namely, that the fertilization of the
egg of purpuratus by the sperm of Asterias only takes
place while both eggs and sperm are in this hyperalkaline
solution. If eggs and sperm are put into these solutions
separately and if then from time to time sperm and eggs
so treated are transferred into normal sea water, as a
rule not a single egg is fertilized; while with the same
material when eggs and sperm are together in the hyper-
alkaline solution as many as 100 per cent. of the eggs may
be fertilized. The effect of the alkali is, therefore, rap-
idly reversible; the eggs when put from the hyperal-
kaline sea water free from sperm into the normal sea
water containing very motile sperm of Asterias can not
be fertilized; when put back into hyperalkaline sea water
containing Asterias sperm they will be fertilized rapidly.

This rapid reversibility of the effect of the NaOH in-
dicates that it must be confined to the surface of the egg
and the spermatozoon or both; and this is corroborated
by the fact that the NaOH does not enter the cells. One
of the forces which determine the entrance of the sper-
matozoon into the egg may be surface tension and the
phenomenon of the entrance may be comparable or pos-
sibly identical with the phenomenon of phagocytosis.

Godlewski mentioned that he occasionally observed a

7 Loeb, Pfliiger’s Arch, IC, 323, 1903; CIV, 325, 1904; Arch. f.
Entweklngsmech., XL, 310, 1914; Science, N. S., XL, 316, 1914.

8 Godlewski, Arch. f. Entwekingsmech., XX, 579, 1906.

264 THE AMERICAN NATURALIST [Von. XLIX

fertilization of the egg of the sea urchin with the sperm
of a crinoid in normal sea water after both had been
treated with hyperalkaline sea water separately. This
observation is correct but finds its explanation in the as-
sumption that in such cases the hyperalkaline sea water
had not had time to diffuse from the jelly of the egg or
from the surface of the egg protoplasm by the time the
‘spermatozoon came in contact with it. In order to test
this view the writer treated the eggs of purpwratus with
a hyperalkaline solution of greater than the optimal con-
centration while the sperm was treated separately with
the optimal concentration (50 c.c. sea water + 0.6 c.c.
N/10 NaOH) and then both were mixed in a little sea
water in a watch glass. In such a case a large number
of eggs were fertilized, but while the fertilization occurred
nominally in normal sea water it really occurred in a
layer of hyperalkaline sea water surrounding the proto-
plasm of the egg.

The conclusion from these experiments is that the block
to the entrance of the spermatozoon of Asterias into the
egg of purpuratus is of a rapidly reversible character,
consisting in some alteration of a physical property of
the surface. On this assumption the factor of specificity
consists of an agency which affects these properties of
the surface of the egg in the same sense as the increase in
the concentration of the alkali. It should be added that
the writer observed also that an increase of the concentra-
tion of Ca in the sea water acts in the same sense as an
increase in the alkalinity; and that if the concentration
of Ca is increased the increase of NaOH may be less than
is necessary otherwise.

2. If the idea was correct that the factor of specificity
contained in the spermatozoon affected only the forces
acting at the surface of the egg; and that the lack of this
factor could be replaced by a rise in the alkalinity of the
sea water, it was to be expected that the reverse should
also be possible: namely, that a change in alkalinity or
the constitution of the surrounding medium should pro-

No. 581] ENTRANCE OF THE SPERMATOZOON 265

duce a reversible block to the spermatozoa of the same
species. That means, it should be possible to find solu-
tions in which the egg does not suffer for a long time, in
which the sperm lives for a long time, and in which the
sperm of the same species is intensely active and attacks
the egg with the greatest eagerness and yet is not able to
enter; while if the medium is but slightly changed the
sperm enters the egg at once. The writer carried out
such experiments a year ago in Pacific Grove and last
summer in Woods Hole and found this to be true.®

For the purpose of these experiments the ovaries and
testes of the sea urchins were not put into sea water but
into a pure m/2 NaCl solution (after several washings in
such a solution) and kept in such a solution. Several
drops of sperm and one drop of eggs were in one experi-
ment put into 2.5 c.c. of a neutral mixture of m/2 NaCl
and 3/8 m MgCl, in the proportion in which these two
salts exist in the sea water. In such a neutral solution
no egg of Arbacia or of purpuratus is fertilized no matter
how long they remain in the solution, although the sperm
is very active. If the eggs and sperm are transferred
into the same solution which contains in addition 1 drop
of a N/100 solution of NaOH (or NH,, or benzylamine,
or butylamine) or 8 drops of m/100 NaHCO,, most and
often practically all the eggs at once form fertilization
membranes and begin to segment at the proper time.

The same result can be obtained if the eggs are trans-
ferred into a neutral mixture of NaCl + MgCl, + CaCl,
(in the proportion in which these salts exist in the sea
water) or into a neutral mixture of NaCl -+ MgCl, +
CaCl, + KCl. In such a neutral mixture the eggs form
fertilization membranes and begin to segment.

The eggs will not be fertilized if transferred into a
neutral solution of NaCl or of NaCl + KCl.

It is, therefore, obvious that if we diminish the alkalin-
ity of the solution surrounding the egg and if we deprive
this solution of CaCl, we establish the same reversible

® Loeb, Science, N. S., XL, 316, 1914.

266 THE AMERICAN NATURALIST [ Von. XLIX

block to the entrance of the spermatozoon of Arbacia into
the egg of the same species as exists for the entrance of
the sperm of starfish into the egg of purpuratus in normal
sea water.

Another form of the experiment may be mentioned.
When we put sperm and eggs of Arbacia (which had been
washed in an m/2 NaCl solution) into a neutral mixture
of NaCl+ KCl no egg can be fertilized although the
sperm may be so active and concentrated that the eggs
roll around in the solution and the chorion (the jelly sur-
rounding the egg) may be filled with spermatozoa. In <
one experiment the eggs and sperm of Arbacia were kept
overnight in watch glasses containing 2.5 c.c. of this mix-
ture of neutral NaCl + KCl. The next morning all the
eggs were intact and not a single one was fertilized. At
that time 20 drops of sea water were added to the mixture
and instantly fertilization membranes were formed and
practically all the eggs segmented.’

It can be shown that in this experiment the sea water
added two important substances, Ca and NaOH. If
NaOH alone is added to the mixture of NaCl + KCl, as a
rule no egg is fertilized or only a few; if CaCl, is added
to a neutral mixture of NaCl+ KCl a number of eggs
are fertilized. If both CaCl, and NaOH are added in the
proper proportion as a rule all the eggs are fertilized.

It is perhaps important to call attention to the fact
that if eggs of Arbacia are fertilized in sea water and if
after repeated washings in a mixture of NaCl + KCl or
of NaCl+ MgCl, they are put into these solutions they
will segment repeatedly in these solutions, thus showing
that the eggs were really not fertilized in these two so-
lutions in the above-mentioned experiments.

The striking fact is again that the block created by the

ves This experiment was carried out with different concentrations of sperm
and it was found that only in the dishes where the concentration of sperm
was sufficiently high were all the eggs fertilized upon the addition of sea
water. This is perfectly natural as the majority of spermatozoa die grad-

ually (as do also the eggs) and hence enough spermatozoa will only be alive
the next day if the concentration of sperm was not too low.

è

No. 581] ENTRANCE OF THE SPERMATOZOON 267

lack of CaCl, or NaOH or both to the entrance of the
spermatozoon is removed immediately after these sub-
stances are added. The block must be due merely to a
change in the physical condition of the surface (which
may be based on a rapidly reversible chemical reaction).

In these experiments the NaCl can not be replaced by
isotonic sugar solutions. The same fact was found by
the writer to be true for heterogeneous hybridization.

It is of importance to call attention to the fact that
the abolition of the block in the case of heterogeneous
hybridization depends upon the same substances, CaCl,
and NaOH (or some other alkali), which make normal
fertilization possible. The influence of electrolytes on
the fertilization of the egg of purpuratus by the sperm of
Asterias is parallel to the influence of the same electro-
lytes on the fertilization of the same egg by the sperm of
purpuratus; only the concentrations differ, and always
in the same sense. The forces at work are, therefore,
apparently the same in both cases; but we can only ex-
press surmises as to their nature. The rôle of salts as
well as the rapid reversibility indicate that they are
forces situated at the surface of the egg and the sperma-
tozoon or both. In the first place we may think of sur-
face tension conditions and in this respect it is possible
that the entrance of the spermatozoon into the egg may
be determined by such forces in a way similar to the
process of phagocytosis. In the second place it may be
that previous to the action of surface tension forces an
alteration in the degree of fluidity of the egg surface may
be required (e. g., that physical change which finds its
expression in the formation of the fertilization cone).
Thirdly, it may be possible that before the surface ten-
sion forces can act the spermatozoon must agglutinate
with the egg surface and that this agglutination is de-
termined by certain specific substances or by certain salts
(CaCl, and NaOH) or by both.

Brief mention should be made of the block discovered
by Godlewski’! to the entrance of a spermatozoon into

268 THE AMERICAN NATURALIST — [Vou. XLIX

the egg if the sperm of the same species is mixed with the
sperm or the blood of a species widely apart. If, for
instance, the sperm of a sea urchin is mixed with the ~
sperm of certain annelids (Chetopterus) or molluses and
if after some time the eggs of the same sea urchin are
added to the mixture of the two kinds of sperm no egg is
fertilized. If the solution is, however, subsequently
diluted with sea water or if the egg that was in this
mixture is washed in sea water, the same sperm mixture
in which the egg previously remained unfertilized will
now fertilize the egg. From these and similar observa-
tions Herlant!? draws the conclusion that the block
existed at the surface of the egg, inasmuch as a reaction
product of the two types of sperm is formed after some
time which alters the surface of the egg and thereby pre-
vents the sperm from entering. This view is not only
supported by all the experiments but also by the observa-
tion of the writer that foreign sperm or blood is able to
cause after some time a real agglutination if mixed with
the sperm of a sea urchin or a starfish.12 We can imagine
that the precipitate forms a film around the egg and
acts as a block which can be removed mechanically by
washing.

It is not impossible that the block which exists in the
fertilized egg is due also to an alteration of the physical
character of the surface of the egg which in this case is,
however, induced from within the egg by changes caused
by the entrance of the spermatozoon, which, however, are
not necessarily identical with those causing development
as was shown by the facts in the second chapter.

EY.

We will now turn to the question whether the motility
of the spermatozoon plays no other rôle than to bring the
spermatozoon so close to the surface of the egg that sur-
face tension phenomena can engulf the spermatozoon
into the egg. It is easy to show that if the spermatozoa

11 Godlewski, Arch. f. Entweklngsmech., XX XIII, 196, 1911.

12 Herlant, Anat. Anzeiger, XLII, 563, 1912.

“hes Loeb, Jour, Exper. Zool., XVII, 123, 1914.

No.581] ENTRANCE OF THE SPERMATOZOON 269

of purpuratus are immobilized by NaCN no egg of the
same species can be fertilized, no matter how concen-
trated the sperm; while the same sperm when it revives
from the effect of NaCN fertilizes the same eggs at
once. This meets with the possible objection that the
motility of the sperm might be only necessary to allow
the latter to penetrate the jelly surrounding the egg
protoplasm. In order to test this objection the writer
freed the eggs of purpuratus from this jelly by treating
them for two minutes in a mixture of 50 c.c. sea water
+ 3 c.c. N/10 HCl in which all the jelly is dissolved. The
eggs were washed afterwards in sea water and it was
found that if sperm was added practically all were fertil-
ized. The writer put such eggs with sperm which was
immobilized by NaCN. The eggs and the sperm were
squirted together with a pipette in order to bring about a
close contact. No matter how concentrated the sperm
was, not a single egg was ever fertilized. As soon as
the spermatozoa recovered and showed only a slight
degree of motility fertilization became possible. This
leaves no doubt that the motility of the sperm is one
of the forces required to bring the spermatozoon into
the egg.

That motility is not the only force was already indi-
cated by the previous chapter which made it clear that
even if the sperm is active it can not enter the egg unless
certain physical conditions at the phase boundaries of
egg, spermatozoon and surrounding solution were right.
In order to leave no doubt about this fact the following
experiments may be quoted. If we put NaCl sperm**
of purpuratus or of Arbacia into a neutral mixture of
NaCl + KCl containing eggs of the same species the —
sperm will sooner or later become very active. Yet not
a single egg is fertilized. If we make the solution
slightly alkaline the sperm becomes at once extremely
active yet with a few exceptions no egg is fertilized; while
much less active sperm will fertilize all the eggs if CaCl,
is added. The second fact is this: that the most active

14 Sperm from testicles washed in m/2 NaCl and kept in such a solution.

270 THE AMERICAN NATURALIST [Vou. XLIX

sperm of Asterias will not fertilize the eggs of purpu-
ratus in sea water while it will do so in hyperalkaline sea
water (50 c.c. sea water + 0.6 c.c. N/10 NaOH).

We, therefore, arrive at the conclusion that aside from
the physical conditions at the surface of the egg and the
spermatozoon the impact of the spermatozoon against
the egg is a prerequisite for the process of fertilization.

von Dungern was, as far as the writer is aware, the
first to call attention to the fact that the egg itself causes
resting spermatozoa to become active,'® but curiously
enough he tried to show that only foreign sperm is
‘¢ stimulated ’’ in this way by the egg (which is, as F. Lillie
pointed out, not correct) and v. Dungern tried to explain
on this basis why it was not possible to fertilize the egg
of the sea urchin with the sperm of the starfish which
had at that time not yet been accomplished.

von Dungern noticed that the egg of the sea urchin
‘‘stimulates’’ the spermatozoon of starfish to greater
action and he concluded that since ‘‘stimulation’’ accord-
ing to Jennings causes ‘‘motor reaction’’ whereby the
direction of the motile organism is changed this very
stimulating influence of the egg of the sea urchin upon
the spermatozoon of the starfish prohibited the latter
from getting into the egg. On the basis of the same idea
von Dungern was consistently led to the further con-
clusion that the egg exercised no ‘‘stimulating’’ influence
upon spermatozoa of its own species and that thereby the
spermatozoon of the same species was enabled to get
into the egg. A year after the appearance of von
Dungern’s paper the writer succeeded in accomplishing
the hybridization of the sea urchin egg with starfish
sperm by a method which contradicted von Dungern’s
theory, namely, by increasing the alkalinity of the sea
water whereby the spermatozoon is ‘‘stimulated’’ to still
greater activity; and on the other hand it is a common
experience that a sea urchin spermatozoon becomes more
active when it comes near an egg of its own species.

The writer was anxious to compare the activating

15 v: Dungern, Ztsch. f. allg. Physiol., I, 34, 1902.

No. 581] ENTRANCE OF THE SPERMATOZOON 271

action of eggs of the same and various foreign species
upon spermatozoa. Since the spermatozoa of the sea
urchins are usually very active in pure sea water (i. e.,
sea water free from egg substance) it was necessary to
find a solution in which these spermatozoa will keep alive
for a number of days without showing any motility.
Such a solution was found in a neutral m/2 NaCl solution
and this led to the method of putting ovaries and testes
directly into such solutions instead of into sea water.1®
The ovaries and testes were first washed repeatedly in
these solutions to free them from the blood or its salts,
and then one drop of eggs and one or more drops of the
sperm suspension were mixed in a watch glass containing
5 c.c. m/2 NaCl (free from egg contents). In one experi-
ment the sperm and eggs of two sea urchins, purpuratus
and franciscanus, and two starfish, Asterias ochracea
and Asterina (at Pacific Grove), were used. None of
the four forms of spermatozoa showed any motility in a
pure NaCl solution (without egg contents). In sea
water (free from egg contents) the spermatozoa of the
two forms of sea urchins were very active, those of the
starfish were immobile. The starfish eggs were imma-
ture and did not mature during the experiment (those of
Asterias were out of season and very small); the sea
urchin eggs were mature. The result is indicated in the
following table.

That there exists no strict specificity is obvious by the
fact that the immature eggs of Asterina activate the
sperm of the sea urchin franciscanus as powerfully as is
done by the mature eggs of the sea urchin purpuratus
and franciscanus. But the spermatozoa of the two
species of starfish show a marked specificity inasmuch as
they are activated strongly only by the (immature) eggs
of their own species and only to a slight degree by the

16 The writer had found previously that the unfertilized eggs of purpuratus
are killed more rapidly in sea water than in a neutral m/2 NaCl solution,
probably on account of the greater alkalinity of the former. The same may
be true for the sperm of this species, athough this has not.yet been tested.
The unfertilized egg of Arbacia is more sensitive to a pure NaCl solution
than that of purpuratus.

272 THE AMERICAN NATURALIST (Vou. XLIX

TABLE I
SPECIFICITY OF ACTIVATION OF SPERM BY EaGes

| Asterias d | Asterina € . | Franciscanus g| Wirpa

Asterias 9 (immature). . | | No Moderately iraa effect
very motile. | activation. active. | medi-

ate eae

egg.
Asterina 2 (immature). Not motile. elie activ- Violent activ- Sticke effect
| ity. | ity. | en near
| |
Franciscanus 9 (mature) Slightly | No motility. Immediately Immediately
motile. | active otile.
Purpuratus 9 (mature). Slightly tien effect Immediately Inm petane
motile afterin immediate
some time. bonte ct wit

eggs of the sea urchin purpuratus. In judging these
results the reader must keep in mind first that all these
experiments are made in a NaCl solution, and second,
that it requires a stronger influence to activate the
spermatozoa of the starfish which are at first not motile
in sea water (free from egg contents) than the sea urchin
spermatozoa which are from the very first very active
in such sea water and which may therefore be considered
as being at the threshold of activity in the pure NaCl
solution.

If instead of the eggs themselves the supernatant NaCl
solution from eggs is added to the sperm it is found that
it requires a very much greater concentration of the
supernatant NaCl solution from Asterias eggs to arouse
the purpuratus sperm in NaCl into activity than if the
supernatant NaCl solution from purpuratus or from
franciscanus eggs is used.

he question now arises whether the relative influence
of the egg on the motility of the sperm bears any relation
to the power of the latter to enter the egg; or in other
words if we can foretell which forms will hybridize by
observing the relative activating effect of the eggs upon
the spermatozoa. This does not appear to be the case
on the basis of our present limited experience, since the
activating effect of the franciscanus egg upon the sperm
of Asterias is just as great if not greater than that of
: purpuratus eggs and yet Asterias sperm can enter the

No.581] ENTRANCE OF THE SPERMATOZOON 273

latter and not the former. Even if we intensify the
activity of the spermatozoon of Asterias by putting it in
hyperalkaline sea water it will not enter the egg of
franciscanus.

If we mix eggs of franciscanus and purpuratus in sea
water and add the sperm of purpuratus the eggs of
purpuratus will be fertilized more quickly than the eggs
of franciscanus; and the reverse is true if the sperm of
franciscanus is added to a mixture of both eggs in sea
water. The writer is not quite certain that this differ-
ence is accompanied by a corresponding difference in the
influence of these eggs upon the motility of their sperma-
tozoa. It is certain, however, that the addition of egg
sea water from Asterias does not help the fertilization of
purpuratus eggs by Asterias sperm, although the egg sea
water from Asterias increases the activity of Asterias
sperm.

The writer is, however, of the opinion that this activat-
ing effect of the egg upon the spermatozoon is of the
greatest importance for fertilization in nature and that
the degree of specificity which exists (although it is far
from absolute) is a means of preventing hybridization.
The writer is under the impression that the eggs which
are naturally fertilized in water are fertilized almost
instantly after they are shed. Thus it is stated at
hatcheries that the egg of the salmon loses its power of
being fertilized in a few minutes and in the case of
Fundulus the egg loses this power also very rapidly.
The ripe egg of starfish dies rapidly if not fertilized. On
the other hand, the writer has often been struck with the
fact that the sperm of most marine forms when put into
sea water is at first practically not motile. When the
eggs have a specific gravity considerably greater than
the water (as is the case for Fundulus) the eggs will sink
very rapidly while the sperm remains suspended for
some time. Now we have mentioned that if the abso-
lutely inactive sperm of Asterias or Asterina comes in
contact with eggs of its own species (even if they are
immature) it is at once aroused into violent activity. If

274 THE AMERICAN NATURALIST [ Vou. XLIX

the same were true for the egg of Fundulus fertilization
could take place probably before the egg reaches the
bottom of the water. If by chance a teleost of a different
species would shed its sperm in the immediate neighbor-
hood and some of it could reach the egg of Fundulus
while it is falling the foreign sperm could probably not
be aroused as quickly by the egg of Fundulus as the
sperm of the Fundulus male and hence no hybridization
would occur. In fish we can see that the male and female
shed their sexual cells simultaneously so that they come
at once in contact. The writer is inclined to believe that
something similar occurs also in Echinoderms. He had
last year a chance to verify once more an observation he
had made for a number of years and which he had already
mentioned in a previous publication.” The sea urchins
at Pacific Grove are found in large numbers on rocks in
certain coves near the shore. Up to a certain day in
March every female of purpuratus was full of eggs. On
the next day the surface of the sea in this region showed
the usual indication of the spawning of large masses of
animals: namely the enormous foam formation in the
little coves although the sea was only moderately agi-
tated. This foam formation is due to an increase of
organic substances which lower the surface tension of
the sea water and make the foam more durable. The
writer realized that this might mean the end of material
for some time to come and indeed not a single female of
purpuratus of hundreds opened on that day had eggs.
The condition was the same for all the sea urchins col-
lected for two miles along the shore. During the next
week immature eggs began to appear again in the sea
urchins and in about ten days ripe eggs were again found.
This indicates that in this region the males and females
shed their eggs and sperm simultaneously. It is not im-
possible that among sea urchins which are found in
colonies on the rocks the shedding of the sexual products
of one or several individuals acts as an incentive for the
whole colony. Since the eggs fall in this case also much
17 Loeb, ‘‘The Mechanistic Conception of Life,’’ Chicago, 1912, p. 196.

No.581] ENTRANCE OF THE SPERMATOZOON 275

more rapidly to the bottom than the spermatozoa it is
also very probable that the eggs are fertilized before they
reach the bottom of the sea. We can understand under
these circumstances that the specificity which exists in
the activating effect of the egg upon the sperm is one of
the safeguards against hybridization for eggs that are
fertilized in the water, inasmuch as this specificity acti-
vates the sperm of the same species much more quickly
than that of a foreign species. Other safeguards are the
phase-boundary conditions which we discussed in the
previous chapter.

If we assume that the spermatozoon bores itself into
the egg by the energy of the vibrations of its flagellum it
is easy to understand the importance of its motility for
this process. It is, however, equally possible that a cer-
tain energy of vibration is needed to make the spermato-
zoon stick to the surface of the egg and that afterwards
forces of a different character bring the spermatozoon
into the egg. The fact that under normal conditions a
very slight degree of motility on the part of the sper-
matozoon allows it to enter the egg seems to favor such a
view.

von Dungern had already discussed the possible rôle
of phenomena of sperm agglutination in fertilization as a
protective agency. F. Lillie discovered the transitory ag-
glutination of sperm induced by a substance from eggs of
the same species.’8 When the sperm of the sea urchin
Arbacia is mixed with the supernatant sea water from
eggs of the same species a cluster formation occurs which
may last a number of minutes and which is essentially a
transitory agglutination. In Arbacia the agglutination
is very striking, in purpuratus the phenomena of agglu-
tination are not lacking but the writer was under the im-
pression that other phenomena of the type of tropisms
might enter. But he was not very certain on this point
ana ph that question open for further discussion. The

. Lillie, Science, N. S., XXXVIII, 524, 1913; Jour. Exper. Zool.,
Sop 523, 1914.

276 THE AMERICAN NATURALIST [ Von. XLIX

writer is, however, under the impression that no proof
for the existence of a positive chemotropism of the sea '
urchin sperm for the eggs of the same species has thus
far been given.

The writer observed that this phenomenon of sperm
agglutination depends on the motility of the sperm:’®
It only appears when the sperm is extremely motile and
it lasts only a number of minutes, often only a fraction
of a minute as Lillie had found. The writer observed
that the duration of the clusters depended to some extent
on the alkalinity of the solution. The more alkaline the
latter the more rapidly the cluster scatters. The pres-
ence of a salt with a bivalent metal, especially Ca, seems
necessary for the cluster formation. Sr and Ba act like
Ca and so does Mg but in the latter case a slightly higher
concentration is needed. The more Ca is added the more
powerful the agglutination becomes. These facts sug-
gest the following origin of the agglutination. From the
jelly surrounding the eggs a certain substance is dissolved
in the sea water which reacts chemically with a certain
substance at the surface of the spermatozoon. If this
reaction takes place in the presence of one of the salts
of a bivalent metal, especially Ca, a sticky precipitate is
formed on the surface of the spermatozoa, which is slowly
soluble in the solution; and the more rapidly the more al-
kaline the solution. If the spermatozoa are very active
the impact with which they strike each other may lead to
their sticking together and this agglutination will last
until the precipitate is dissolved again.”

The writer mentions this fact here because it might give
us a clue to the rôle of the motility of the spermatozoon
for its entrance into the egg. One can imagine that the
spermatozoon must stick to the surface of the egg in or-
der to be taken into it and this sticking may not come
about unless the spermatozoon strikes the surface of the

19 Loeb, Jour. Exper. Zool., XVII, 123, 1914.

20 Lillie measures the degree of agglutination by its duration; if our as-
sumption is correct he really measures the time required for the solution of
_ the sticky precipitate on the surface of the spermatozoon by the sea water.

No. 581] ENTRANCE OF THE SPERMATOZOON 277

egg with a certain velocity. This is, however, merely a
suggestion. The really serious difficulty of such an as-
sumption lies in the fact that the specific and transitory
cluster formation or agglutination of the spermatozoa is
not a general phenomenon. It may even turn out to be
confined to sea urchins and certain annelids. It is prob-
ably lacking in all cases of hybridization. Yet this would
not necessarily speak against the possibility of an ag-
glutination of the spermatozoon to the egg as a prerequi-
site of fertilization.

This latter idea receives some support in the writer’s
experiments on heterogeneous hybridization. He was
able to show that both NaOH as well as CaCl,, which
render possible the fertilization of the eggs of certain sea
urchins through the sperm of starfish, also favor the ag-
glutination of that sperm to the chorion of the egg. This
leads to the peculiar phenomenon of mere membrane for-
mation in the egg by the living spermatozoon without the
entrance of the latter into the egg.*!

Lillie seems to take it for granted that the substance of
the egg which causes sperm agglutination is identical
with the substance which stimulates the spermatozoa into
greater activity. If this were correct the conditions for
the two phenomena should be identical, which is however
far from being the case.

The writer showed that if we deprive the eggs of pur-
puratus of the jelly which surrounds them and if we wash
them afterwards a few times in sea water to deprive them
of the last vestiges of jelly substance which may still ad-
here to them they have lost completely and permanently
the power of forming clusters with the sperm of their
own species. Such eggs were washed four times in m/2
NaCl and when a drop of the supernatant NaCl solution
was added to NaCl sperm of purpuratus which was not
motile it activated the sperm very powerfully.

The writer had found that the egg sea water of S. fran-

21 Loeb, Arch. f. Entweklngsmech., XL, 310, 1914.

278 THE AMERICAN NATURALIST [ Von. XLIX

ciscanus does not give a trace of agglutination with the
sperm of purpuratus but if the experiment is made in
m/2 NaCl solutions it can be shown that the franciscanus
egg NaCl solution activates the NaCl sperm of pur-
puratus in an m/2 NaCl solution very strikingly.

The immature eggs of Asterias ochracea activate the
otherwise non-motile sperm of the same species, but the
eggs of this starfish do not give any agglutination reac-
tion with their own sperm and Lillie found the same for
the starfish in Woods Hole. It might be said that all this
only proves that the activating effect requires a smaller
concentration than the agglutinating effect, but may yet
be caused by the same substance. This objection is, how-
ever, not tenable in the following case.

Purpuratus sperm washed in m/2 NaCl is as a rule
more active in a mixture of NaCl + KCl than in a mixture
of NaCl + CaCl, (if both solutions are free from egg con-
tents); yet in the latter solution the agglutination reac-
tion upon the addition of egg-NaCl is very strong while
in the former it is lacking (unless the sperm or testicles
or ovaries give off some CaCl, to the surrounding solu-
tion). Again it might be argued that the activation of
the spermatozoon might be induced by the same sub-
stance as the agglutination, but that the agglutinating sub-
stance in both cases reacted with different constituents
of the spermatozoon. While this may be admitted, it
must also be conceded that with the facts which we have
at our disposal at present we can not be certain that the
aggultinating and activating substances are identical.

VII

Lillie?? not only takes the identity of the two substances
for granted but he assumes that without the agglutinating
substance in the egg (to which he gives the somewhat
prejudicial name ‘‘fertilizin’’) no fertilization is possible.
Fertilization in his opinion consists in the combination
of the spermatozoon with a molecule of or in

22 Loe. cit.

No.581] ENTRANCE OF THE SPERMATOZOON 279

the egg, whereby the fertilizin molecule undergoes a
change in the other end and this change causes the egg to
develop. The fertilizin is thus an ‘‘amboceptor”’ in the
sense of Ehrlich’s side-chain theory.

The side-chain theory was invented by Ehrlich for an
altogether different purpose. Bordet had found that for
certain phenomena of immunity two substances’ were
needed (which Ehrlich named amboceptor and comple-
ment, respectively). Ehrlich assumed that they were
bound chemically by the antigen (the substance against
which the organism was immunized) but found that while
the antigen (4) was able to bind B (the amboceptor) in
the absence of C, it was not able to bind the complement
C in the absence of B. From this Ehrlich concluded that
of the two possible modes of linkage between the three

bodies 44 a and A—B—C the latter was the one which

a
really occurred. Since in this case C is not directly
linked with A but through the intermediation of B he
called B the ‘‘ amboceptor ’’ and the scheme of linkage a
** side-chain °’ linkage.

Lillie applies this theory (which covers the two possible
modes of linkage of two chemical compounds to a third
one) to the entrance of the spermatozoon into the egg, by
calling the egg an antigen A and the spermatozoon a com-
plement C and assuming the existence of a hypothetical
amboceptor B in the form of the substance that causes
agglutination, the ‘‘ fertilizin.’’ Even if we are willing
to overlook the fact that the egg and the spermatozoon
are cells and not simple organic compounds and if we are
willing to overlook the further fact that the assumption
of an amboceptor as a connecting link between the two
is arbitrary we can not overlook the fact that the sperma-
tozoon does not combine chemically with the egg but that
it actually enters into the egg and attaches itself to the
egg nucleus. It seems then futile to discuss whether the
Spermatozoon combines with the egg in side-chain fashion
(namely, Egg—Fertilizin—Spermatozoon) or in direct

280 THE AMERICAN NATURALIST [ Vou. XLIX

fashion, namely,

/ Fertilizin

\Spermatozoon

since the engulfing of the spermatozoon into the egg is a
physical process which bears no relation to either possi-
bility.

It has been stated that the ‘‘ fertilizin theory ’’ explains
also the phenomena of artificial parthenogenesis just as
well as any other theory. In a recent book on artificial
parthenogenesis the writer has given the results of a
large number of experiments and he has tried to explain
some of them; the reader would, however, vainly look
for a ‘‘theory”’ of artificial parthenogenesis. A theory
in a scientific sense consists in the presentation in
mathematical or numerical form of a phenomenon as
the function of its variables. The writer has tried to pre-
pare the ground for such a treatment of the phenomena
of fertilization and of the first development of the egg by
working out those variables which permit a quantitative
treatment, but even if the exploration had been advanced
further than it actually has been, it would not be possible
to ever expect that a single theory could cover all the
phenomena of fertilization and development, since under
these two headings so many physically and chemically
different processes are included (of which one follows the |
other) that they can not be covered by one theory. It is
true the writer had in former publications océasionally
used the term ‘‘ lysin theory of fertilization ’’ but only to
express the fact that cytolytic agencies induce membrane
formation and that the membrane formation induced by a
spermatozoon might also be due to a cytolytic agency con-
tained in the spermatozoon; but he has dropped this term
in his recent book on the subject.

While the writer does not desire to enter into a further
discussion of the side-chain theory of fertilization he
wishes to point out that it rests on the claim that that
substance which causes sperm agglutination is contained

Egg

No. 581] ENTRANCE OF THE SPERMATOZOON 281

in the unfertilized egg and that the egg can only be fer-
tilized as long as this ‘‘ fertilizin ’’ is present in the egg.
It is obvious that such an assumption demands for its
proof that in all cases in which an egg can be fertilized it
must contain the agglutinating substance. There is only
one test for the presence of this substance, namely the
cluster formation of the sperm in the presence of egg sea
water. This proof can not be furnished since, as the
writer had shown in a former paper, the reaction is lack-
ing in many cases of hybridization; it is also lacking in
the case of the starfish.” It is not impossible that if the
theory is tested further it will be found lacking in a con-
siderable number of cases. To this objection Lillie re-
plies that it is not necessary that the eggs should actually
give the agglutinin reaction, it is sufficient that the ag-
glutinating substance is contained in the egg. But how
can we tell that it is contained in an egg which fails to
give the agglutination reaction as long as this reaction is
the only reliable test for the presence of the agglutinating
substance in the egg? Rigorously speaking, even if all
eggs of every species gave the agglutinin reaction it would
still be necessary to furnish a direct proof that the ag-
glutinin has anything to do with fertilization and develop-
ment.
It may be possible that Lillie considers such a proof
to be contained in the following statement.
I adopted then the working hypothesis that this substance?* is neces-
sary for fertilization and there followed immediately three corollaries,
iz.: (1) if it were possible to extract this substance from eggs they
would no longer be eapable of fertilization; (2) fertilized eggs are inca-
pable of uniting again with spermatozoa, hence if the hypothesis is
correct they could no longer contain free fertilizin; (3) eggs in which
membranes have been formed by methods of artificial parthenogenesis
become incapable of fertilization; such eggs must also therefore be de-
void of free fertilizin after they have reached the non-fertilizable con-
dition if the eta: is correct. These consequences were actually
found to be true.?
23 Lillie, Biol, Bull., XXVIII, 18, 1915.
24 The ‘‘ fertilizin.’’
25 Lillie, Jour. of Bio, Zool., XVI, 523, 1914.

282 THE AMERICAN NATURALIST [Vov. XLIX

Of these three ‘‘ corollaries ’’ the first one is the most
important, since it claims that the power of the eggs of
being fertilized varies with their contents of fertilizin.
The proof consisted in this: that eggs were washed a
number of times during three consecutive days and after
two days the percentage of eggs that could be fertilized
were diminished to about one third.

There is thus the anticipated decrease in the percentage of fertilizations.

It is a well known fact that the unfertilized eggs of the
sea urchin (in fact of all marine animals) perish when
they lie for some time in sea water and one of the main
causes of this phenomenon is also known, namely oxida-
tions. If the oxidations are inhibited through the removal
of oxygen or the addition of KCN the life of the eggs can
be prolonged.’ In the mature starfish egg this death
which is accelerated by the temperature (and has the high
temperature coefficient of many life phenomena) takes
place in a few hours,” while it begins a little later in the
egg of the sea urchin. After the artificial membrane for-
mation it takes place very rapidly also in the sea urchin
egg (coincident with the enormous increase in the rate of
oxidations caused by the artificial membrane formation)
and in this case the death of the egg can also be retarded
by the withdrawal of oxygen or the addition of eyanide.*$
In view of these facts the objection can not be avoided .
that in Lillie’s experiment the number of eggs which could
be fertilized fell off after two days to one third not on
aecount of the loss of ‘‘ fertilizin’’ but because of the
fact that two thirds of the eggs were dead by that time.
That this assumption is well grounded is testified by
Lillie’s own remarks:

Concomitantly, with these effects of the series of washings the devel-
opmental energy becomes greatly reduced. This was very obvious from
the second fertilization.2® On August 24 (48 hours after fertilization)
a large quantity of living material was contained in the second A fertili-

26 Loeb and Lewis, Am. Jour. Physiol., VI, 305, 1902.

27 Loeb, Biol. Bull., III, 295, 1902. _

28 Loeb, ‘‘ Artificial Parthenogenesis and Fertilization.’’

29 Which occurred on the second day.

No. 581] ENTRANCE OF THE SPERMATOZOON 283

zation but none had even approximately pluteus:structure. The most
common form was a stereoblastula. In the second B fertilization there
were a few abnormal prismatic plutei, while the majority were gastrulae.
The third fertilization resulted in extremely abnormal ciliated types.
The fourth and fifth did not proceed beyond abnormal cleavage stages.
From this and similar experiments Lillie draws the fol-
lowing conclusion:

The eggs have evidently lost something which affects their power of
fertilization. Table 3 shows the measure of loss of the sperm agglu-
tinating substance and justifies the general conclusion that this is a
factor in the result. The loss of other substances may also combine in
the decrease of fertilizing power, but of this we know nothing definite.
As a matter of fact, fertilizing power is gradually lost with decrease of
fertilizin content of the egg.

It seems to the writer that in these experiments the
power of being fertilized was gradually lost by the death
of the eggs. And an additional justification of this criti-
cism is given by the following fact, that if we deprive
fresh eggs of purpuratus permanently of their power of
giving off ‘‘ fertilizin ’’ their power of being fertilized is
not only not lost but is entirely unaltered. The writer
has shown that if the eggs of purpuratus are treated for
two or three minutes with a mixture of 50 c.c. of sea water
+3 c.c. of HCl (whereby the jelly surrounding the egg
is dissolved) and if the eggs are washed they give no
trace of a fertilizin reaction but 100 per cent. of the eggs
can be fertilized.* ;

It might be argued that the supernatant sea water from
these eggs had not lost all power of causing agglutination
of the sperm. This the writer must deny but for argu-
ments’ sake he will admit that a trace near the ‘‘ psycho-
logical limit ’’ might have been overlooked where a ‘‘ fer-
tilizin ” partisan might have declared that he still could
perceive a faint indication of a ‘‘ fertilizin ” reaction.
In that case only a few eggs should have been fertilized—
the fertilizin theory rests on this assumption; in reality,
however, practically one hundred per cent. were fertilized
in every case (provided the eggs had not been lying in the
see water too long, i. e., more than a day or two).

30 Loeb, Jour. Exper. Zool., XVII, 123, 1914.

284 THE AMERICAN NATURALIST — [Vou. XLIX

To this Lillie replies that perhaps the sperm of pur-
puratus is not so delicate an indicator for agglutinin as
the sperm of Arbacia—but as long as the agglutination
reaction is the only test for the presence of fertilizin in
the egg, such an answer begs the question.

From the fact that the power of agglutinating the sperm
is lost if the egg of purpuratus is deprived of its jelly by
acid treatment the writer drew the conclusion that in this
egg the ‘‘ fertilizin’’ does not come from the unfertilized
egg but only from its jelly and that this was contrary to
Lillie’s assumption. To this Lillie?! replied by pointing
out that the immature eggs of Arbacia do not give the ag-
glutination reaction while the mature Arbacia egg gives
the reaction very powerfully, and that we must conclude
from this that the ‘‘ fertilizin’’ contained in the jelly
comes from the egg and is given off during the period of
the maturation divisions (the latter statement, however,
is after all only an assumption though a probable one).
But this does not meet the question at issue, namely that
in the egg of purpuratus at the time of maturity the fer-
tilizin which is given off is contained exclusively in the
jelly and not in the egg, as it should be if the presence
of fertilizin in the egg were a prerequisite for its ability
of being fertilized. It is true that if we repeat this ex-
periment in the egg of Arbacia we find that after the re-
moval of the jelly by HCl a trace of the agglutinating sub-
stance may still be given off by the egg, although little in
comparison with that given off by the jelly. But this
does not alter the facts as they are found in the egg of
purpuratus.

As far as the two other proofs of Lillie are concerned,
we have already touched upon them in the previous parts
of this paper. The fact that the fertilized eggs of Ar-
bacia (and of purpuratus) cease to give the agglutinin
reaction is due to the loss of the jelly on the part of the
fertilized egg to which in Arbacia should be added the
fact that some of the material of the cortical layer is given —

31 Lillie, Biol. Bull, XXVIII, 18, 1915,

No. 581] ENTRANCE OF THE SPERMATOZOON 285

off during the process of membrane formation. The
writer has pointed out in former papers that the cortical
layer of the egg which undergoes liquefaction in the
process of membrane formation behaves towards reagents
very much like the jelly which surrounds the egg.*? But
since in the egg of purpuratus the loss of this agglutina-
ting power on the part of the egg is not necessarily accom-
panied by the loss of the power of being fertilized—e. g.,
in the HCl experiment—we are inclined to believe that
there must be another reason that an egg fertilized by
sperm can not be fertilized a second time.

As far as the statement is concerned that the egg can
no longer be fertilized after artificial membrane forma-
tion by butyric acid the writer can not admit the correct-
ness of this statement (see Chapter III). In the eggs in
which artificial membrane formation has been called forth
by butyric acid the main if not the only block to a subse-
quent fertilization is the membrane itself.

This can be proved by a very simple experiment. If
we call forth the membrane formation in the egg of
purpuratus in a neutral or faintly alkaline solution of m/2
(NaCl + KCI + CaCl.) (instead of in sea water) a very
thin membrane is formed, which is easily torn and offers
no resistance to the spermatozoon. All the eggs treated
in this way can be fertilized by sperm. The agglutinin
reaction of such eggs is, however, permanently lost.

The facts thus far known seem to force us to the conclu-
sion that no adequate proof has been offered thus far for
the connection between the power of an egg of being fertil-
ized by sperm and its power of causing a cluster formation
of the sperm. The writer has pointed out in a previous
paper that it is difficult to see why there should exist
such a relation, since sperm agglutination can only in-
. hibit the entrance of the spermatozoon into the egg.

82 ‘í Artificial Parthenogenesis and Fertilization,’ Chicago, 1913, pp.
210-14, University of California publication, Physiology, Vol. 3, p. 1, 1905.

GERM CELLS AND SOMATIC CELLS?
LEO LOEB

Resvtts obtained in the field of experimental pathology
and especially in cancer research have an important bear-
ing on certain problems of general biology. In the follow-
ing I wish to consider connectedly some of these facts
from this point of view.

I. A sharp distinction between germ cells and somatic
cells has become clearly established, especially through
the writings of Nussbaum and Weismann. More recent
results which demonstrated that the differentiation of
germ cells from the somatic cells at a very early stage
of embryonic development and their non-participation in
the formation of somatic tissues exists in various species,
tended to emphasize this sharp distinction between
somatic and germ cells.

Weismann? especially insisted on the radical difference
between germ cells and somatic cells, inasmuch as he
attributed potential immortality to the former and only a
temporary existence to the latter. And Weismann re-
gards this difference as essentially founded in the struc-
ture of both kinds of cells and fundamentally connected
with the functioning of the somatic cells; this difference
was obtained through selective processes as an adaptation
in the struggle for existence. He does not regard the
death of somatic cells as an accidental occurrence due to
unfavorable conditions which it might be in our power to
change, but as an inherent characteristic of somatic cells.
He mentions, though casually, that the life of the cock’s
comb might be prolonged by grafting it on another fowl—
but only to dismiss this idea as having no important theo-

1 From the Department of Pathology, Barnard Free Skin and Cancer Hos-
pital, St. Louis.

2A,

Weismann, ‘‘ Ueber Leben und Tod,’’ Jena, 1884; ‘‘ Ueber die Verer-
bung,’’ Jena, 1883

286

No. 581] GERM CELLS AND SOMATIC CELLS 287

retical bearing. R. Hertwig? also regards the death of the
somatic cells as unavoidably determined by their organi-
zation which precludes necessary readjustments.

Minot? likewise held the life of somatic cells to be
limited in duration and he ascribed this limitation to
changes in cell structure, leading to a differentiation of
the cytoplasm during the process of life; a change which
he designated as cytomorphosis.

Within the last 14 years certain facts have been estab-
lished which are contrary to this conception of a radical
difference between germ and somatic cells as far as their
potential immortality is concerned. Experimental inves-
tigations in tumor growth have furnished these facts.
Before we state these results we have first to consider,
how far tumor cells can be regarded as somatic cells. We
consider here malignant tumors (cancers). They origi-
nate at various parts of the body, often under the influ-
ence of long-continued irritation. In many cases we can,
if we obtain sufficiently early tumors, trace the trans-
formation of the normal into the abnormally proliferating
(tumor) tissue. This has been, as was to be expected,
especially observed in the case of superficial cancers,
where early stages of tumors are most likely to be en-
countered, for instance, in cancers of the skin, of certain
mucous membranes, but also occasionally in internal can-
cers as in those of the stomach. In such cases the cancer
cells are undoubtedly the offspring of ordinary somatic
cells. There are, however, tumors, so-called teratomata,
which in all probability take their origin in the germ
glands and other parts of the body from parthenogenet-
ically developing ova. But while these latter tumors do
not originate from somatic, but from germ cells, the tumor
cells themselves are no longer germ cells, but somatic cells
in the same sense as the ordinary tissue constituents
which also are derived from germ cells. We can therefore

3 R. Hertwig, Biol. Centralblatt, Bd. XXXIV, No. 9, 1914.

+C. S. Minot, ‘‘The Problem of Age, Growth and Death,’’ New York,
1908,

288 THE AMERICAN NATURALIST [Vou. XLIX

without doubt regard tumor cells as a kind of somatic
cells.

One of the most characteristic properties of cancer cells
is their ability to grow after transplantation into other
animals of the same species. This applies not to all, but toa
certain number of spontaneous cancers; the majority of
spontaneous tumors are not transplantable into other in-
dividuals of the same species. They grow, however,
usually after transplantation into the same individual in
which they originated. There does not exist as far as
their origin is concerned any essential difference between
these two kinds of cancers—those transplantable and not
transplantable into other individuals. The cancers used
in experimental tumor investigation take their origin
from somatic cells; but it appears some are less sensitive
to the difference in the chemical composition of the body
fluids which exists between different individuals of the
same species than others, and those less sensitive can be
transplanted, while others can not.

In those tumors which are transplantable, relatively
few tumor cells give after inoculation into other animals
origin to the new tumors, and the tumor cells after the
first transplantation not rarely multiply with greater
vigor than they did in the original animal, an effect caused,
as I could show, through the stimulating influence of the
cutting and otherwise manipulating the tumor cells. In
each animal therefore there are produced many successive
generations of tumor cells, and after transplantation into
another individual each surviving cancer cell produces
again new generations. Consecutive transplantations into
many individuals have been carried out with the same
tumor. The potential proliferative power of the cancer
cells is therefore enormous. It is, however, not so much
the intensity of the proliferative power of the tumor
cells which we wish to consider as the potential duration
of their life. It has been shown that epithelial, as well as
connective tissue tumors can be transplanted through
many generations and can survive for a long time the
animal in which the tumor originated. Thus I was able

No. 581] GERM CELLS AND SOMATIC CELLS 289

to transplant the connective tissue cells of a rat sarcoma
through forty successive generations of animals, and it
was merely the result of accidental bacterial infection
due to the unfavorable conditions under which the work
had to be carried out which caused the ultimate death of
the propagated cells.

An epithelial tumor found by Jensen in a mouse has
been propagated in various laboratories through a period
of almost fifteen years, and another epithelial tumor of
the mouse we have been propagating in mice for a period
of seven or eight years, without any sign of diminishing
vitality in the propagated cells being noticeable.

In all these transplantations of tumor cells, be they of
connective tissue or epithelial origin, it could be shown
that the peripheral cells remain alive and from these sur-
viving cells the cell growth starts. These observations
suggested to me in 1901 the conclusion that tumor cells
may have a potential immortality in a similar manner as
germ cells,® and inasmuch as tumor cells are only modified
somatic cells, I furthermore concluded that the same
statement holds good in the case of somatic cells.” Fur-
ther experiences in the field of experimental tumor inves-
tigation during the following years confirmed this conclu-
sion and permitted its enunciation with greater definitive-
ness. The potential immortality of the somatic cells of
course can only be made probable, it can never be defi-
nitely proved, inasmuch as our experience merely deals
with finite periods. But the same restriction holds good
in the case of the germ cells in which the potential immor-
tality is likewise merely a strong probability and not a
definitely proven fact.

Weismann believed that protozoa are in the same sense
potentially immortal as germ cells, in contradistinction to
somatic cells which do not possess potential immortality.
Some facts were, however, discovered which, according to

5**On the Transplantation of Tumors,’’ Jour. Medical Research, Vol. VI,
No. 1, 1901, p. 28; Virchow’s Archiv, Bd. 167, 1902, p. 175.

6‘‘Tumor Growth and Tissue Growth,’’ Am, Philosophical Society,
XLVII, 1908, and at other places.

290 THE AMERICAN NATURALIST [Von. XLIX

the interpretation given them, seemed to contradict
Weismann’s conception. Thus Maupas found that vari-
ous kinds of infusoria did not propagate by fission indefi-
nitely, but that a sexual process, conjugation, was neces-
sary at certain times, and Calkins showed that there were
regular periods of depression, and while a spontaneous
recuperation from the effects of certain depressions could
take place and in still other cases artificial stimulation
would aid the animals in overcoming the critical periods,
at other times depressions proved fatal without an inter-
vening conjugation. Woodruff, however, by choosing
conditions of environment more in accordance with the
conditions found in nature, could keep a strain of Par-
amecium apparently indefinitely alive without any inter-
vening periods of copulation being required. This seemed
to point to a potential immortality of protozoa in the
sense of Weismann. Recently, however, Woodruff and
Erdmann‘ found that the recovery from depression which
takes place is accomplished through nuclear changes com-
parable to, but not identical with, those observed during
copulation. This seems in some respects to agree with
R. Hertwig’s previously enunciated theory according to
which depressions and senility in cells are due to a dis-
proportion between the nuclear and cytoplasmic material,
and that recovery from such unfavorable conditions de-
pends upon the reorganization of the nucleus, essentially
consisting in a diminution of the mass of the latter. Inas-
much as in metazoa—he concluded, further—such a rear-
rangement between nuclear and cytoplasmic masses can
only take place in the case of germ cells, but not somatic
cells, only germ cells are immortal, while somatic cells
are necessarily mortal. While Weismann regarded the
unavoidable mortality of the somatic cells as a secondary
acquisition, the result of a process of selection, the death
of somatic cells being of advantage to the propagation of
the race, Richard Hertwig’ regards the death of somatic
cells as inherent in their structure, which precludes the

7 Biol, Centralblatt, Bd. 34, August, 1914, p. 484.
8 Biol. Centralblatt, Bd, XXXIV, 1914, No. 9.

No. 581] GERM CELLS AND SOMATIC CELLS 291

possibility of nuclear reorganization of the cell necessary
for continued life. In a somewhat related way, Minot
considers, as mentioned above, the death of somatic cells
as inevitable and as the result of cytomorphosis, which
means the relative increase in size and differentiation of
somatic cells during life. In this connection it is inter-
esting to note that while R. Hertwig considers (in pro-
tozoa primarily, but secondarily also in other cells) an in-
crease in the size of the nucleus—the result of the activity
of the cells—as the cause of functional disturbances lead-
ing to senility, Minot on the other hand connects senility
with a relative decrease in the size of the nucleus and an
increase in the mass of the cytoplasm. Now as far as the
protozoa are concerned, the controversy does not seem to
concern so much the potential immortality of these organ-
isms as the problem as to whether the individual pro-
tozoon corresponds to a germ cell or to a somatic cell of a
metazoon, or whether it partakes of the character of both.
There can be little doubt that individual protozoa possess
potential immortality, a conclusion which would not be
invalidated through a loss of certain parts of the pro-
tozoon body at certain periods of its life cycle.

We may therefore conclude that all three kinds of cells,
protozoa, germ cells, as well as certain somatic cells of
metazoa, possess a potential immortality.

Tumor cells are somatic cells in which such secondary
changes leading to a cessation of proliferation as take
place under certain conditions in all the individual cells in
some kinds of somatic tissues, are affecting only a certain
number of cells. In the case of some somatic cells, as, for
instance, those of the epidermis, it is evident that the
secondary changes in structure and metabolism, which
lead to a cessation of proliferative power, are due to un-
favorable conditions of blood-supply. What Minot calls
eytomorphosis can therefore in this case be referred not
to necessary transformations inherent in the cells, but to
unfavorable environmental conditions into which the
cells are placed as a result of their multiplication. Such
secondary degenerative changes take place also in tumor

292 THE AMERICAN NATURALIST [Vou. XLIX

cells under similar defective conditions of blood-supply.
Here also degenerative changes entail a cessation of pro-
liferation in a similar manner as in ordinary tissue cells.
While farther away from the blood-vessels the tumor cells
degenerate and die, near the blood-vessels they continue
to live and to multiply.

While from a theoretical point of view, therefore, the
question as to the potential immortality of somatic cells
has through the experiments on tumor cells been answered
in a decisive manner, it was nevertheless of interest to
extend these investigations to ordinary tissues. Such
investigations we undertook in the course of the last
eight years, and while certain obstacles were encountered,
which prevented the continued life of ordinary tissues, the
results were of interest in giving an insight into some of
the conditions which determine the growth, life and death
of somatic cells. These investigations have shown that if
tissues are transplanted into another individual of the
same species, under the influence of the constitution of
the body fluids, which differs in different individuals of
the same species, the metabolism in the transplanted
tissues is interfered with as shown, for instance, in patho-
logical differences in pigmentation seen in black skin of
the guinea-pig after transplantation into other animals of
the same species. After transplantation of pigmented
skin into the same individual in which it originated, such
pathological changes do not occur. As I have previously
pointed out, a certain adaptation exists between the
tissues and body fluids in animals of the same species, and
even between the tissues and body fluids of the same indi-
vidual. Thus it comes about that the interaction of tis-
sues and body fluids of the same species leads to different
and less toxie products than those produced through the
interaction of the body fluids of one with the tissues of
another species. Even the interaction of tissues of one
animal with the body fluids of another animal of the same
species leads to more toxic products than the interaction
of body fluids and tissues of the same individual. In the
latter case toxic products, interfering with the life of

No. 581] GERM CELLS AND SOMATIC CELLS 293

normal tissues, are not produced, while in the former cases
such products acting directly or indirectly are formed.
As an illustration of such a specific relationship between
body fiuids and tissues, I cited the specifically adapted
effect which tissue coagulins exert on the constituents of
blood plasma.’

Now as the result of these differences in metabolism
induced through the differently constituted body fluids the
lymphocytes begin to invade the transplanted tissues, and
the invading connective tissue does not preserve, as it
does after auto-transplantation, its young and cellular
state, but produces fibrous bands which contract around
the parenchyma after homoiotransplantation, and thus
exert pressure. Both connective tissue and lymphocytes
destroy thus the homoiotransplanted tissue, while they
usually spare the autotransplanted tissue the metabolism
of which is normal. In the case of certain tissues, as, for
instance, kidney, however, even after autotransplantation
into the subcutaneous tissue the metabolism of the trans-
planted cells becomes abnormal under the abnormal con-
ditions under which they now live, and here the lympho-
cytes and connective tissue destroy, therefore, even the
autotransplanted tissue, although at a later date than the
homoiotransplanted kidney tissue.

The fitness of a tissue in an individual determining its
power to live or to grow depends, therefore, on two factors:
(1) on the specific adaptation existing between tissues and
body fluids, and (2) on the way in which various sub-
stances are carried to the tissue. A perfect nutrition im-
plies the carrying of the food substances to the tissues in
the normal way through blood-vessels. It is probable that
on the intact relations between capillary endothelium and
parenchyma cells depends such a sifting of various food
substances and waste products as is best suited to the
normal metabolism of the cells. If, as in the case of the
kidney tissue, this mechanism is disturbed, abnormal sub-
stances are produced notwithstanding the specific adapta-
tion existing in this case between tissue cells and body

? ‘“‘Immunity and Adaptation,’’ Biol. Bulletin, Vol. IX, 1905, p. 141.

294 THE AMERICAN NATURALIST (Von. XLIX

fluids after auto-transplantation. And it seems that a
perfect fulfilment of the second requirement might even
be able to overcome a deficiency in the first condition, the
specific adaptation between tissues and body fluids. This
seems at least to be the case, whenever a kidney is suc-
cessfully transplanted into another individual of the same
species and lives here for a long period of time.

A peculiar resistance to foreign body fluids is appar-
ently shown by the germ cells. They represent in reality
individuals residing in a host organism of the same spe-
cies. In this case the host organism is nearly related to
but not identical with the individuality of the germ cells.
In some respects we have, therefore, here a condition com-
parable to one existing after homoiotransplantation of
tissues. And still the germ cells do not show any signs of
injury. There is, therefore, in germ cells as yet lacking
that substance which has a specific affinity to certain parts
of the body fluids, or through their situation the germ
cells are somehow protected against the injurious influ-
ence of these substances.

Thus it comes about that through transplantation into
other individuals of the same species the potential im-
mortality of the ordinary tissues can not be demonstrated.
This applies to the tissues investigated so far.’ How-
ever, it is quite possible that we may yet find that in the
case of certain tissues the life may be permanent even
after homoiotransplantation. It was furthermore think-
able that through serial transplantation, retransplanting
the tissue at an early date before the lymphocytes and con-
nective tissue had had a chance to seriously injure it,
better success could be obtained. In the case of the skin
I have undertaken such serial transplantations some years
ago; our investigations have, however, in this case shown
that it was not possible to retransplant this particular

10 Leo Loeb u. W. A. F. Addison, Arch. f. Entwicklungsmechanik, Bd.
XXVII, 1909, p. 73; Bd. XXXII, 1911, p. 44. Max W. Myer, Bd.
XXXVIII, p. 1, 1913. Llewellyn Sale, Bd. XXXVII, 1913, p. 248, M'G.

Seelig, Bd. XXXVII, 1913, p. 259. Cora Hesselberg, Journ. Experimental
Medicine, Vol. XXI, 1915, p. 164.

No. 581] GERM CELLS AND SOMATIC CELLS 295

tissue indefinitely ;'! these experiments ought to be ex-
tended; especially might it be of interest to use the direct
descendants as hosts for the tissues of the parent. It is to
be expected that the quality of the parents which makes
the body fluids suitable for their own tissues might make
them likewise suitable for certain of their offspring. Such
experiments I began some time ago and I expect to
continue them if opportunity should present itself.

The growing of tissues in culture media, which excludes
attack on the cells by connective tissue and lymphocytes,
may also serve the same purpose and quite recently has
been used through a larger number of generations. But,
as stated above, we have in these experiments merely to
deal with an attempt to confirm the potential immortality
of the somatic cells which had in principle been estab-
lished through previous investigations on the life of tumor
cells.

While thus various kinds of tissues of an organism have
the potentiality of an immortal life, separated from the
organisms to which they belonged, the organism as a
whole invariably dies and with it its component tissues.
This is evidently due to the interdependence of various
parts of an organism and to the death of certain sensitive
cells, especially the ganglia cells of the central nervous
system. We might therefore be inclined to conclude that
these ganglia cells do not possess the potentiality of im-
mortal life. But even in the case of the ganglia cells,
which are of such significance for the life of the organism
as a whole, we can at present not deny the possibility that
they also may have the potentiality of immortality and
that they merely succumb under the influences of certain
injurious conditions arising in the organism. On the
other hand, fully developed ganglia cells have apparently
lost the power to multiply; they are furthermore sensi-
tive to certain insults to which other tissues show resist-
ance. Thus the unfavorable condition prevailing during
the process of transplantation into another organism and
directly afterwards seems to be sufficient to cause the death

11 Leo Loeb, Archiv, f. Entwicklgsmch., Bd. XXIV, 1907, p. 638.

296 THE AMERICAN NATURALIST [ Vou. XLIX

of certain ganglia cells when other tissues would survive.
But neither of these facts prove that under favorable con-
ditions very much differentiated cells, like ganglia cells,
might not have the power to live indefinitely, although they
have lost the power to multiply. Sensitized connective
tissue cells of the uterus may after transplantation into
the same or another individual in which corpus luteum
substance is circulating grow very energetically and pro-
duce placentomata, while after transplantation of fully
developed deciducomata no further growth can be ob-
tained and all or almost all the cells die. Here also the
fully ‘‘differentiated’’ cells have lost the power to multi-
ply and at the same time they have apparently become
more sensitive to the effect of injurious influences than
young and yet undifferentiated predecidual cells of the
uterine mucosa.

But here we can observe that some strands of fully
differentiated placentoma tissue may survive even under
those unfavorable conditions—without, however, resuming
growth or returning to the undifferentiated condition—
namely, such strands of tissue as are situated under the
best environmental conditions, in close proximity to the
host tissue, at places most accessible to the foodstuffs or
oxygen supplied by the circulating blood or by the peri-
toneal fluid. This suggests that even much differentiated
cells which have lost their power to propagate may still
have the power to live, when kept under favorable condi-
tions, and that their death is the result of unfavorable
environmental influences. Thus we must at least admit at
the present time the possibility that also the ganglia cells,
while they do no longer multiply, may still possess a
potential immortality; that cellular differentiation pre-
eludes the latter possibility has as yet not been demon-
strated. We must therefore sharply distinguish between
the power of cells to grow and their power to live; while
the former seems to be destroyed through differentiation
—at least in some cases—the latter may still exist. In
the case of other tissue cells we have it to a certain
extent in our power through experimental conditions

No. 581] GERM CELLS AND SOMATIC CELLS 297

to prevent those changes which lead to differentiation
and death; thus in the case of tumor cells through con-
stant transfer into a new host we can enormously in-
crease the number of as yet less differentiated cells
on which the propagation depends by causing an in-
tense multiplication of these tumor cells, many of which,
if left in the same organism, would have undergone sec-
ondary degenerative changes. Through experimental
means, viz., through the transplantation into different
kinds of hosts the relative preponderance of propagation
and differentiation of tumor cells can be varied. We fur-
thermore know that through ‘‘chemical’’ sensitization
combined with mechanical stimulation, the same effect can
be produced, at least temporarily, in the connective tissue
cells of the uterine mucosa. Under those conditions a
large number of cells are induced to propagate and re-
main young while their offspring gradually change into
fully differentiated cells. If we grant the possibility that
differentiation of such cells as ganglia cells, while it en-
tails loss of the power to propagate and greater sensitive-
ness to insults, does not necessarily mean the necessity to
die, then the problem of prolongation of life would to a
great extent depend upon the possibility of preventing
injurious influences which at present disturb the function
of ganglia cells from attacking these cells and causing
their death.

II. The germ cells are potentially immortal, but this
potential immortality can only be realized, if at certain
periods certain changes take place within the cells, which
concern especially the nucleus, phenomena consisting in
maturation, followed by fertilization or parthenogenetic
development. In a similar manner the observations of
Woodruff and Erdmann suggest that at the time of de-
pressions in the life of the protozoa, possibly similar nu-
clear phenomena take place at least in certain cases. We
have seen that in the case of somatic cells there are also
indications of the existence of potential immortality. The
question may therefore be raised, whether similar periodic
rearrangements of the nucleus, as in the case of the germ

298 THE AMERICAN NATURALIST [ Vou. XLIX

cells and protozoa, may not also take place in the case of
the somatic, especially of tumor cells. Without consider-
ing any connection with the problem of the immortality of
the somatic cells, Bashford, Murray and Bowen"? stated
on an empirical basis that in charting the number of suc-
cessful inoculations of a mouse carcinoma in mice, in
different generations and in different strains of the same
generation, they noticed definite rhythmic variations in
the number of successful inoculations, a maximum of suc-
cessful inoculations in one generation being followed by
a minimum in the succeeding generation. However, they
also state that parallel strains of the same tumor did not
show maxima or minima at the same time. Bashford,
Murray and Bowen, in order to explain these observations,
assumed that different parts of the same tumor show
different degrees of gsowth energy at the same time; this
would imply that such areas differing in growth energy
at the same period are separated through transplanta-
tion; so that in one generation mainly energetically grow-
ing pieces are used for transplantation in the succeeding
weakly growing pieces almost altogether, an assumption
which does not appear very probable.

Calkins! held that there occur in succeeding genera-
tions not so much rhythmic variations in the number of
successful transplantations as in the growth energy of
the tumors. He compared these rhythmic variations with
the rhythms observed by him in the case of Paramecium.
However, such rhythms were not noticeable in the mouse
carcinoma which we have propagated for a number of
years in our laboratory, as has been shown by Moyer S.
Fleisher.‘* He furthermore shows that even in the case
of those tumors, on the study of which Bashford and his
collaborators and Calkins base their conclusion, it is very
probable that the variations which these authors ob-
served do not represent definite rhythms, but are, as far

12 Bashford, Murray and Bowen, Zeitsch. f. Krebsforschung, 1907, Bd. 5,
Heft 3.

18 Calkins, Jour. Exper. Med., 1908, X,

14 Moyer S, Fleischer, Zeitschrift f. So Se ae Bd. 14, Heft 1, 1914.

No. 581] GERM CELLS AND SOMATIC CELLS 299

as their conclusions are not based on methods of deter-
mining the growth energy of tumors, not suitable for this
purpose, in all probability merely the expression of the
existence of a number of uncontrolled variable factors.
Such factors are numerous and they may explain certain
variations observed in growth energy and number of takes
in transplantations undertaken at different times or in
different mice. There exists, therefore, at the present time
no evidence making even probable the existence of
rhythms of growth and vitality in somatic cells com-
parable to those found in protozoa; neither have thus
far been found in somatic cells indications of nuclear
changes similar to those periodically occurring in germ
cells and probably also in some protozoa and apparently
bearing some relation to variations in growth and vitality
in these cells. At present we must therefore reckon at
least with the possibility that the immortality in somatic
cells is not connected with rhythms in vitality and in nu-
clear changes of such a character as observed in the other
two kinds of potentially immortal cells.

II. External factors acting on an organism may exert
an influence on its germ cells and here produce certain
changes which may be transmitted to the following gen-
erations, and thus through a number of generations the
offspring may show deviations from the type, although
the character of the lesions appearing in different gener-
ations may not be identical. This has been observed as
a result of the action of poisons such as lead and alcohol.
Especially the extensive investigations of Stockard on the
action of aleohol in guinea pigs demonstrate conclusively
that defects appear through several generations. In the
case of these injuries transferred to the offspring, it is
doubtful whether and to what extent these inheritable
defects are characteristic for a certain poison, or whether
we have to deal with traumatisms which might be caused
in a similar manner by many poisons or even by injurious
physical agencies. There is some evidence tending to
show that most diverse chemicals may influence embryonic
development in a similar manner, that they may produce

300 THE AMERICAN NATURALIST [Von. XLIX

identical defects in the developing organisms. It seems
to be otherwise in the case of certain external conditions
which produce first changes in some somatic cells which
on their part apparently induce such secondary changes
in the germ cells that in the offspring, not exposed to the
external conditions that affected the parents, again
changes appear in some somatic cells similar to those
produced in the parents through the external conditions.
Such results were published by Kammerer. In the latter
ease the changes produced and transmitted to the off-
spring showed evidently a characteristic specific relation-
ship to definite external conditions. Such a specific rela-
tionship is, as far as we can judge at present, lacking in
the case of defects or deficiencies produced through the
action of poisons.

I wish to report briefly on a change which my collab-
orators Moyer S. Fleisher, Miguel Vera and myself!® have
produced in somatic cells, a change which is transferab'e to
the following cell generations and is therefore hereditary,
and which, while it would be pronounced non-specific, if
we should use ordinary criteria, can through the use of
special methods be shown to possess a definite, character-
istic relationship to the external factor that caused this
change and must therefore be called specific. The obser-
vations on which these conclusions are based are briefly
as follows:

If we inoculate mice with the mouse carcinoma used by
us in our experiments and from the ninth to the fourteenth
day after inoculation give on successive days four intra-
venous injections of such substances as colloidal copper
or hirudin, a marked inhibition in growth takes place
during the period of injection. The intensity of this
inhibition varies in different cases and it is possible for
us by using a combination of two substances to cause a
retrogression of a considerable number of tumors. Now,

15 Moyer S. Fleisher and Leo Loeb, Jour. Exper. Med., Vol. XX, 1914,
p. 503. Moyer S. Fleisher, Miguel Vera and Leo Loeb, Jour. Exper. Med.,

Vol. XX, 1914, p. 522. Moyer S. Fleisher and Leo Loeb, Jour. Exper. Med.,
Vol. XXI, 1915, p. 155.

No. 581] GERM CELLS AND SOMATIC CELLS 301

if we inject daily from the second to the fifth day after in-
oculation mice with either of these two substances, tumor
growth is not noticeably retarded through the injections.
Tumors, in an early stage of development, are resistant
to this inhibiting effect. If we now give four intravenous
injections, one on each successive day, from the second to
the fifth day, and later again to the same mice four injec-
tions from the ninth to the thirteenth day, the latter series
of injections which had been effective in other animals not
previously injected from the second to the fifth day, have
now almost completely lost their efficacy. The mice have
through the first set of injections become immune against
the action of colloidal copper and hirudin as far as the
effect of these substances on tumor growth is concerned.

We next inquired into the mechanism of this immunity,
and especially were we concerned with the place where
this immunity is produced. It was conceivable that this
took place either in some organ of the injected animal or
in the tumor cells themselves. The following experiments
showed that both possibilities were realized: If we inject
the mice on the four days preceding inoculation with
tumor, immunity is produced, at least in the case of col-
loidal copper. This proves that some organ in the host
animal contributes to the immunity, inasmuch as in this
case the preliminary injections exerted their influence
without having had a chance to act on the tumor cells.
But the tumor cells themselves also become actively im-
mune, as shown in the following manner: We inject ani-
mals with either colloidal copper or hirudin from the
second to fifth day, and again from the ninth to the
thirteenth day after inoculation. Two days after the last
injection we used the tumors of some of the injected ani-
mals for reinoculation into a new set of mice. Nine to
thirteen days after this second inoculation the mice be-
longing to the second set are injected with colloidal copper
and hirudin, respectively. Now we find that these tumors
also are almost entirely resistant against the effect of
colloidal copper and hirudin, although i in this case no pre-
liminary injections had been given to the animals which

~

302 THE AMERICAN NATURALIST [Vou. XLIX

are now the bearers of the tumors. It is therefore neces-
sary to conclude that the tumor cells used for transplanta-
tion are in this case the sole bearers of the immunity, and
that the tumor cells themselves have been actively immu-
nized. From the latter experiments we may furthermore
conclude that the immunity acquired by tumor cells is
transferred to the following generations of tumor cells,
and that therefore a hereditary transmission of a char-
acter acquired by somatic cells under the influence of
external conditions takes place. The conclusion is based
on the following consideration: The process of tumor in-
oculation consists in the transfer of a very small particle
of tumor. Very soon after transplantation most of the
transplanted tumor cells become necrotic and only a rela-
tively small number of peripheral tumor cells remain
alive. These very soon begin to proliferate, and through
their proliferation give origin to the developing tumor.
If therefore the tumors developing after transplantation `
in the new hosts are immune, the immunity must have
been transmitted from the few cells remaining alive after
inoculation to the new cell generations to whom they give
origin. A fully developed tumor represents a combina-
tion of a large number of generations of tumor cells and it
may be assumed that the later generations of tumor cells
preponderate numerically very much over the earlier
generations. Through how many generations of tumor
cells this transmission of the acquired immunity can be
propagated remains yet to be determined. From our
preliminary experiments, which are, however, not yet
definite, it appears not improbable that it extends at least
through several series of transplantations.

Both colloidal copper and hirudin inhibit tumor growth.
The sign by which we judge the effect of these substances
is therefore essentially the same in both. We might thus
be inclined to conclude that their action is identical and
that likewise the immunity which they produce is the
same; that animals having received preliminary injections
of colloidal copper would therefore be immune not only
against the action of colloidal copper but also against

No. 581] GERM CELLS AND SOMATIC CELLS 303

hirudin, and that those having received preliminary in-
jections of hirudin would also be immune against colloidal
copper. We wished to test this conclusion and undertook
therefore experiments in which we immunized animals
with one substance and examined later their immunity
not only against the substance with which they were
immunized, but also against the other substance. These
experiments showed that animals immunized with col-
loidal copper are essentially only immune against the
effect of colloidal copper, not of hirudin, and those immu-
nized with hirudin are immune against hirudin, but not
noticeably (very weakly, if at all) against colloidal copper.
The acquired immunity is therefore a specific one. This
specificity can be shown to exist if after preliminary in-
jections given from the second to the fifth day after inocu-
lation the immunity is tested through cross injections
given from the ninth to the thirteenth day. It can also
be demonstrated in the tumors transplanted into other
animals after preliminary injections in the first set of
animals. The specificity concerns therefore the immunity
which is produced in the tumor cells and probably also the
immunity in the organism of the injected animals. These
investigations prove then (1) that an acquired immunity
against the injurious. action of certain substances can be
localized in the cells concerned; (2) that this immunity
can be transferred to later cell generations; and (3) that
although the effect of two substances on the cells is appar-
ently the same, the mechanism through which this effect
is produced differs in the case of each substance, and that
therefore the immunity produced against the injurious
action of these substances is a specific one for the sub-
stances injected. We see therefore that in a similar
manner as in germ cells an effect produced through an
external agency can be transmitted to later generations;
a transmission of changes produced through an external
(chemical) agency may be transmitted to later genera-
tions also in the case of somatic cells. But the further
results we obtained in the case of somatic cells suggest the
question whether the lesion produced and transmitted

304 THE AMERICAN NATURALIST [Vou. XLIX

in germ cells may not also be specific, although different
chemicals produce apparently the same results. May
there not in germ cells, just as in somatic cells, exist a
difference in the mode of production of these lesions
through the different substances and consequently a speci-
ficity in the acquired lesion, notwithstanding the appar-
ently unspecific character of the lesion? Our work makes
it possible that this question which seems of considerable
theoretical interest may be solved in a similar manner as
in the case of the somatic cells, viz., through testing the
immunity produced through the action of chemical sub-
stances.
SuMMARY

1. It is shown that evidence similar to that which makes
probable the potential immortality of protozoa and germ
cells also exists in the case of somatic cells of metazoa.
As far as the protozoa are concerned, the discussion does
not, as seems to be assumed, concern their potential im-
mortality so much as the question whether protozoa cor-
respond to germ or to somatic cells of metazoa or repre-
sent perhaps a combination of both.

2. While in the case of tumor cells the potential immor-
tality of somatic cells has been demonstrated as definitely
as the character of the problem will ever permit, the diff-
culties standing in the way of a similar demonstration in
the case of certain other somatic cells and the means of
overcoming these difficulties are analyzed. It is partic-
ularly shown that chemical differences existing between
the body fluids of individuals belonging to the same spe-
cies are the basis of these difficulties, and that as a result
of these differences the metabolism of cells in a new envi-
ronment is modified in such a way that the behavior of
connective tissue cells and of lymphocytes is altered, and
that as a result of these alterations the death of the tissue
is brought about where it could probably have lived indef-
initely. Besides altering the character of the body fluids,
transplantation of tissues furthermore has usually an
additional injurious effect; it changes the way in which
nourishment is carried to the transplanted cells, and this

No. 581] GERM CELLS AND SOMATIC CELLS 305

may also lead to alterations in metabolism calling forth a
destructive activity of connective tissue and lymphocytes.
Thus there exist difficulties in the case of certain tissues
in demonstrating through serial transplantation their
potential immortality, in a similar manner as it has been
demonstrated in the case of strongly proliferating somatic
cells (tumor cells).

It is also shown that in the case of the germ cells which
really represent a foreign individual within a host organ-
ism mechanisms exist which prevent these injurious agen-
cies from becoming effective.

3. We must sharply distinguish between the power of
cells to grow and their power to live. While the former
seems to be destroyed through differentiation, the latter
may still exist, and we can therefore at present not deny
the possibility that even highly differentiated somatie
cells may still possess the potentiality of immortal life.

4. While in the case of protozoa and germ cells definite
cycles exist, manifesting themselves either through the
oceurrence of rhythmically occurring depressions of vital-
ity or of typical changes in the nuclei, such cycles have so
far not been demonstrated in the case of somatic cells,
particularly of tumor cells.

5. In the case of germ, cells external factors can pro-
duce certain changes, and these changes (not necessarily
identical with those originally produced through the ex-
ternal factors) can be transmitted to the offspring. It is
shown that in a similar way in the case of somatic (tumor)
cells a transmission of characters acquired under the in
fluence of external agencies to the succeeding cell genera-
tions may take place. It can be shown in the case of the
somatic cells that apparently similar changes produced
through different external agencies are really not iden-
tical, but specific. It is suggested that such a specificity
of transmitted characters may also exist in the case of
germ cells despite the apparent identity of changes pro-
duced through different external agencies.

MENDELIAN INHERITANCE OF FECUNDITY IN
THE DOMESTIC FOWL, AND AVERAGE
FLOCK PRODUCTION 1

Dr. RAYMOND PEARL

In 1912 I showed,? from extensive experimental data
that, in certain breeds of domestic poultry, winter egg
producing ability is inherited in a strictly Mendelian
manner. It was pointed out that there was much evidence
indicating that winter production was, on the whole, a
rather reliable index of total fecundity capacity. As was
to be expected, the novelty of the results presented in the
papers referred to led to their criticism from various
points of view, including that of the practical poultryman.
Most of these criticisms have been based upon some mis-
understanding of the nature of the results themselves.
Others, and particularly those of the poultry press, have
apparently been based on a purely conservative instinct
to resist the intrusion of any new idea which seems to
threaten those solid personal and editorial assets of (re-
puted) infallibility and ‘‘safe and sane’’ judgment.

It has seemed to the writer more likely to conduce to
the advancement of knowledge in this field if he went
steadily about collecting more and more concrete objec-
tive evidence rather than engaging in polemic disputa-
tions with everyone whose opinion in regard to the valid-
ity or interpretation of the earlier results chanced to
differ from his own. As a result of this policy there has
accumulated a large mass of additional experimental data
confirming and extending the results of the earlier work.

1 Papers from the Biological Laboratory of the Maine Agricultural Ex-
periment Station, No. 81.

2 Pearl, R., ‘‘The Mode of Inheritance of Fecundity in the Domestic
Fowl,’’? Jour. Exper. Zool., Vol. 13, pp. 153-268, 1912. Cf. also ‘‘The Men-

delian Inheritance of Fecundity in the Domestic Fowl,’’ AMER. NAT., Vol.
XLVI, pp. 697-711, 1912.

306

No. 581] INHERITANCE OF FECUNDITY 307

This material will be published as opportunity offers.

It is the purpose of the present paper to record certain
facts which are pertinent to a general consideration of
the problem of inheritance of fecundity, but at the same
time do not fall in the direct line of the experimental
inquiry. They are matters, in other words, which are
essentially by-products of the investigation but still have
a more or less important bearing on the interpretation,
in a broad sense,-of the whole.

I. Tue Seasonat DISTRIBUTION or A Frock Ece Propvc-
TION UNDER A MENDELIAN SYSTEM OF BREEDING AS
COMPARED WITH SIMPLE Mass SELECTION

The mean egg production per bird in the different
months of the laying year has been given by Pearl and
Surface’ in an earlier paper. Those results are based on
the weighted mean production of the flocks of Barred
Plymouth Rocks at the Maine Agricultural Experiment
Station during the ten years that a system of mass-selec-
tion was followed in breeding for egg production.

It is an obvious deduction from the results of the Men-
delian experiments recorded in the earlier papers already
referred to, that by their application it should be possible
to modify the average production of a flock over a rather
wide range, the modification being of a fixed and perma-
nent character under any definite conditions of environ-
ment and breeding. To many practical poultrymen the
only test of the validity of the conclusions reached which
has any significance, is that of average flock production.
It is obvious that from a technically critical point of view
such a test has, of itself, relatively small value in helping
to judge of the correctness of a Mendelian interpretation.
At the same time it is clear that if one takes a flock of
poultry of mixed genetic constitution in respect of fecun-
dity and aims to preserve in his breeding only animals
carrying both the factors L, and L, necessary for high

3 Pearl, R., and Surface, F. M., ‘‘A Biometrieal Study of Egg Production
in the Domestic Fowl.’’ II. Seasonal Distribution of Egg Production,’’
U. S. Dept. of Agr., B. A. I. Bull. 110, Pt. II, pp. 81-170, 1911.

308 THE AMERICAN NATURALIST [Vou. XLIX

production, there ought to result a marked and immediate
improvement in average flock production no matter what
the size of the flock.

This, as a matter of fact, is exactly what has been done
in the breeding of the flock of Barred Plymouth Rocks at
the Maine Station for several years past. No attempt has
been made to propagate low fecundity strains, after it had
once been demonstrated that this could be done. In the
work since 1912 the experimental aims have been such as
not to be at variance with the practical one of getting the
most eggs with the least trouble and expense, so far as
has concerned the Barred Plymouth Rock stock. Conse-
quently in making the matings from which the founda-
tion Barred Plymouth Rock stock was being maintained
I have each year endeavored to keep a number of differ-
ent blood lines comparatively pure for the factors L, and
L,, and then intercross these lines with one another.

The results have been highly successful from a prac-
tical point of view. This is indicated by the figures shown
in Table I and graphically in Fig. 1. These compare the
mean egg production per bird month by month under the
old system of mass-selection and under the new system
of breeding which recognizes the Mendelian inheritance
of fecundity with sex-linkage of the factor on which high
production depends. The figures for the new system are
those of the laying year 1913-14. In the laying year
1912-13 the flock had not yet attained any considerable
degree of homogeneity in respect of fecundity factors
since up to and including the preceding year low produc-
ing genetic combinations had been deliberately propa-
gated and therefore an average which included all birds
in the flock would be manifestly unfair as a test of the
practical worth on a large scale of the new systems of
breeding. The laying year 1913-14 is then the first com-
pleted year on which records are available for a fair test
of the Mendelian plan on a total flock scale.

The Barred Rock flock of the year 1913-14 included 192
birds which completed the year’s work. A number of
other birds (about 20) began the year but died before its

No. 581] INHERITANCE OF FECUNDITY 309

completion. These 192 birds were divided among three
flocks of 125 each, the other birds in each flock being cross-
breds of various sorts.

It is possible to compare these 1913-14 flock with the
old records during nine months of the year only. The
reason for this is found in the fact that the trap-nesting
season is, under the present system of management,
brought to a close with August. Furthermore a record is
now kept of the laying of the pullets in October at the
beginning of the year, whereas formerly the season’s
records did not begin until November 1. This comparison
is made in Table I. Also in this table the production for
1913-14 is compared with the best single year during the
mass selection experiment, when anything approaching a
corresponding number of birds were included,‘ and for
which all environmental conditions may be regarded as
approximately normal.’ The single year records which
come nearest to fulfilling all the conditions for a fair com-
parison with 1913-14 are those for the 100-bird pens in
the laying year 1905-06. There were two such pens and
182 birds survived through the year. There was one
small environmental accident in that year which reduced
the production in May somewhat. There were adverse
environmental influences in 1913 probably quite as effect-
ive in reducing production as anything that operated in
1905-06. The seasonal conditions, size of flock, ete., were
all fairly closely comparable with those obtaining in 1913-
14. At that time (1905-06) the flock had been under con-
tinuous mass selection for eight years.

There are a number of difficulties in the way of making
a comparison between any single year now, and the ‘‘best

4 The absolutely best single year under mass selection was 1901-02. That
year there were only 48 birds for which records are available. These were in
several respects a special lot, and can not fairly be compared with large
flocks kept under ordinary flock conditions. Cf. Pearl and Surface, loc cit.

5 Cf. Pearl, R., and Surface, F. M., ‘‘A Biometrical Study of Egg Produc-
tion in the Domestic Fowl. I. Variation in Annual Egg Production,’’ U. $.
Dept. Agr., B. A. I. Bull. 110, Pt. I, pp- 1-80, 1909, for an account of the
environmental difficulties in certain of the earlier years.

ê See Pearl and Surface, loc. cit., p. 18.

310 THE AMERICAN NATURALIST (Von. XLIX

year’’ made under the mass-selection system. In the first
place in order to make the comparison at all fair the
flocks in which the birds were kept when the records were
made must be of approximately the same size. It has
been conclusively demonstrated by earlier work in the
laboratory that egg production becomes reduced as flock
size increases. It would be idle to compare the results
now where the birds run in flocks of 125 to 150 birds per
pen with the ‘‘best’’ of those prior to 1904, when the
flocks were never larger than 50 birds each and were some-
times smaller. This restricts single year comparisons
then to the period after 1904.

In the second place, if we take the year when the total
production was highest, as the ‘‘best’’ year, we shall find,
in practically every case, that some particular month or
months of this year.will fall below the average for that
month or months. There are then two alternatives,
either, on the one hand, to take for comparison with a
single year now that year under the old system of breed-
ing which, on the whole, is the best and then make allow-
ances for disturbing factors in particular months, or, on
the other hand, to compare a single year now, month by
month, with an artificial year’s record made up by pick-
ing out the best record of each individual month regard-
less of the year in which it occurred or of the size of
flock. The second of these comparisons is obviously
artificial, since it is continued high production month
after month in the same laying year which is important.
It is of interest, however, to see the results of the com-
parison on both bases. These comparisons are made in
Table I.

The best single year 100-bird pen record in the earlier
period is, as already pointed out, that for 1905-06, having
regard to the months here compared (November to July,
inclusive). The 100-bird pen of 1904-05 made a better
record during the summer than the corresponding pens
of 1905-06, but fall considerably below in winter pro-
duction. In 1905-06 there was an environmental accident

No. 581] INHERITANCE OF FECUNDITY 311

TABLE I

MONTHLY DISTRIBUTION OF MEAN EGG PRODUCTION PER BIRD UNDER DIF-
FERENT BREEDING SYSTEMS

Weight- 1918. 14 (oe |
ed Mean org

Month Under | zed Best Month in Any Year of Mass Selection; Any | Year
ont Mass |Flocks Un- Size Flock 1913-14
elec | der Mass
tion Selection
1905-06
| 100-bird
ens )

November.| 4.63 | 5.38 6.45 (1904—05, 100-bird flock) 10.76

i ; | 9.91 (12.02 (1901-02, only 48 birds in small flocks); 14.19

January...| 11.71 | 13.27 {15.21 (1901-02, only 48 birds in small flocks)| 13.88
ocks) 13.

1

|

February. .| 10.87 | 13.39 14.46 (1905-06, n 87
March. ...| 16.11 | 17.33 18.29 (1905-06, 50-bird flocks) 19.22
April, os... 15.85 | 16.48 487 nc 50 (1901-02, oe 8 birds in small hata 18.44
May... T: | .02 (1902-03, 147 birde in small floc 16.88
R ANS ar 477 #4 "88 (1901-02, only 48 birds in small pee 14.56
Wy 0. a7 | 10.498 [14.90 (1901-02, only 48 birds in small flocks)| 14.62

(overfeeding of green food) in the latter part of April.
This adversely affected the May production. The 100-
bird pens were more affected than the 50-bird pens. Con-
sequently, in order to give every possible advantage to the
earlier period of the work, I have taken the 50-bird pen
averages for April, June and July and have graphically
interpolated the figure for May in the diagram.

The data in Table I are set forth graphically in Fig. 1.

From the table and the diagram the following points
are to be noted:

1. It is apparent that the laying in the part of the lay-
ing year covered by the statistics was distinctly better
in 1913-14 than either the weighted mean of the whole
period of mass selection, or than in the best comparable
year of the earlier period.

2. The difference is somewhat more pronounced in
respect of winter production (i. e., the laying prior to
March 1) than for any other cycle. Under the earlier
plan of breeding the average winter production was 36.12
eggs. This production corresponds reasonably closely to
the division point at 30 eggs between genetically high and

7 Average from 50-bird pens of same year (1905-06). See text.
8 Average omitted because of abnormal conditions. See text.

312 THE AMERICAN NATURALIST [ Vou. XLIX

genetically mediocre winter producers which was used in
the Mendelian analysis. In the year 1905-06 the mean
winter production was 41.95 eggs. In 1913-14 the pro-

S

EN

;

EGG PRODUCTION
hg
S

®

~

NOV DEC JAN FEB MAR APR MAY JUNE JULY

Fic. 1. Diagram comparing mean monthly egg production under different sys-
tems of breeding. The light continuous line gives the weighted means for the
earlier years, the heavy continuous line the means for 1913-14, and the dotted
line the means for 1905-06 100-bird pens. The cross-hatched area in comparison

with the Bys area indicates in the increase of the 1913-14 averages over the
earlier figure
duction in the corresponding months was 51.20 eggs
per bird.

3. It was shown by Pearl and Surface? that, on the
average, a flock of hens produces 81.73 per cent. of their
total annual yield between November 1 and August 1.
Applying this figure to the 1913-14 nine-month total of
135.82 èggs, we get for the probable production of this

9**Biometrical Study of Egg Production in the Domestic Fowl. II. Sea-
sonal Distribution of Egg Production,’’ U. S. Dept. of Agr., B. A. I. Bull.
110, Pt. II, p. 89, 1911

No. 581] INHERITANCE OF FECUNDITY 313

flock of 192 birds from November 1 to November 1 a total
of 166.18 eggs. This value, as a matter of fact, is very
close to the average production per bird of those (53)
out of the 192 which were kept over for experimental pur-
poses a second year. The corresponding total for the
weighted mean annual production over the whole period
is 128.86.

4. Taking the artificial year given in next to the last
column of the table it is seen that in 1913-14, with 125-bird
flocks, the November, December and March averages were
higher than the highest made in the corresponding months
during the mass-selection period, regardless of size of
flock or other conditions. The April, May and July aver-
ages in 1913-14 were substantially equal to the highest
made in the corresponding months under mass-selection.
The highest January, February and June averages in the
mass-selection period were from 1 to 2 eggs higher than
the corresponding months in 1913-14. Taking the totals
of the whole 9-month period compared, we have for the
artificial year, made up of the highest mean monthly pro-
duction under mass selection for each month regardless
of the year or the flock size, a total of 133.73 eggs per
bird, while that for the single year 1913-14 is 135.82.

Another comparison, which brings out some additional
facts, is set forth in Table II. Any bird laying 18 or more
eggs per month in the months November, December, Jan-
uary and.February may certainly be regarded as a high
winter producer. The proportion of such high pro-
ducers in the whole flock gives valuable additional infor-
mation to that furnished by the means, since the monthly
egg production variation curves are distinctly skew. The

TABLE II
SHOWING PROPORTION OF FLOCK LAYING 18 oR MORE EGGS IN THE SPECIFIED
MONTHS

a ` | Total Flocks 1899-1907,| _ 100-bird Flocks Flock of 1913-1914,
Month | 7 Pie Sent; 1905-1906, Per Cent. Per Cent.
November....... | 7.0 5.5 26.0
December........| 19.0 30.2 47.4
amay ee | 24.2 36.3 42.2
February........ | 22.6 36.3 31.8

314 THE AMERICAN NATURALIST [ Vou. XLIX

mean and median do not coincide. In Table II is shown
the percentage of the whole flock laying 18 or more eggs
in the months specified.

This table shows in an even more striking way than
the means in Table I the marked difference between the
flocks of the present time and those of the earlier years.
In 1913-14 nearly half the flock laid 18 or more eggs each
during December and January.

The data presented in this paper establish, I think,
the following facts:

1. There is a marked difference in the average produc-
tion per bird of Barred Plymouth Rock pullets of the
Maine Station strain at the present time, as compared
with what obtained in the earlier trap-nesting work of the
Station described by Pearl and Surface (loc. cit.).

2. This difference is in the direction of a substantially
higher mean flock production at the present tame.

3. The increase in flock production is most pronounced
in respect to winter production.

The most probable explanation of the above results
appears to the writer to be that the plan of breeding now
followed is more nearly in accord with the biological facts
regarding the inheritance of fecundity than was the plan
followed in the earlier years.

The reasons for this opinion, while not constituting
complete proof of the suggested explanation, certainly
make a strong body of evidence in its favor. They are,
summarily stated:

(a) That the increases in flock productivity have been
synchronous with changes in breeding practise.

(b) That the increases give every indication of being
permanent, there having been no tendency towards a de-
cline in flock productivity since 1908, when the simple
mass selection was stopped and breeding begun on a
progeny-test basis.

(c) That there have been no changes in management or
environmental circumstances synchronous with the in-
creases in flock production and capable of accounting for
them. The hens are housed to-day in the same houses

No. 581] INHERITANCE OF FECUNDITY 315

that they were in 1904; are fed substantially the same
feed, the only modification of the ration having been in
the direction of one less stimulating to production than
the one formerly used; are hatched in the same sort of
incubators; reared in the same yards, ete.

(d) That the most marked gains have been in that
cycle of production (winter laying) to which especial
attention was paid in the breeding.

(e) That when analyzed in terms of individual matings
the results obtained in egg production have been the re-
sults to be expected on the Mendelian hypothesis of the
inheritance of this character earlier set forth, with only
minor exceptions for which the explanation is in nearly
all cases apparent.

Il. Aw InpEPENDENT CONFIRMATION OF THE SEX-LINKAGE
OF THE FACTOR ror Hıcam FECUNDITY

Besides the results with large flocks which have fol-
lowed the practical application of the Mendelian hypoth-
esis of fecundity inheritance at this Station, numerous
poultrymen in various parts of the world have obtained
similar results. Several instances of this sort might be
cited from private correspondence. The writer has felt,
however, that such cases really contributed nothing new
in principle, and that therefore there was no special need
of calling attention to them.

There lately appeared, however, in an English poultry
paper, a note which seemed to me to be of interest on
several grounds. In the first place, it is evident that the
writer, Mr. E. N. Steane, is a careful observer, and an
experienced poultryman. In the second place, his ob-
servations on inheritance of egg producing ability appear
to be, from his point of view, entirely original and unin-
fluenced by any earlier work.

The parts of Mr. Steane’s note” which are pertinent in
the present connection are these:

10 Steane, E. N., ‘‘The Production of ‘Best Layers,’’’ The Feathered
World (London), Vol. 52, p. 285, 1915.

316 THE AMERICAN NATURALIST [ Von. XLIX

My own experience, and that of many other breeders, tends to show
that the birds hatched from high pedigree hens are not such prolific
layers as those hatched from healthy hens of an indifferent laying strain
mated to high pedigree cockerels.

For three or four seasons I bred from two-year-old white Leghorn hens
of a gold-medal laying strain mated to a cockerel of equally-good descent,
and the results, to my mind, were disappointing, and did not yield an
adequate profit on the money spent. The pullets were less prolific than
their parents, and inclined to be delicate and more or less undersized,
while the percentage of fertile eggs was lessened.

Then by a lucky chance one season I had not enough eggs from a pen
of Rhode Island Reds to fill up an ineubator, and I made up the defi-
ciency from a pen of good-sized healthy Leghorn hens of no particular
laying strain mated to a pedigree cockerel. Practically every egg from
this pen was fertile, the chickens proved strong, and the results seemed
in every way satisfactory,

This, of course, led to my systematic mating of healthy, well-grown
birds of indifferent laying strain to high pedigree cockerels, with ve
successful results. The fertility of the eggs was extremely satisfactory,
the chickens turned out strong and healthy, and the pullets on arriving
at maturity were highly prolific layers, each pullet averaging 200 eggs
and over during the first twelve months, as against about 130 from the
pullets of the high pedigree hens, many of whom also died off. In the
second year the birds did equally well, the number of eggs being main-
tained and all being of a good size.

Later, I tried the result of mating high pedigree hens to a healthy

cockerel of no special laying strain, but without success, the chickens
being healthy, but the laying results much below the average, so that
nothing was to be gained by further trials in that direction.

While being quite aware that many breeders do not agree with my
conclusions, and that a great deal also depends on the condition and en-
vironment of the birds—prolificacy being always greatly improved by
the birds having a free range, I am myself firmly convinced that such
mating makes for the production of best layers. All my experiments
were, of course, carried out under the same conditions in each ease, the
birds being kept in runs of 20 yards by 10, on well-drained, sandy soil,
with a house and scratching shed attached, and fed on the same diet as
that adopted in the recent laying competitions.

It is evident that Mr. Steane’s experience was exactly
parallel to the results of the present writer’s investiga-
tions reported in earlier papers. High producing females
did not transmit that quality directly to their daughters.
The character is sex-linked.

he only point of difference is that noted in the second

No. 581] INHERITANCE OF FECUNDITY 317

paragraph of the quotation, and I think that the explana-
tion of the discrepancy there is contained in the closing
words of the paragraph where Mr. Steane says:

The pullets were .. . inclined to be delicate and more or less under-

sized, while the percentage of fertile eggs was lessened.
This would indicate that other causes besides the breed-
ing operations were working to bring about a poor physio-
logical condition of the progeny, which is of course incon-
sistent with high productivity. Lowered fertility of eggs
is one of the best indicators of reduced vitality which can
be found.

We appear to have, in this case, a rather complete inde-
pendent confirmation by a practical poultryman of one of .
the present writer’s chief results in regard to the inher-
itance of fecundity.

II. Summary

In this paper it has been shown that:

1. There is a marked difference in average egg produc-
tion per bird of Barred Plymouth Rock pullets of the
Maine Station strain at the present time as compared
with what obtained during the period of simple mass-
selection for this character.

2. This difference is in the direction of a substantially
higher mean production at the present time, when tested
on flocks of large size.

3. The increase in flock average productivity is most
pronounced in respect to winter production, which is the
laying cycle to which especial attention has been given in
the breeding.

4. The cause of this increase in flock productivity ap-
pears, with a degree of probability which is very high and
amounts nearly to certainty, to be that the method of
breeding the stock now followed is more closely in accord
with the mode of inheritance of fecundity than was the
simple mass-selection practised in the earlier period.

5. The result announced in earlier papers that high
fecundity is a sex-linked character, for which the female
is heterozygous, has been confirmed by practical poultry-
men in their breeding operations.

SHORTER ARTICLES AND DISCUSSION

THE APPEARANCE OF KNOWN MUTATIONS IN OTHER
MUTANT STOCKS

In Drosophila ampelophila the reappearance of known muta-
tions in stocks that appear to be uncontaminated is a not un-
familiar occurrence, but we discount all such cases unless in some
way the occurrence can be controlled, because the chance of con-
tamination even with extreme care might be claimed to be greater
than the chance of mutating.

In the stock of sepia-eyed flies a few individuals with very pale
(yellowish red) eyes appeared. Sepia eyes are very dark or
black brown in color. So that the flies with the new eye color
stood out conspicuously amongst the dark-eyed sepias.

From the color of the eye it was suggested that it might be
vermilion-sepia. If this were the case it should give, when bred
to vermilion flies, vermilion-eyed offspring, because the factor for
vermilion would be common to both stocks and the stock ver-
milion would carry the normal (dominant) allelomorph of sepia.
When the test was made the offspring were vermilion. These F,’s
inbred gave in F, 122 vermilion to 39 vermilion sepia, approxi-
mately 3:1, which is the expectation for one factor difference.
The result shows that a mutation to vermilion eyes had taken
place in stock that had already sepia eyes. The resulting flies
were the double recessive vermilion sepia.

That the result is not due to contamination is evident, for had
a vermilion-eyed fly got into the sepia stock it would have
produced red-eyed (wild type) females or vermilion males. As
no red- or vermilion-eyed flies were present this explanation is
excluded.

A similar mutation took place in stock having purple eyes.
Like sepia the eye color of these flies is dark but in this case
has a distinct purplish-red tinge. Among the offspring from a
cross of a female heterozygous for purple with a pure purple
male a fiy with very pale orange-colored eyes appeared. This fly,
which was a male, was also conspicuously unlike the remainder
of the red or purple offspring of this pair.

It was at first thought possible that this was the appearance
of a mutation to cherry in a mutant stock. If this had been the
case we should expect that in a cross to a cherry female all of F,
offspring would be cherry. The test showed instead all cherry

-818

No. 581] SHORTER CONTRIBUTIONS AND DISCUSSION 319

males and all red females—the normal dominant color. This
proved then that the new fly did not contain the factor for cherry.

It was then suggested that this was a second case of the appear-
ance of vermilion in mutant stock. Since the fly was a male
several matings were possible, and it was therefore crossed to a
vermilion female. As in the previous case if vermilion were
common to both stocks, the offspring should be all vermilion.
This condition was actually found in all the F, offspring of this
second cross, and the F,’s gave vermilion and the orange-eyed
fly—now shown to be vermilion purple—in approximately 3:1
classes. This demonstrated that a mutation to vermilion had
taken place in a fly already having purple eyes, for as the cherry
cross indicated, the pure mutant stock contains the normal
dominant allelomorph to every factor except the one it shows.

In order to demonstrate that the purple factor was still pres-
ent unchanged in the germ cells of the double recessive (ver-
milion purple) fly, the original male was also mated to pure
purple-eyed stock. As in the other cases if both parents contain
the factor for purple all the offspring should be purple. This
was the actual result obtained, and it proves the original mutant
to have been the double recessive vermilion purple.

In this case also it is not possible that the result could have
been due to contamination. This would indeed be highly im-
probable when the original parents were a single isolated pair,
the female of which was a virgin when first mated. But even
had a vermilion male been able to mate with the heterozygous
female only red-eyed flies could have been produced. The ver-
milion purple combination could not occur because the germ
cells of each animal carry the normal dominant allelomorph of
the mutation in the other.

Our results in these two cases show that mutations within other
mutant stocks occur, and they also indicate that in the case of
vermilion we have a mutation which has recently reappeared
twice, T. H. MORGAN,

COLUMBIA UNIVERSITY HAROLD PLOUGH

THE EVENING PRIMROSE VARIETIES OF DE VRIES

No explanation of the ‘‘variation’’ of heterozygous plants had
presented itself until Mendel went back to the haploid generation,
and referred the differences in the progeny of heterozygotes to the »
Segregation of differences in the pollen-grains and embryo-sacs
from which the plants had arisen. He dealt, however, with plants

320 THE AMERICAN NATURALIST [ Vou. XLIX

in which all the young pollen-grains and embryo-saes had pre-
sumably an equal chance to mature.

Perhaps a parallel state of affairs exists to-day. De Vries and
others have brought to light, in the progeny of Œnotheras, a cer-
tain amount of ‘‘variation,’’ by no means so striking as can be
seen in the progeny of many variety or species crosses, but re-
markable chiefly because it could not be explained. In my opin-
ion, this ‘‘variation’’ can perhaps be explained by going back,
as Mendel did, to the haploid generation. We may, I think, pre-
sume that there are certain genetic factors concerned with the
development of the young microspore into a complete pollen-
grain, and the young megaspore into a normal embryo-sac. If
the plants are heterozygous for one or more of these factors, we
get definite ratios of normal and aborted pollen-grains, or normal
and aborted embryo-sacs. We thus have a population of haploid
individuals (microspores or megaspores) which show segregation
into viable and non-viable. We may expect ratios of normal and
aborted haploid individuals of 1:1; 1:3; 1:7; 1:15, ete., apply-
ing to either pollen-grains or to embryo-saes, or to both. When
we thus have heterozygosity of one or more factors essential for
the development of the individuals of the haploid generation, the
laws and ratios for certain characters of the diploid generation
may become very different.

According to Geerts, who seems to have made the only accu-
rate study of the point, the typical Gnothera lamarckiana aborts
one half its microspores (two from each tetrad) and one half
its embryo-saes. The simplest hypothesis demands two factors,
one essential for the development of pollen-grains, and one for
the development of embryo-saes, which factors show complete
(or nearly complete) repulsion. Then, of course, the offspring
will be permanently heterozygous for these two factors, and
also more or less for any other factors which may be linked with
them. (Linkage has been shown to exist, I think, in all plants
where it has been looked for.) With more than two such fac-
tors heterozygous, the aborted pollen-grains or embryo-sacs may
increase, and the ratios in the progenies of crosses become al-
tered. Hence, in my opinion, a promising step towards the in-
vestigation of inheritance in Œnotheras is the correct determina-
tion of the ratios of aborted and normal pollen-grains and
embryo-saes in the different amet of @nothera lamarcki-
ana, and related species. JOHN BELLING

FLORIDA AGRICULTURAL EXPERIMENT STATION

E, 1915

VOL. XLIX, NO. 58

The American

Naturalist

intended for pepek and books, etc., intended for review should be

MSS
sent ae Enur of THE AMER

rticles containing summ

RICAN ssir

RALIST, Garrison-on-Hudson, New York.
ries of research wor bearin on th

s
problems of Babagan evolution are cepectatly welcome, and will be given preference

in prere

ea reprints of ip esaa are supplied to authors free of charge.

aoe
Further i "reprints “will be supplied at cos

tions and ioien andea should be sent to the e The

dilao ipion price is four dollars a
Canadian postage

ear. Foreign postage
twenty-five pala additional.

s fifty cents and
The papas for single copies is

forty cents. The advertising rates are Four Dollars for

THE SCIENCE PRESS

NEW ee Sub-Station 84

Lancaster, Pa.

Ent d as second-class matter, Ap

* thea Pact OFF

Garrison, N. Y.

t Lancaster, Pa., under the Act of

tes of March 3, 1879.

FOR SALE
ARCTIC, aor and GREENLAND
S’ SKINS,
Well oe Low Prices
Particulars of
G. DINESEN, Bird Collector
Husavik, North Iceland, Via Leidle, England

JAPAN NATURAL HISTORY grepa
Perfect Condition and Lowest Pri
Specialty: Bird Skins, Oology, AETR E Marine
Animals and others. Catalogue free. Correspond-
ence solici
T. FUKAI, Naturalist,
Konosu, Saitama, Japan

WANTED

BIRDS OF AMERICA by J. J. Audubon, 7 volumes,
please report cash price, stating condition, binding and dates
of volumes.
F. C. HARRIS,
Box 2244 Boston, Massachusetts

The — of Chicago

provide course:
— Literature, y se ate
ommerce and Administra-

Marine Biological Laboratory
Woods Hole, Mass.

sable Facilities for research in Zoology:
ones aang tata Physiology and Bot-

po 300 pita for not over three
months. Areata tables are avail-
on for beginners in resear
o ak p under the direction

of vgpetetae ae of the taft.
for such a table is $50.0 00.
INSTRUCTION instrostion
July—August
Ph pieng and
arkie Sivas d and Bog

osophical Aspects es Biol
ie Allied Sciences is also oife ie

pasi; onl sage
es d

i Fu , Liver-
~— pe a v Pinem of Algae isnt a “ior

SUPPLY Animals and
DEPART poe oer saxt

MENT
Open the Entire Year

GEO. M. GRAY, Curator, Woods Hole, Mess
The annual announcement will be sent on application to
The Director, Marine Biological La oods Hole

.

THE
AMERICAN NATURALIST

Vout. XLIX June, 1915 No. 582

THE naira Stal OF CERTAIN INTERNAL
CONDITIONS OF THE ORGANISM IN
ORGANIC EVOLUTION

First Parer. Tue REGULATION oF THE PuHysico-CHEM-
ICAL CONDITIONS OF THE ORGANISM!

F. H. PIKE AND E. L. SCOTT
DEPARTMENT OF PHYSIOLOGY, COLUMBIA UNIVERSITY

I. Introductio

II. The anias hinis of internal ipasa in the higher organisms.
. Thermo-regulation in the higher organis

The cardiac and vascular mechanisms.

=

The concentration of sugar in the blood.
The osmotic pressure of the body fluids.
The digestive tract.
The epidermis an nd sere mechanical protective mechanisms.
. The internal secretions.
The mechanisms of er
IKL The laws of chemical equilibrium.
1. The law of mass action.
- Van’t Hoff ’s law.
. The phase rule.
IV. The interpretation of the regulatory mechanisms in terms of chemical
equilibrium.

ee T

w bo

. Applications of Van’t Hoff’s law.

. The general conditions of the reactions in the cells.
3. Stimulation in terms of chemical equilibrium.

V. General considerations and summary.

bo =

Tue desirability of an attack on the general problems
of evolution from the point of view of physiology, as well
as the general deficiency of literature in physiology bear-

1 Read before Section F of the American Association for the Advance-
ment of Science, December 30, 1914.

321

322 THE AMERICAN NATURALIST [Vou. XLIX

ing on these problems, has been recognized for some time.?
The considerations to be presented in the papers of this
series are in part the outgrowth of experimental work
on the central nervous system and its relation to the proc-
esses of evolution ;? they were in part suggested by Black-
man’s‘ paper on the manifestations of the principles of
chemical mechanics in living matter; and they have grad-
ually grown up in our minds as our attention was at-
tracted more and more to the questions involved. The
bearing upon the processes of evolution of certain of the
facts drawn from the experimental study of the compara-
tive physiology of the nervous system will, for the most
part, be presented in separate papers embodying the
experimental data. The relation of the physico-chemical
conditions of the organism, together with those nervous
mechanisms with which the maintenance of the physico-
chemical conditions is inextricably bound up, to the ques-
tions of adaptation and fitness of the organism will con-
stitute the greater part of the subject matter of this and
following papers on the general problems of evolution
from the point of view of the physiologist.

Since the inception of the work, there have appeared,
in addition to Blackman’s paper, several other papers of
interest to physiologists on certain phases of the problem
of evolution.

Woods?’ has collected the better-known cases of modi-
fication in animals and plants induced by changes in the
environment or in the general conditions of existence, as
shown by changes in the rate of growth, changes of the
external form of the body, the occurrence of artificial

2 Howell, ‘‘Problems of Physiology of the Present Time,’’ Congress
of Arts and Sciences, Universal Exposition, St. Louis, 1904, Vol. V, p.
11 of the reprint.

3 Pike, American Journal of Physiology, 1909, XXIV, p. 124; Ibid.,
1912, XXX, p. 436; Quarterly Journal of Experimental dareng AN
VII, p- 1; Popular Science Monthly, 1914, p. 403; Science, N. S., Ey.

805.

4 Blackman, Nature, 1908, LXXVIII, p. 556; AMERICAN NATURALIST,
1908, XLII, pp. 633-664,
5 Woods, Popular Science Monthly, 1910, p. 313.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 323

parthenogenesis, the modification of mental and moral
traits, and the extent of regeneration of lost parts.
When arranged in the order of their position in the
taxonomic scale, organisms show a steadily decreasing
response, with reference to these phenomena, to changes
in the environment. Woods does not indicate, except in
very general terms, the physiological mechanisms in-
volved in bringing about this diminishing effect of the
environment. Some years earlier Donaldson® had called
attention to the general lack of influence of formal edu-
cational training upon the course of later life of talented
individuals.

Julian Huxley’ in his discussion of the individual in
the animal kingdom, points out that the individual
acquires an increasing independence of the environment,.
or that the environment has a diminishing effect. Hux-
ley assigns mere increase in size and the increasing com-
plexity and efficiency of the nervous system as two of the
factors involved in the attainment of the freedom from
mere accidents.

Mathews’ has pointed out more specifically some of the
various internal mechanisms which are involved in the
acquisition of independence of the environment on the
part of the higher animals. These are, according to him,
(1) the heat-regulating mechanism, (2) the mechanism
of immunity, (3) the mechanism for rendering animals
independent of external conditions of moisture, (4) the
mechanism which renders them independent of barometric
pressure, (5) the mechanisms for reproduction and car-
ing for the young, (6) the alimentary mechanism, and (7)
the nervous system.

Henderson? has shown that the environment is an ex-
ternal thing in which certain physico-chemical conditions
are kept relatively constant, while others may vary

® Donaldson, ‘‘Growth of the Brain,’? London and New York, 1895, pp.
347, 355, 360, 365.

t Huxley, J. S., ‘‘The Individual in the Animal Kingdom,’’ Cambridge
and New York, 1912.

8 Mathews, AMERICAN NATURALIST, 1913, XLVII, pp. 90-104.

? Henderson, ‘‘ Fitness of the Environment,’’ New York, 1913.

:

324 THE AMERICAN NATURALIST [Vou. XLIX

widely, even in the same region. As long as organisms
live in the ocean or in the water generally, they are sub-
jected to certain relatively constant conditions dependent
upon the physico-chemical properties of the environment.
But it should be recognized that with organisms which
have migrated out upon the land the case is somewhat
different, as all conditions, except oxygen concentration
and pressure, vary more markedly on the land than in the
water. Temperature, in particular, varies greatly over
the land, and the poikilothermal animals have their activ-
ities greatly limited by the temperature conditions.
From Henderson’s premises, the conclusion follows
that life is what it is because the environment is what it
is. A different environment might, and in all probabil-
ity would, have resulted in, or been associated with, a dif-
ferent form of life on the earth. Certain characteristics
commonly called adaptations are, as Henderson shows,
the automatic and inevitable results of the physico-chem-
ical conditions obtaining in the environment. If suc
a view assumes that an adaptation ceases to be an adap-
tation when the manner of its origin is discovered, it
would seem that we were in need of a more precise defini-
tion of adaptation. The question arises whether other
similar characteristics that have been called adaptations
are not also the inevitable and automatic result of the
physico-chemical conditions of the environment. We may
therefore consider the question of the reality of adapta-
tion, and also whether some of the characteristics of or-
ganisms are not really adaptations,’° even if they arise
from the action of physico-chemical conditions in the en- °
vironment. Again, since, in a given environment, with es-
-= sentially the same physico-chemical conditions for all its
inhabitants, there are many animal types, it would appear
that there were some influences operative within the or-
ganism itself to produce certain characteristic reactions,
known as adaptations, to the environment. As we will
show subsequently, these facts do not in any way pre-
clude an explanation of their origin on some hypothesis
10 Mathews, loc. cit.

No.582] SIGNIFICANCE OF INTERNAL CONDITIONS 825

other than vitalism. The organisms may be, under cer-
tain conditions at least, more variable than the environ-
ment. These considerations apply also to those simple
organisms which are the cellular constituents of a larger
body. Certain of the higher vertebrates, in which, as
has been indicated, the influence of the environment is
probably less than in some of the lower forms, are more
variable than certain of the lower vertebrates.

It is a fair inference from the facts cited in the various
papers referred to, that certain characteristics of living
matter, such as its slight degree of alkalinity, and its rel-
atively great specific heat, may be regarded as the direct
automatic and inevitable results of the properties of the
environment. Other characteristics, as the peculiar type
of the nervous system, are not so obviously the direct and
inevitable result of the environment, and these may be
regarded, for some time to come, as adaptations to the
environment or to the general conditions of existence in
which the particular organism is able to live and perpetu-
ate itself. .

A statement of the general problem may now be made.
Two general classes of organisms live in a relatively con-
stant environment, so far as the general internal condi-
tions of the organism are concerned. (1) For example,
the lower marine organisms of fairly limited distribution
live in an external environment which changes but little
in temperature, osmotic pressure, inorganic salt content,
neutrality or faint alkalinity, oxygen and carbon dioxide
concentration, and soluble nitrogen compounds. Tem-
perature and amounts of light may vary somewhat
throughout the year, but the temperature changes in the
given region of the ocean are less in magnitude than the
temperature changes over a corresponding area of land.
The osmotic and inorganic salt relationships of organism
and environment alike vary but little. Nor does the or-
ganism, in general, maintain within itself, any purely
physico-chemical condition differing greatly from that in
the general external environment. This does not pre-

326 THE AMERICAN NATURALIST [Vou. XLIX

clude the origin within the marine organism of stereo-
chemical isomers, the great importance of which has re-
cently been pointed out by Reichert. Stability of con-
ditions would even favor the perpetuation of such com-
pounds as could be formed under a given set of condi-
tions. (2) On the land, as has already been indicated, the
higher animals—birds and mammals—have acquired a
relative independence of the environment as shown by
their wide distribution. Their internal conditions of tem-
perature, moisture and the like, may not only differ
greatly from the same conditions in the environment at
any given time, but the internal conditions, as we shall
show, do not change greatly when the external conditions
change. Intermediate between these two types is a third
type which lives in an environment subject to wide varia-
tions of temperature and moisture, and whose internal
temperature varies with the change of external tempera-
ture. While some of the internal conditions of these or-
ganisms remain relatively stable, others are subject to wide
variation. We have then to explain, in terms of function,
(1) what are the various mechanisms by which the higher
animals have attained this relative independence of the
environment and (2) what has been the rôle of these
mechanisms in organic evolution.

II. THE GENERAL Constancy or INTERNAL CONDITIONS IN
THE HIGHER ORGANISMS

The study of the internal physico-chemical relationships
of higher animals has led to the acquisition of a consid-
erable mass of facts concerning this phase of animal or-
ganization. These facts show that in the higher animals
there exist a number of mechanisms which interact to
bring about a remarkable constancy of internal physico-
chemical conditions during the life of the animal. It ap-
pears justifiable again to call attention to some of these
facts with a brief description of some of the mechanisms
involved in order better to emphasize some of the inter-
pretations of, and inferences drawn from, them in so far

11 Reichert, Science, 1914, N. S., XI, pp. 649-661.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 827

as these interpretations and inferences have to do with
the general problem of evolution. We may, then, first of
all briefly review the more salient of these facts, and after-
ward attempt their interpretation in terms of well-known
physico-chemical laws—e. g., the law of mass action and
the phase rule.

1. Thermo-Regulation in the Higher Organisms
We may divide the vertebrates into two general groups
on the basis of their internal or body temperature; (1)
those which maintain a relatively constant internal tem-
perature and (2) those whose internal temperature is
variable. These groups may be tabulated as follows:

TABLE I
I. Animals with a constant body temperature.
about 42° C. or 107° F. Birds.
a bi about 39° C. or 102° F. Mammals.
yeri about 37° C. or 98.6° F. Man.
II. Animals with an inconstant body temperature.

Adults or partly grown mam-
rds

(a) Animals which die when
the temperature falls? Newborn mammals and birds,
below 20° C.

(b) Animals which become
korpit Npe te me Hibernating forms—some mammals.
perature falls below

L

a Aan O Reptiles, batrachians, fishes, amorg
vertebrates; molluses, insects, and

the invertebrates generally.

active when the tem-
perature falls below
20° C. L

Looking now to the differences between lower and higher
vertebrates, to restrict ourselves at the start to a rela-
tively small part of the animal kingdom, one of the most
noticeable changes has been the development of a very
constant body temperature in the so-called warm-blooded
or homoiothermal animals (Table II). The detailed
enumeration of the body temperature of all the animals
so far observed would require too much space, but the fol-
lowing data will be sufficient to show upon what basis of
fact the statements rest:
12 Richet, Dictionnaire de Physiologie, 1898, t. III, pp. 85-86.

328 THE AMERICAN NATURALIST [Vou. XLIX
TABLE IT
TABLES OF BODY TEMPERATURE
1. Birds
No. of
Genus or Species | Observa- Temperature Observer
tions
Spairow....:-.. 1 42.1° O Davy.
UEROy A 1 42.7 avy.
Peacock. i543 2% 1 40.5 to 43.0 Davy.
Guinea fowl..... 1 43.0 Davy
Duck, domestic 110 (Mean) 42.07 Martins
Mas S RE 179 (Mean) 42.3 Marti
Te POE E, (Maximal) 43.45 Martins
ew E (Minimum) 40.8 Martins
saN Davy, Eydoux,
Mikao SE A 9 40.6 rown-Sequard,
longipennes 69 40.6 Souleyet.
Peasant). 23. . 5 42.5 ichet.
yo Ro E S 17 42.5 Mantegazza, De
rquay, Dumeril,
Davy, Prevost and
Dumas.
Pigeon oaa 600 (Mean) 41.9 Chossat.
(Noon) 42.22
(Midnight) 41.48
Piom. ese 10 ( 41.2 Corin, Van Beneden.
(Maximum) 43.6
(Minimum) 39 Daily variation 2.2
Pigeon 355.5 2: 31 (Mean) Zander.
(Maximum) 44
42.0

Pee We aie ec aD ee Se a ee

Richet!® considers it probable. that the temperature of
birds is always above 40° and never, except under ex-
treme conditions, above 44° C.

Further measurements are given by Sutherland

Simpson.'4
2. Mammals15
Genus or Species. Temperature.
Pe oe ess ee ee ee ee 89.7° C.
SOOO siao iae es ea pe pe ee 39.6
Oe co eee ee ks Ua ee 39.5
Rabla -oe Fs ies es a ee es 39.5
OGG ie ce ree b eee r tree teas ety oboe 39.2
es i cossos irriparre or re a xy ein Fk 39.2
Monkeys ...... PVub a betas E eu eee 38.3
Horsen i a a ee ee 37.7
DEOHOUKENIRG soenan As a a ee ok 30.0

13 Richet, loe.

, loe. cit., p. 87.
14 Proceedings of the Royal Society of Edinburgh, 1911-12, XXXII, pp-

19-35.
15 Richet, loc. cit., t. TII, p. 91.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 329

The monotremes occupy a peculiar position in the scale
(Table III). Their internal temperature, while relatively
constant and considerably above that of the external air, is
notably lower than that of mammals generally.'® This
group probably presents a transition stage between the

thermal animals and the homoiothermal animals
with a higher body temperature.

TABLE III

SHOWING Bopy TEMPERATURE OF THE MONOTREMES

Temperature

| Cloacal Peritoneal | External

Echidna aculeata, var. typ............ e E. WC: SLS C:
Echidna aculeata, var. typ............ | 29.5 31.5 22.0
Echidna aculeata, var. typ............ | 380.5 31.5 18.0
hidna aculeata, var. typ............ | $81.5 31.5 18.0
Echidna aculeata, var. typ. young..... 31.0 31.5 24.0
Echidna aculeata, var. typ. young..... 34.2 31.5 22.5
Echid Cuba VOR ty. oe i 34.0 36.0 31.5
Echidna aculeata, var. typ..........-. 28.3 30.0 31.5
Echidna a op WEP EV Pet es 28.3 26.9 20.0
rnithorh radost ne 24.4 26.9 20.0
ertain paradou sin es i 25.2 25.2 23.0

The normal diurnal variation in temperature of the
human has been studied by many observers. The temper-
ature varies not only with the time of day, but with the
taking of food, with age, with external temperature and
with other conditions to which we will return later. On
the average, the highest temperature is attained in the
evening between five and eight o’clock, and the lowest in
the early morning between two and six, with upper and
lower limits of 37.5° C. (99.5° F.) and 36.3° C. (97.4° F.),
respectively. The body temperature may, of course, be
greater or less without any necessary implication of dis-
ease processes, but the figures given may be considered as
fairly representative.’7 It seems well established that, in
conditions of health, the daily variation in temperature
is about one per cent. of the mean—98.4° F.—expressed in
the Fahrenheit scale. It is a peculiar fact also that the

16 Richet, p. 90, Table II, and C. J. Martin, Philosophical Transactions of

the Royal Society, London, 1903, Series B, Vol. CXCV, p. 1
17 Stewart, ‘‘ Manual of Physiology,” 6th ed., New York, 1910, p. 605.

330 THE AMERICAN NATURALIST [Vou XLIX

normal body temperature is fairly close to the upper
limits of body temperature compatible with life. A varia-
tion of five per cent. from the mean, expressed in degrees
Fahrenheit, is often cause for grave anxiety, and a varia-
tion of ten per cent. is usually fatal. Compared with this,
the variation in body temperature of fishes which live in
the water only a few degrees above the freezing point
in the winter and in the tepid water of a shallow stream
in the summer is enormous.

The highest temperatures recorded in the human, after
which recovery has occurred, are shown in Table IV

vae IV
Temperature
AE RN o e rA erino dai, ‘cited by Lom-
broso.
AB ohea enn kiar Oe Alvarenga
SEI aan Searlat Bouveret.
SSO fi. eera earlat Vicente et Bloch.
CES. ee Periostitis and pyemia Weber
43.3 .......Intermittent fever Meter cited by Seguin.
ee eee Intermittent fever Hirtz.
eens Intermittent fever Alvarenga.
Ee Intermittent fever Riess.
J Peeper ne Intermittent fever Bassanovitz.
B eee Intermittent fever Diez Obelar.
Me wis Intermittent fever ay eo cited by Richet.
BA Bo i hin Intermittent fever
OSB) Cece, Rheumatism Witson Tos cited by Seguin.
449 nan Hysterical icterus Loren
dileas Hysteria miem ka by Gilles de la
Tourette.
SD eens Hysteria Lombroso
SO Rush . Hysteria R. Visioli
BS Ge Hii . Hysteria Sciamanna.
REO ix, Cane Fracture of the cervical Brodie, Lorain.
vertebra
Se A Fracture of the 12th a Simon.
vertebra and deliriu
tremens
S98 ae a Fracture of the 6th cer- Frerichs.
vieal (19 hours after
traumatism)

The somewhat alarming but relatively innocuous mani-
festations of hysteria are attended by a high temperature
without the complicating factor of an infection, and pres-

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 331

ent, therefore, the effects of temperature alone. It will
be noted that the body temperature attained in hysteria—
45° C. or 113° F.—and 46° C. in a case of intermittent
fever with recovery is nearly as high as we have found
recorded—45.7° C. or 114.3° F. after death in a case of
tetanus, and is next to the highest recorded, as far as
our observation goes, with recovery of the patient. The
temperature of the body may, however, rise still higher
under the influence of external agents. Thus, in death
from strong electric currents, Klein'® reports a body
temperature of 132° Fahrenheit (55.5° C.) or even 140° F.
(60° C.) immediately after death.

The effect of temperature upon the contractile mani-
festations of protoplasm has been summarized by
Schäfer :19

In warm-blooded animals the phenomena cease altogether to be ex-
hibited if the protoplasm which is under observation is cooled to below
a temperature of about 10° C., although they will be resumed on warm-
ing the preparation again, and this even if it has been cooled to 0° C.,
or a little lower. And when warmed gradually, it is found that the
movements become more active as the temperature rises, attaining a
maximum of activity a few degrees above the natural temperature of the
body, although if maintained at an abnormally high temperature they
are not long continued. A temperature a little above this maximum
rapidly kills protoplasm, at least that of vertebrates, producing a stiff-
hess or coagulation in it (heat-rigor), which is preceded by a general
contraction; from this condition of rigor. the protoplasm can not be
recovered. But the protoplasm of some organisms will stand tempera-
tures approaching that of boiling water without passing into heat-rigor.
Freezing may cause destruction of protoplasm in higher animals, but
that of certain of the lower animal and plant organisms is capable of
resisting extreme cold, apparently for an indefinite time. This has also
been found true for seeds of plants (Dewar).

Frog’s muscle (gastrocnemius) reaches its maximum
efficiency at about 35° C., after which a falling off occurs
as the temperature is increased. Heat-rigor makes its
appearance at about 41° C.—about two degrees Centi-
grade above the usual body temperature of a dog (an

18 Klein, New York Medical ‘Journal, May 30, 1914.
19 Schäfer, ‘*Text-Book of Microscopic Anatomy,’’ London and New
York, 1912, pp. 68-69,

332 THE AMERICAN NATURALIST [Vou XLIX

average of 39.38° C. for 176 measurements), or slightly
less than the usual body temperature of birds (42° C.).”

The maintenance of a constant temperature is depend-
ent not upon one mechanism alone, but upon the coordi-
nated interaction of several mechanisms. The presence
of a coat of fur or feathers has long been recognized as a
factor in maintaining the constant temperatures of mam-
mals and birds. The development of such a protective
covering has many times been emphasized by evolution-
ists, and seems to be well accounted for on the theory of
natural selection. The fur or feathers tend to diminish
heat loss from the surface of the body, but have nothing
to do with the production or distribution of heat within the
body. A subcutaneous layer of fat may still further re-
duce the heat loss from the surface, as in the Cetacea.

The production of heat is directly dependent upon oxi-
dation in the muscles and glands of the body. A fall of
general body temperature is attended by increased muscu-
lar activity, as shivering, when the temperature tends to
fall unduly low. The muscular activity is dependent in
its turn upon the nervous system, and upon the supply of
oxygen and oxidizable substances through the blood.

The effect of the blood in maintaining a more nearly
constant temperature of the muscles, as well as the pro-
duction of heat by the muscles themselves, is shown by
Meade Smith’s experiments on mammalian muscles.”
Although more heat is produced in a muscle which is ĉon-
tracting than in a resting muscle, there is still some heat
production while at rest. When the artery going to a
resting muscle was tied off, the difference in temperature
between muscle and blood due to heat production in the
muscle might be as much as 0.6° C. at the end of a five-
minute period. When the circulation is intact, this dif-
ference in temperatures does not become so great. Te-
tanic stimulation of a muscle may lead to a considerable
increase in the temperature of the venous ee oe

20 Richet, ‘‘ Dictionnaire de Physiologie,’’ 1898,

Up
21 Meade Smith, Archiv für (Anatomie und) ‘Physiologie “(Du Bois
Reymond), Physiol. Abt., 1881, pp. 105-152.

No. 582] SIGNIFICANCE OF INTERNAL. CONDITIONS 333

directly from the muscle as compared with arterial blood.

The distribution of heat is accomplished by the circu-
lating fluids of the body, and particularly by the blood.
When the heat loss by radiation from the surface of the
body becomes too rapid, the contraction of the walls of the
peripheral blood vessels cuts down the quantity of blood
going to the surface and, hence, the loss of heat as well.
The constriction of the peripheral blood vessels and the
contraction of the muscles tend to restrict the lower limit
to which the body temperature may fall. The lower
the external temperature, the greater the supply of
heat from internal combustion needed to maintain the
usual temperature of the body unless the radiation be
checked by clothing or by artificial heat. It is generally
stated that the temperature of the unclothed human body
at rest may be maintained until the external temperature
falls to 27° C. (Senator). This statement, as will be
shown in a later paper, may be open to question. When
the external temperature falls below this point, shivering
or other involuntary muscular movement begins. This
relation between temperature and metabolism accounts in
large measure for the large amounts of food sometimes
consumed by Eskimos. A young vigorous Eskimo may
eat as much as four kilograms (nine pounds) of meat in
a day.??

K. E. Ranke gives another illustration of the effect of
climate upon diet in Germany and in Brazil. Allowing
himself a free choice of food, the controlling influence
being his appetite, his food requirements were 3,300 to
3,500 calories a day, when the external temperature range
was from 15° C. to 22° C. Ina dry atmosphere at 25° C.,
the fuel value of the diet fell to 2,800 calories. In an at-
mosphere with a temperature of 25° C. to 28° C. and a hu-
midity of eighty-three per cent., the heat value of the diet
fell to 1,970 calories in.a day. This diet was insufficient

22 Rink, cited by Lusk, ‘‘ Fundamental Basis of Nutrition,’? New Haven,
1914, p. 28

334 THE AMERICAN NATURALIST [Vou. XLIX

to maintain his body weight, and disturbances of his gen-
eral health appeared.?*

The cold-blooded and warm- keki animals react dif-
ferently to changes in external temperature. Thus, the
carbon dioxide output of a frog rose from 0.015 gram
per kilogram of body weight per hour when the external
temperature was 1.6° C. to 0.639 gram when the external
temperature was increased to 34° C. (H. Schultz). But
as was first shown by Pflüger and his pupils, the metab-
olism of a warm-blooded animal increases as the external
temperature is lowered. If, however,the body tempera-
ture is raised there is likewise an increase in the metabo-
lism of the warm-blooded animals. Pflüger regarded the
increase in metabolism of the warm-bloodedanimalsaccom-
panying the decrease of external temperature as a later
acquisition or as a mechanism which has gradually been
evolved in the special interest of a constant temperature.
Rubner’s measurements of the metabolism in a dog showed
an increase from 30.8 calories an hour when the external
temperature was 27.4° C. to 40.6 calories an hour when the
external temperature was lowered to 11.8° C.,—an in-
crease of about thirty-three per cent. These considera- -
tions are sufficient to suggest that the effect of similar
changes in the environment may not only not have effects
of equal magnitude upon organisms at different levels in
the taxonomic scale, but may even have opposite effects at
the two extremes of the scale.

The upward march of the body temperature is re-
stricted by the greater access of blood to the periphery
and the increased loss of heat by radiation from the sur-
face. A still greater loss of heat, in addition to the rather
constant amount lost by evaporation of water from the
lungs, is brought about by evaporation of water from the
skin or other surface of the body, such as the tongue in
dogs. The amount of water on the skin is regulated by
the activity of the sweat glands, and these, in their turn,
are under nervous control.

23 Tigerstedt, ges Book of Physiology,’’ translated by Murlin, New
York, 1906, p.

No.582] SIGNIFICANCE OF INTERNAL CONDITIONS 385

Animals unable to maintain this constant body temper-
ature throughout the year may maintain a constant tem-
perature during the warmer seasons of the year and hiber-
nate during the winter. This is particularly the case with
small mammals, whose relatively large ratio of surface to
mass may greatly facilitate heat loss, and with animals
whose supply of food is difficult or impossible to obtain
during the colder season. The body temperature of the
animal falls greatly during hibernation. The importance
of size in its relation to metabolism may be shown from
the specific energy requirements of various animals. In
general, the heat requirement of all well-nourished warm-
blooded animals is about 35 calories per hour for each
square meter of body surface. But since the surface va-
ries as the square of the dimensions of the body, and the
mass varies as the cube of these dimensions, the ratio of
surface to mass is much greater in small animals than in
large. A mouse requires 452 calories per kilogram of
body weight in twenty-four hours, while a horse requires
but 14.5 calories and a man about 24 calories in the same
time.24 To sustain a number of mice equal in weight to
a man would require more than eighteen times as much
food, measured in calories, as a man would need; and
more than thirty horses could subsist upon the same
amount of food that would be necessary to sustain a num-
ber of mice whose aggregate weight was equal to that of
one horse. This is very different from saying how much
food one mouse as large as a man or a horse would need,
and should not, under any conditions, be confused with
such a statement.

Milne-Edwards observed that in a small bird such as
the sparrow, the body temperature might be lower in
winter—40.8° C.—than in summer,—43.77° C.—the dif-
ference in temperature amounting to about 3° C.

The heat regulating mechanism of the body is, then,
not a simple one but a complex one, involving muscle and
gland, food supply and distribution of the blood, the nerv-

24 Lusk, loc. cit., p.10. `

336 THE AMERICAN NATURALIST [Vou. XLIX

ous system and the oxygen tension in the blood. We
may next consider some of these subsidiary mechanisms.

2. The Cardiac and Vascular Mechanisms

The mechanism for the distribution of the blood is in
itself a complex one, and involves (1) the mechanism con-
trolling the rate and force of the heart beat and (2) the
mechanism controlling the caliber of the blood vessels.
When the cardio-regulatory and vasomotor nervous
mechanisms are intact, a fall in blood pressure is attended
by an increase in the rate of the heart beat; and, con-
versely, when the blood pressure tends to rise, the rate
of the heart decreases. When the extrinsic nerves to the
heart are cut, these changes in the pulse rate no longer
occur as an accompaniment to the changes in blood pres-
sure. The importance to the animal of these changes in
heart rate with changes in blood pressure is shown by
the fact that rabbits and dogs whose extrinsic cardiac
nerves have been cut soon get out of breath on attempting
to run.2° Through the combined agencies of the vaso-
motor and the cardio-regulatory nervous mechanisms,
the blood pressure in all mammals so far investigated?’
and in some birds, e. g., ducks and fowls,?7 is very much
the same, that is, about equal to a pressure of one hun-
dred and twenty-five millimeters of mercury. Changes
in the caliber of the blood vessels or in the rate of the
heart beat equalize the local changes of pressure. due to
changes in muscular activity. Working glands or muscles
receive, as a rule, more blood than similar organs at rest.
This increase in blood-supply may be due in part to the
action of metabolites upon the walls of the blood vessels
of the active structure (Barcroft).

The blood pressure in the homoiothermal animals, or
the blood pressure during the periods of activity of such

25 See Guthrie and Pike, American Journal of Physiology, 1907, XVIII,
pp. 27-28, for the earlier literature.

26 Porter and Richardson, American Journal of Physiology, 1908, XXIII,

p: 131.
27 Riddle and Matthews, American Journal of Physiology, 1907, XIX,
108.

p.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 837

of them as hibernate, is significantly higher than in the
poikilothermal animals. The precise significance of this
higher pressure in the homoiothermal animals is un-
known. It has been suggested that a certain pressure is
necessary to overcome the friction of the blood against
the walls of the blood vessels. It would appear that fully
as much friction might be encountered in the vessels of a
turtle weighing thirty or more kilograms as in the vessels
of a guinea-pig weighing less than one kilogram. Yet
the guinea-pig has the higher blood pressure. Nor does
the difference in blood pressure appear wholly due to
mere differences in viscosity of the blood of the two
forms.

The general stages, from the point of view of function,
in the phylogenetic development of the vascular system
have been indicated elsewhere.?®

In connection with the question of the rôle of a more
or less constant blood pressure in the animal economy,
we may mention the experiments of Legallois, Schiff and
Goltz.?° These investigators found that, while the cells
of an animal do not die immediately after the blood pres-
sure reaches a low level, the life of the cells in such an
animal as the frog is not possible for indefinite periods of
time, and in rabbits or dogs, death is a matter of hours.
We may, perhaps, imagine that the low blood pressure
may give rise to changes in the chemical systems in the
cell that are incompatible with indefinite existence.

3. The Respiratory Mechanism

The respiratory movements in mammals, and probably
also in birds, are, as Haldane and Lorraine Smith have
shown, kept up at such a rate as will maintain a constant
tension of carbon dioxide and oxygen in the alveoli of the
lungs, and presumably in the blood leaving the lungs.
The oxygen and carbon dioxide content of arterial blood
may be supposed to be fairly constant in any one indi-

28 Pike, Quarterly Journal of Experimental Physiology, 1913, VII, p. 23.

29 The literature on the effects of low blood pressure has been given
in the American Journal of Physiology, 1912, XXX, pp. 444-446.

338 THE AMERICAN NATURALIST [ Vou. XLIX

vidual. The reaction of the body fluids is likewise de-
pendent, in some degree at least, upon the tension of
oxygen and carbon dioxide in the blood.

The reaction of the body fluids, particularly the blood,
remains remarkably constant during the life of the ani-
mal. This is not so much the peculiarity of the higher
organisms as is the constant temperature. The process
of regulation of neutrality, as Henderson has shown, is a
physico-chemical process and depends upon the proper-
ties of carbon dioxide, bicarbonates and phosphates in so-
lution. The changes in concentration of hydrogen and
hydroxyl ions in the blood are, in their turn, related to
the respiratory rhythm.

The whole subject has been so well summarized by
Haldane* that I venture to quote his statement entire.

To illustrate this point I may perhaps refer to a subject which we
have recently been investigating at Oxford. We have found that the
respiratory center is so extremely sensitive to any increase or diminu-
tion of the partial pressure of carbon dioxide in the blood that a
diminution of 0.2 per cent. of an atmosphere, or 1.5 mm. of mereury will
cause apnea, while a corresponding increase will double the breathing.
The recent researches of Hasselbach have afforded experimental evi-
dence of what had already seemed very probable—that the stimulus to
which the center responds is the difference in hydrogen ion concentra-
tion, or acidity, brought about by the very slight deficiency or excess of
earbon dioxide. He has also investigated quantitatively the effect on
the hydrogen ion concentration of the blood of varying the partial
pressure of carbon dioxide. From his results and ours it follows that
the hydrogen ion concentration of the blood during rest is extraor-
dinarily constant, and remains so day by day and year by year. As
the amount of acid and alkali passing into the blood from the food and
other sources is constantly varying, it follows that the regulation of
hydrogen ion concentration is mainly brought about by the kidneys.
It has been known for long that the urine varies in acidity or alkalinity
aceording to the diet; but Hasselbach has measured the actual varia-
tions in hydrogen ion concentration. Putting together his conelusions
and ours, it appears that during ordinary resting conditions the varia-
tions in hydrogen ion concentration of the urine are about a hundred
thousand times as great as those of the arterial blood.

Thus the kidney epithelial cells react so delicately to variations in

30 Haldane, ‘‘ Mechanism, Life and Personality,’’ New York, 1914, pp.
9-51.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 389
hydrogen ion concentration of the blood, that the very smallest varia-

liquid which is, filatinniy speaking, intensely acid or alkaline, the net
result being that the normal hydrogen ion concentration of the blood
remains practically constant.

en we have such figures before us we realize the marvellous fine-
ness of the regulation by the kidneys and respiratory center. Physi-
ologists are still so much under the influence of the old gross mechanical
theories of secretion that attempts at exact measurements of the delicacy
of regulation by the kidneys have hitherto scarcely been made in the
ease of regulation in other directions, though we have every reason to
believe that similar delicacy exists as regards the regulation of the water,
salts, and other blood constituents. It is hard to realize that something
which looks under the microscope like nothing more than a somewhat
indefinite collection of gelatinous material can react, and continue
throughout life to react, true as the finest mechanism of highly tem-
pered steel, to the minutest change in its environment.

TABLE V
SOLUBILITY OF GASES IN WATER; VARIATION WITH THE TEMPERATURE
The table gives the weight in grams of the gas which will be absorbed

n 1000 grams of water when the partial pressure of the gas plus the vapor
pressure of the liquid at the given temperature equals 760 mm.

Gas 8°: C: 10°. C. 20° C. 80° C. 40° C. 50° C.
Ope .0705 0551 0443 .03 1 0263
mo .00192 00174 .001 .00147 00138 00129
Mona, .0293 0 .0189 .0161 0139 0121
Bae a 431. 248 & :

Cl 9.97 7.29 5.72 4.59 3

Oe opr is: 3.35 2.32 1.69 1.26 0.97 0.76
5...) 7.10 5.30 3.98 —- TOTA
Na 987. : 535. 422. — —
DO oa. 228. 162. 113. 78. 54. VEGRE

Whether such a constant O, and CO, content of the
blood of all poikilothermal animals is to be found under
all conditions of existence is unknown, but the coefficients
of absorption of oxygen and carbon dioxide at varying
temperatures would indicate that the carbon dioxide and
oxygen held in solution in the body fluids would be sub-
ject to change with a change in temperature." In the
higher forms, the iron-bearing respiratory pigment, hemo-
globin, absorbs all the oxygen with which it combines

31 Table V, from Table No. 130, p. 142, Smithsonian Physical Tables.

340 THE AMERICAN NATURALIST [Vou XLIX

within wide ranges of barometric pressure. There is
some difference in the processes of absorption at high and
low barometric pressures (Haldane), but the end result is
essentially the same. No more oxygen is taken up by the
hemoglobin at pressures of two or three atmospheres
than at a pressure of half an atmosphere although the
amount of oxygen dissolved in the fluid portion of the
blood niay be greater at the higher pressures. This dif-
ference is insignificant compared with the total oxygen in
the blood. In the lower forms, manganese in the Echino-
derms, and copper in certain Crustaceæ take the place of
the iron in the respiratory pigment. The general condi-
tions under which the oxygen is carried in the blood are,
however, essentially the same in the various forms.

4. The Concentration of Sugar in the Blood

It has been stated since the time of Claude Bernard
that the blood in the portal vein at the height of digestion
contains more sugar in unit volume than the blood in the
hepatic vein. The liver of the homoiothermal animals
converts sugar to glycogen and stores it up in that form.
Further extensive storage of glycogen occurs in the
muscles.

While, under ordinary circumstances, there is good
reason to believe that the concentration of sugar in the
blood of a given animal is fairly constant for a given set
of conditions, it is known that the amounts of sugar in the
blood may vary under other conditions. Excessive
amounts of sugar in the blood are eliminated by the kid-
neys. If the concentration of sugar falls, the transfor-
mation of glycogen to sugar makes up the deficiency. The
concentration of sugar in the blood is not necessarily the
same for any two species of animals under essentially the
same conditions. Nor does the concentration of the sugar
in the blood remain constant in any one individual, cat,
dog, or human, under all conditions.**

32 Cannon, American Journal of Physiology, 1914, XXXIII, p. 257;
Seott, Ibid., 1914, XXXIV, p. 271; Shaffer, Journal of Biological Chem-
istry, 1914, XIX, p. 297:

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 341

It is probable that an organ or organs having a glyco-
genic function exist in some invertebrates, e. g., the
oyster, since glycogen may be obtained from these ani-
mals. Whether such a delicate adjustment similar to
that obtaining in the higher animals between stored glyco-
gen and circulating carbohydrate exists in these forms is
unknown to us.

5. The Osmotic Pressure of the Body Fluids

In many species of homoiothermal animals, the osmotic
pressure of the blood, as measured by the depression of
the freezing point, is very constant, nor can it be easily
influenced by changes in the environment. The ingestion
of large quantities of water may be followed by the secre-
tion of large quantities of dilute urine or by profuse per-
spiration. A shortage of water to drink leads to the se-
cretion of small quantities of more concentrated urine.
Both qualitatively and quantitatively, the physico-chem-
ical constitution of the blood varies within relatively nar-
row limits. ne boa

6. The Digestive Tract 4

The mechanical and chemical processes of the alimen-
tary tract reduce all the food, or its digestible portions,
to water-soluble substances, in which form they are ab-
sorbed. The proteins of the food are reduced to the
amino acids or polypeptids, the starches and sugars to
monosaccharides and the fats to soaps before being ab-
sorbed.

The introduction of foreign proteins, e. g., egg albumen,
as such directly into the blood of the host is injurious.
But the amino acids and polypeptids to which this pro-
tein is hydrolyzed by the action of the digestive enzymes
are, in general, qualitatively the same as those in the blood
of the host. The difference in proteins depends upon the
difference in the quantitative relations of the amino acids
in the protein molecule as well as upon the qualitative
differences. The hydrolysis of all proteins to their

342 THE AMERICAN NATURALIST [Vou. XLIX

simpler splitting before absorption is a mechanism of
protection against the entrance of foreign protein into
the blood. ; :

The alimentary tract to a certain extent renders the
individual independent of the quantitative or stereo-
chemical constitution of the proteins of its food, the only
necessity being the presence of certain amino acids in the

od.

But efficient as the alimentary tract is, it does not ex-
clude all foreign or inutilizable substances. Cholohema-
tin for example, from sheeps’ bile, characterized by a
peculiar spectrum and hence susceptible of easy identifi-
cation, is absorbed by the dog and at least some of it is
excreted in the dogs’ bile without change.

The protection due to changes in the mucous membrane
of the alimentary tract is well shown in the case of ar-
senic. It was long supposed that the arsenic-eating
peasants of the Austrian Tyrol had become immune to
the effects of a considerable concentration of arsenic in
the blood, since they were able to take large quantities of
it by mouth without suffering the usual effects of an over-
dose. It was shown by Cloetta? that the apparent im-
munity arose from changes in the intestinal mucosa which
prevented the absorption of all but a small percentage
of the arsenic taken.

If the lower members of the fatty acid series be fed to
a dog the body fat which is stored up is softer and of a
lower melting point than usual. Excess fat in the food
may be stored up as body fat, but under certain condi-
tions, the animal may use as much fat as is taken in the
food, and the animal remains in metabolic equilibrium so
far as the fat is concerned.

Although the amount of protein necessary to sustain
life in an adult or to provide for growth in a young animal
varies with the nature of the protein taken in the food,
the animal is able to build up its own body protein from

33 Cloetta, Archiv für Experimentelle Pathologie und Pharmakologie,
1906, LIV, p. 196.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 343

any food which contains the requisite amino acids. The
power of synthesis of new amino acids in the body is
limited. Glycocoll, for example (amino-acetic acid), may
be synthesized in the body. More protein is needed as
food if its content in certain of the amino acids is low
than if its content of these necessary acids is high.** But
under all the varying conditions, within wide limits, the
animal may maintain itself in nitrogen equilibrium,
neither storing up any nitrogen in its flesh nor losing any
from its tissues. The excess of any particular amino
acid above that necessary for tissue growth or repair is
eliminated from the body. Voit’s famous 35 kilo dog
was able to maintain nitrogen equilibrium when fed on
amounts of protein varying from 480 grams of meat a
day as a lower limit up to 2,500 grams a day.

The quantities of fluid, fat and protein in the blood,
while undergoing some changes with the varying condi-
tions of nutrition, starvation and fasting, remain close to
a standard concentration as long as life lasts. Variations
of great magnitude are incompatible with the prolonged
life of the organism.

7. Epidermis and Other Protective Mechanisms

In addition to absorption, the limiting membranes of
the body have other and important functions. The en-
trance of many harmful foreign substances into the or-
ganism is prevented by the protecting epidermis, the al-
veolar mucous membrane, and the intestinal mucosa, al-
ready noted.

But, however important the exclusion of harmful sub-
stances may be, the protection of the organism through
the retention of valuable material is also of importance
and must be provided for. —

The escape of useful substances is prevented by the
skin, the liver, the alimentary tract, the lungs and the kid-
neys. The splitting products of the hemoglobin arising
from the death of the red blood cells are separated by

34 Lusk, loc, cit.

344 THE AMERICAN NATURALIST [ Vou, XLIX

the liver in the form of bile pigments, but the iron is con-
served for use in the formation of new hemoglobin. But
little iron is eliminated by the liver, although much iron
may be stored there. The milk of higher animals con-
tains but little iron, and the young mammal needs much
iron for the formation of new blood cells during the rapid
growth of early life. Provision is made for this by the
storage of iron in the liver of the fetus. As postpartum
growth proceeds, the amount of iron in the liver is re-
duced and the amount occurring in hemoglobin is in-
creased.
8. The Internal Secretions

The internal secretions form another group of chemical
conditions which vary within narrow limits in a state of
health. Many years ago Caspar Friederich Wolff ex-
pressed the idea that each organ of the body stood in the
relation of an organ of internal secretion to some other
organ in the body. The liver liberates sugar which is
necessary for the action of the muscles, and it also lib-
erates urea which has a definite relation to the action of
the kidneys. The muscles set free carbon dioxide which
acts upon a particular group of cells in the nervous sys-
tem, and nitrogenous waste products, some of which are
transformed to urea in the liver. But apart from these
more general relationships, a system of ductless glands
sets free a variety of chemical substances which bear a
rather different relation to the development and activity
of many other organs and tissues of the body.** The
relation of the adrenal gland to the myo-neural junction
of sympathetic and smooth muscle fiber*® is well known.
The dependence of the development of the secondary
sexual characters upon the internal secretion of the sexual
glands is also familiar to biologists. The internal secre-
tions in general have become a part of the internal con-
ditions of the organism and of the chemical environment
of the cells of the organism.

35 Mathews, Science, 1897, N. S., V, pp. 683-685.
86 Elliott, Journal of Phisiologs. 1905, XXXII, p. 401.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 345

9. The Mechanisms for Elimination of Waste Products

The elimination of waste products occurs promptly and
surely. The liver of higher animals not only separates
out the useless portions of the hemoglobin derived from
worn-out corpuscles (with retention of the iron-bearing
part), in the form of bile pigment, but also transforms
the ammonia compounds arising from nitrogenous metab-
olism to the relatively non-poisonous urea. The known
chemical functions of the liver? are numerous. The kid-
neys promptly remove the urea and uric acid from the
blood as these and other products of metabolism accumu-
late following activity of the cells. Certain mineral salts,
e. g., iron, are excreted by the intestinal mucosa. Some
fat, in addition to that which is not absorbed from the
food, is also excreted from the intestine.

The attempt to tabulate the various excretions of the
body reveals the fact that, although the qualitative com-
position of the bile and the urine is relatively constant,
the quantitative variations are very wide under different
conditions. The quantity of carbon dioxide exhaled dur-
ing twenty-four hours depends upon the diet and upon
the amount of muscular activity during the day. Al-
though somewhat disconcerting at first sight, in view of
the constancy of internal conditions, this very inconstancy
of the excretions is to be regarded as a consequence of the
maintenance of constant internal conditions. For it is
only by the conservation of those necessary elements or
substances whose supply is limited, and the free elimina-
tion of waste products or of other substances in excess
that the constancy of internal conditions can be main-
tained. Intake and outgo of matter and energy are
such as to maintain relatively constant physico-chemical
conditions within the organism. The greater the energy
requirements of the organism, the greater must be the
intake to meet these requirements, and the cba the
amount of waste products eliminated.

87 Hofmeister, ‘‘Chemische Organization der Zelle,’’ ein Vortrag; Braun-
schweig, 1901,

346 THE AMERICAN NATURALIST [Vou. XLIX

The list of mechanisms, organs and tissues of the animal
body and of specific organic or inorganic substances
within the body, whose conditions and concentration re-
main relatively constant, could be extended. The descrip-
tion of these phenomena bulks large in the extensive lit-
erature of physiology of to-day. Sufficient data have been
adduced to show that there is a considerable basis of fact
for the interpretations which we now wish to present.
A summary of the known facts of the internal conditions
of the organisms permits of the statement that in the
homoiothermal animals, there are, then, several mech-
anisms of extreme delicacy and great constancy under
similar conditions and varying but little under wide
changes of external conditions. The tracing out of their
development constitutes a large part of the subject matter
of comparative physiology. But their interpretation
rather than their development is the thing of main in-
terest at present. We will return to the questions of the
origin and development of these mechanisms in later
papers.

What point of view will give us the best insight into
the rôle of these mechanisms in the evolution of the verte-
brate phylum? Of what use have they been? Or are
they simply mechanisms which have arisen in the course
of evolution apparently through correlation with other
phases of development but without obvious significance
to the organism?

HOI. Tue Laws or CHEMICAL EQUILIBRIUM

Before attempting the interpretation of these mech-
anisms, or pointing out their rôle in evolution, we may
very briefly review the laws of chemical equilibrium as
exemplified in the ‘‘slow’’ reactions of the physical chem-
ist. For our present purpose, these laws may be included
under the law of mass action, Van’t Hoff’s law and the
phase rule. A fuller statement is given in Blackman’s
paper and in the text-books of physical chemistry.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 847

1. The Law of Mass Action

The mass law or law of mass action expresses the re-
lationship between the molecular concentration and the
speed of the reaction. The concentration of the substance
is usually expressed in terms of the number, either whole
or partial, of gram molecular weights or gram molecules
present in one liter of the solution. On the basis of Avo-
gadro’s law, there is the same number of molecules of
sodium chloride in a gram molecule (58.4 grams) of
sodium chloride as there are molecules of cane sugar in a
gram molecule (342 grams) of cane sugar or molecules
of oxygen in a gram molecule (32 grams) of oxygen. If
equal fractions of the gram molecular weights of two dif-
ferent substances are each dissolved in a liter of water,
there will be the same number of molecules in each of
the two solutions.

Let two substances, A and B, be present in a solution
in the concentrations (expressed in gram molecules) of
c, and Cz, respectively, at any stage of the reaction
A+B—>C-+D and let the temperature remain constant
throughout the action. The speed (S) of the forward ac-
tion expressed in gram molecules of A and B transformed
in unit time is defined by the relation c, X¢sXF=S
where F is the affinity constant. As the reaction pro-
ceeds, c, and cs and, hence S, steadily decrease, since A
and B are being continually used up. S may therefore be
taken at any time as the quantity of A and B which would
be transformed in the unit of time if the concentration
Ca and cp were maintained at a constant value by the con-
tinual addition of new substance. F is the measure of
the intrinsic activity (affinity) which is the driving force
in the reaction, and is independent of the concentration.
If unit concentrations are taken, c, = Cs =1 and F =S.
The activity, F, is thus represented numerically by the
number of gram molecules transformed in unit of time
when each reacting substance is present in unit concentra-

348 THE AMERICAN NATURALIST [Von, XLIX

tion. Since c4, cz and S may be measured at any time,
F may be calculated for any action.*®

The law of molecular concentration or law of mass ac-
tion is: In every chemical experiment, the speed of the
action at any moment is proportional to the first, or some
higher, power of the molecular concentration, at that time,
of each interacting substance and to the intrinsic activity
(affinity) of the substances.

2. Van’t Hoff’s Law

But the speed of any reaction at any concentration
varies with the temperature. In general, the increase in
speed is about ten per cent. for each increase of one de-
gree Centigrade, or, as it is sometimes expressed, the
speed of the reaction is doubled when the temperature is
increased ten degrees Centigrade. This is known as Van’t
Hoff’s law. The actual change in the speed of the re-
action may be greater or less than ten per cent. for each
change of one degree Centigrade, and is usually expressed
by a coefficient. When the coefficient is 1.2 or less, that
is, when the change in speed is two per cent. or less for
each change of one degree Centigrade, the action is
usually considered to be a physical and not a chemical
action. When the temperature coefficient is greater than
1.2, the action is commonly considered to be a chemical
action. No theoretical explanation of Van’t Hoff’s law
of change in speed with change in temperature has so far
been advanced.

These laws apply to reactions which go on at a measur-
able speed and which have been called ‘‘slow’’ reactions
by the physical chemists. These ‘‘slow’’ reactions are to
be distinguished from those reactions which proceed so
rapidly that no measurement of their speed at different
intervals is possible, or reactions of the explosive type.

38 Smith, ‘‘ Gereral Inorganie Chemistry,’’ 1st ed., New York, 1906, p. 251.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 349

3. The Phase Rule

One other principle of physical chemistry finds frequent
application in biology, and that is the phase rule devel-
oped by Gibbs. The phase rule defines the condition of
equilibrium existing in a system by the relation between
the number of coexisting phases and components. ‘‘Ac-
cording to it a system made up of n components in n + 2
phases can only exist when pressure, temperature and
composition have definite fixed values; a system of n
components in »+1 phases can exist so long as only
one of the factors varies and a system of n components
in n phases can exist while two of the factors vary. In
other words, the degree of freedom is expressed by the
equation

PPaCtoor P= C+3—2

where P designates the number of phases, C the number
of components, and F the degree of freedom.’’*® In other
words, F' represents the number of conditions which may
be varied without causing one of the phases to disappear.

An example of the phase rule, based upon the proper-
ties of a familiar substance, is that of ice, liquid water and
water vapor existing together in a closed vessel from
which the air has been exhausted. Ice, liquid water and
water vapor each constitute a phase of the system, and
there is but one component or substance—water—present.
Here, one component exists in three different phases.
We have, then, n components and n +2 phases. The es-
sential conditions for the existence of the system are
temperature and pressure of the water vapor. In the
notation quoted above, P=3, C=1. Hence, F=1+2
—3=0. Neither of the conditions—temperature or
pressure—of the system can be changed without causing
one of the phases to disappear. There is no degree of
freedom, or, as it is sometimes expressed, the system is a
non-variant system. The exact conditions for stability

39 Morgan, ‘‘ Physical Chemistry,’’ New York, 1911, p. 119.

350 THE AMERICAN NATURALIST [Voi XLIX

of such a system are a pressure of water vapor equal
to 4 mm. of mercury and a temperature of .007° C. above
0° C.—the freezing point at atmospheric pressure.

Many applications of the phase rule to living matter
have been made. We will cite but one. The globulins—
typical proteins found in the blood of animals—are in-
soluble in distilled water, but are soluble in dilute solu-
tions of the inorganic salts, such as sodium chloride. The
globulin may exist in a system of water, sodium chloride
and globulin, as globulin in solution or as precipitated
globulin. The globulin is the only component existing
in more than one phase under the conditions of the ex-
periment.*** Addition of water to the system to such a
degree that the concentration of the inorganic salts falls
below a certain minimum leads to a precipitation of part
of the globulin in solution. The removal of the mineral
salts, keeping the volume of the solution constant, will
also lead to precipitation wholly or in part, of the globulin
in solution. But whether water be added or salt removed,
the essential condition which undergoes changes is the
concentration of the salt.. Pressure does not enter in as
one of the conditions of the system. And if the tem-
perature of the system be raised above a certain point,
depending upon the globulin present in the system, the
globulin will be precipitated. In this system, the number
of components is one (globulin) and the number of phases
is also one, dissolved globulin. We have, therefore, a
system of n components with n phases, and two of the
conditions may vary, within certain limits, at the same
time, viz., the concentration of the sodium chloride and
the temperature. In the terminology of the quotation
above, P=1, C=1 and F=1+2~—1, or 2. This 16
also expressed by saying that the system is divariant.
This system is of interest because of the fact that it also
illustrates the phenomena of maximum points.

39a While globulin is not the only component entering into the system, we
have restricted the discussion to the department of the globulin for reasons
of space and simplicity.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 351

Further details should be sought in the text-books of
physical chemistry, and especially those by Bancroft and
Findlay on the phase rule.*°

IV. THE INTERPRETATION OF THE REGULATORY MECHAN-
ISMS IN TERMS OF CHEMICAL EQUILIBRIUM

But what evidence is there that the laws of mass action
or of chemical equilibrium apply to living matter? Is
there any evidence that the reactions occurring in the
cell are ‘‘slow’’ reactions similar to those of the physico-
chemical laboratory? The answer to these questions is
decidedly in the affirmative. Much evidence in favor of
such a view was presented by Blackman. Hofmeister,
Bredig and others regard the cell as a congeries of en-
zymes, each one, according to Hofmeister,“ acting in its
own compartment upon its own peculiar substrate.

1. Applications of Van’t Hoff’s Law

As further evidence of the nature of the reactions in
living matter, we may cite the work of Shelford‘? on tiger
beetles, in which the length of the combined quiescent
periods of the pupal and the prepupal stages was in-
creased from four or six weeks at a temperature of 28°
to 30° C. to ten or twelve weeks at a temperature of 15°
to 17° ©. Riddle! found that the temperature coeffi-
cients for digestion in Amia, Rana, Necturus and the
common turtle (Emydoidea) ranged from 0.93 in Nec-
turus to 7.81 in the turtle. Rogers and Lewis* have re-
cently shown that the temperature coefficient of the rate
of contraction of the dorsal blood vessel of the earthworm
is of the order of magnitude to be expected if the processes

40 Bancroft, ‘‘ The Phase Rule,’’ Ithaca, New York; Findlay, ‘‘ The Phase
Rule and Its Applications,’’ 3d ed., 1911, London and New York.

41 Hofmeister, loc. cit.

42 Shelford, Linnean Society’s Journal, 1908, XXX, p. 176.

43 Riddle, American Journal of Physiology, 1909, XXIV, pp. 447-458.

44 Rogers and Lewis, Biological Bulletin, 1914, XXVII, p. 269. See also
Lehenbauer, ‘‘ Physiological Researches,’’ 1914, I, pp. 247-288,

352 THE AMERICAN NATURALIST [VoL XLIX

concerned are of the nature of slow chemical reactions.
The application of Van’t Hoff’s law in these instances
is sufficiently plain.

Considering the processes in living matter, from this
point of view, we may gain some insight into the reason
why so many of the factors or conditions entering into
the reactions occurring in the body of a higher organism
should be kept as nearly constant as possible.

2. The General Conditions of the Reactions in the Cells
_ In determining the velocity of a reaction, we may deter-
mine (1) what quantity of the reacting substances com-
bine or react in unit time, the usual method of the labora-
tory, as has been shown above, or (2) we may determine
what quantities of material must be added in unit time
to keep the reaction going at a constant rate. Recalling
now the nearly constant factors in the higher mammalian
organism, the oxygen content, the temperature, and the
hydrogen ion concentration all varying within relatively
narrow limits, and the variations usually being in such a
direction as to get more material to an active or working
structure in unit time, we can see that there are certain
very effective devices for maintaining a reaction at a
constant speed, which are the counterparts of the appara-
tus employed in the chemical laboratory. But the mech-
anisms in the living organism are capable of reguiating,
with a great degree of exactness, more conditions than
any artificial mechanisms so far devised in the laboratory
can control.

In the evolution of the organism the development of
the various regulating mechanisms which we have de-
scribed has brought about a set of conditions which tend
to keep the environment surrounding the cells relatively
constant. The analogy beween the reactions in the cells
and the slow reactions of the physical chemist becomes
clear. The temperature of the body being constant, the
reactions in the cells, dependent as they are, upon a con-
stant supply of material, go on at a relatively constant
rate, or at such a rate as is determined by the needs of

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 358

the organism, and a rate which is provided for by the
changing distribution of the blood.

Not all the physico-chemical conditions of cell activity
are as constant as those discussed in the second division
of this paper. Nor does the experimental interference
with certain body structures leading to known departures
from the usual conditions always entail serious results.
As an instance of this we may cite the experiments of
Ogata, who investigated the rate of absorption of pro-
tein when fed by the mouth as compared with its rate of
absorption when introduced directly into the intestine
through a fistula. Taking the nitrogen output in the
urine as the expression of the rate of absorption, the
nitrogen output rose much more rapidly after direct in-
troduction of the meat into the intestine than it did when
the meat was fed by mouth. Although the absorption of
the food was apparently more rapid than usual, the capac-
ity for adjustment on the part of the organism was not
exceeded. We may mention in passing that one function
of the stomach may be to act as a storehouse and provide
for a more gradual absorption of food than would other-
wise occur. In the terminology of this paper, there is a
less sudden entrance of constituents tending toward a
disturbance of the equilibrium when the stomach is
present than when it is absent. If food is administered
in small portions and in a finely divided state after com-
plete removal of the stomach, life goes on as usual
(Czerny). But one is hardly justified in saying that, be-
cause great and profound changes do not occur in the
organism after extirpation of the stomach, the stomach
has no important function.

A detailed consideration of the inconstant or variable
conditions and of the manner and extent to which changes
in the environment can influence all internal conditions,
must be deferred for another communication. Enough
has been said in these pages to show, in outline at least,
the essential uniformity of some important internal con-

45 Ogata, Archiv fiir Anatomie und Physiologie, 1883, p. 89.

354 THE AMERICAN NATURALIST [Vor. XLIX

ditions of the higher organism and to indicate their rôle,
on the assumption that the internal mechanisms of the
organism are physico-chemical mechanisms.

In the response of the respiratory mechanism to the
increased concentration of carbon dioxide or to lack of
oxygen in the blood, we have an instance of adaptation
which is not at once seen to be an obviously automatic
and inevitable result of the physico-chemical properties
of the environment. A striking characteristic of the re-
spiratory center is at once its sensitiveness to slight
changes in the concentration of carbon dioxide and its
tolerance to the accumulation of carbon dioxide in the
blood. The respiratory cells react to an extremely slight
increase of carbon dioxide which is insufficient to affect
the other cells, and remain sensitive to this increase after
the concentration has risen so high that the visible re-
sponses of certain other cells have ceased. The common
excitability of the respiratory and other motor nerve cells
to carbon dioxide may ‘be supposed to result from the
disturbance or change produced in a complex system by
the accumulation of one end product of the reaction, and
to this extent to be an automatic result of the physico-
chemical constitution of the cell.

The question raised by Mathison*® as to whether car-
bon dioxide is a stimulant for all nerve cells is of interest
in this connection. Carbon dioxide is certainly a stimu-
lant for the central nerve cells of the respiratory mech-
anism, but it is not necessarily a stimulant to the same
degree for all nerve cells. It is probable that all living
matter is more or less sensitive to the accumulation of
carbon dioxide since it is one of the waste products of
all destructive metabolism. The cell bodies of the respira-
tory neurones, by reason of the development of this
common property of excitability to carbon dioxide, have
become especially adapted to respond to slight variations
in carbon dioxide. The adaptation undoubtedly depends
upon a physico-chemical change in the respiratory neu-

46 Journal of Physiology, 1910, XLI, p. 448.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 3855

rones. The persistence of the common property accounts
for the asphyxial convulsions of the spinal animals and
for the movements which are sometimes considered to be
respiratory movements and commonly attributed to so-
called respiratory centers. But whether we consider that
the cells of the respiratory group have gradually acquired
a lower threshold value for stimulation by carbon dioxide
than the other cells of the nervous system, or that the
cells of the respiratory group have simply retained the
the common excitability of protoplasm in general to car-
bon dioxide, and the remaining cells have undergone modi-
fications which have raised their threshold value, makes
little difference from the theoretical point of view. In
an environment which is, so far as one can determine,
uniform, certain quantitative variations have occurred
which have resulted in a differential sensitiveness to car-
bon dioxide or to lack of oxygen. The changes have not
been qualitative, since asphyxial convulsions involving
muscles innervated from other parts of the central nerv-
ous system, may be brought on by a reduction of the
oxygen supply.

The usefulness of the lower threshold value for lack
of oxygen in a particular group of cells is at once ap-
parent. Oxygen generation of the blood and elimination
of carbon dioxide proceed without attention, and without
noticeable excitation of any other group of nerve cells.
There is no disturbance of the precision of movement of
any group of muscles aside from those actually engaged
in the respiratory act, nor the slightest effect upon the
neurones involved in mental processes, resulting from the
decreased oxygen or increased carbon dioxide tension in
the blood sufficient to provoke a respiratory response.

3. Stimulation in Terms of Chemical Equilibrium

This brings us to a consideration of the nature of stimu-
lation in general. Lack of space precludes all but the
briefest mention at this time. We may here simply in-
dicate the consideration in terms of the laws of mass

356 THE AMERICAN NATURALIST (Vou. XLIX

action and of the phase rule. That changes in the speed
of reaction depend upon the concentration of the reacting
substances or of the end products of the reaction has
been shown in the discussion of the laws of mass action.
It is not difficult to see that the respiratory movements
owe their origin largely to changes in concentration of
carbon dioxide and oxygen, and, since these changes result
in a slight change in the concentration of the hydrogen
ions it is not difficult to imagine that the law of mass
action may be involved in the stimulation of the respira-
tory cells in the medulla oblongata. We have given in the
third section of this paper two illustrations of conditions
coming under the operation of the phase rule. It is true
that living matter undoubtedly comprises vastly more
complex systems than those described, but that the general
principles underlying the reactions are similar in most
important respects to the systems employed in labora-
tory experiments is scarcely to be doubted. The with-
drawal of water from a cell or nerve fiber by osmosis or
drying, entailing a quantitative change in the amount of
water in the cell, is followed by other changes in the cell
which tend to bring about a reestablishment of conditions
in accordance with the laws of chemicalequilibrium. That
a change of phase of some of the components occurs in
the process is probable.

Such influences as drying, applications of heat or me-
chanical pressure, whether occurring in the laboratory or
in nature, are known as stimuli, and the changes asso-
ciated with their operation in the organism are responses
to stimulation. As Jost? has pointed out, the formal
conditions of existence may act as stimuli to organisms.
Although we must admit that a wide gap still exists, it
seems to us that the discussion of stimuli and the proc-
esses of stimulation in terms of the law of mass action
and the phase rule will enable us to meet in some degree,
however small, Haldane’s objection*® that no causal con-

47 Jost, ‘‘ Pflanzenphysiologie,’’ zweite Aufl., p. 618, Jena, 1908.

48 Haldane, loc. cit., p. 37.

No. 582] SIGNIFICANCE OF INTERNAL CONDITIONS 357

nection has been shown between stimulus and response.

And we may hope that here as elsewhere in biology the

limits to our knowledge of nature will gradually be broken
own.

The accumulation of waste products in the blood or body
fluids through increase of their concentration in these
fluids, leads to modified activity of the excretory and
other organs. We cite a few examples.

The kidney, in addition to the elimination of water, fol-
lows the law of mass action in other ways. The volume
of urinary secretion, other things being equal, is propor-
tional to the volume of blood flow through the kidneys in
unit time. The greater volume of blood carries with it a
greater volume of waste products in unit time, and hence
a greater volume of secretion is the result to be expected if
urinary secretion is a physico-chemical process following
the general provision of the mass law.

The accumulation of waste products arising from the
slow reactions in the cells gives rise to the phenomena of
fatigue, and the general slowing down of the cell proc-
esses, just as the accumulation of the end products of
any slow reversible reaction decreases the amount of
chemical transformation in unit time, in accordance with
the mass law.

Excess of carbon dioxide, a typical waste product, even
of the activity in nerves,*® decreases or abolishes the con-
ductivity of a nerve fiber. A stimulus (geotropic) may be
applied to a plant in an oxygen-free atmosphere, but the
responses will not occur until the plant is moved to an
atmosphere containing oxygen.*°

But even waste products in a certain concentration
may be necessary for the optimum conditions of activity
of an organ. Baglioni®! points out that the selachian
heart maintains its activity better in a solution of the
inorganic salts containing two per cent. urea—the normal

49 Tashiro, American Journal of Physiology, 1913, xxxii, p. 107.

50 Jost, ‘‘ Pflanzenphysiologie,’’ zweite Aufl., p. 524, Jena, 1908.

51 Baglioni, Zeitschrift fiir allgemeine Physiologie, 1906, VI, p. 71.

358 THE AMERICAN NATURALIST [Von XLIX

concentration of this substance in selachian blood—than in
similar solutions without the urea.

V. QENERAL CONSIDERATIONS AND SUMMARY

The higher organisms have, therefore, developed a sys-
tem of regulation by means of which internal conditions
are kept relatively constant. This mechanism consists
essentially of a physical means of distribution of material
and heat—the circulatory organs and fluids—whose com-
position varies within narrow limits, a muscular, a glandu-
lar and a nervous mechanism for regulating the tempera-
ture, and a system of excretory organs for removing the
waste products from the circulating fluids. Both chemical
and nervous mechanisms of coordination are involved.
The variations in the composition of the circulating fluids
are such as will provide greater quantities of easily utiliz-
able material at a time when it is needed. The internal
secretions are important agents in maintaining the organ-
ism at a high pitch of efficiency through their influence
upon the neuro-muscular apparatus and the general me-
tabolism of all the tissues and organs. Regardless of the
variations in external conditions, so long as these do not
transcend the limits within which life is possible, and
barring physical accidents or disease, the internal mech-
anisms keep it always fit, whether for work or rest, for
battle or for play.

We have heard much about the survival of the fittest,
and about the rôle of the strong jaw and powerful teeth
and other physical characteristics in the struggle for
existence. The doctrine of evolution, so far as its morpho-
logical side is concerned, may be regarded as fairly well
founded. A little reflection, however, will show that the
morphological aspect is only one phase of the problem.
What profits it an animal to possess strong muscles and
sharp teeth unless these muscles shall be at all times ready
to contract quickly and surely? What if it become en-
gaged in combat with an adversary and its muscles be

No.582] SIGNIFICANCE OF INTERNAL CONDITIONS 359

sluggish from cold? Or, supposing the temperature to
be favorable, it be not able to control those muscles accu-
rately and sink its teeth into the vital spot of the enemy?
The answer is simple; another skeleton will soon lie
bleaching. Somewhere or other evolution must have been
concerned with the functional side. One protective mech-
anism has been suggested by the slow action of muscles
in the cold and their more rapid action at higher tempera-
ture. The combat. between a dog and a snake may be a
fairly even one when the weather is warm, and very
much in favor of the dog when the weather is cold. There
is a strong presumption that the elaborate and compli-
cated nervous vascular and glandular mechanisms, some
or all of which are developed in birds and mammals, have
some bearing on the general problem of evolution. It has
rendered them far more independent of the environment
than poikilothermal animals are. There is not so much
necessity of hibernation during the winter, and a frosty
morning is as good as any for hunting.

And if we consider that the changes of energy and
material underlie all the other changes in the organism,
regardless of the source from which they arise, it will be
apparent that at least one part of the final discussion of
evolution will be in terms of the changes of matter and
energy within the organism.

The problems of the general processes of evolution—
the adjustment of the animal to its environment or re-
sponses to changes in it, variation, adaptation, heredity
and geographical distribution, and even the biochronic
equation (De Vries) may all be approached from the point
of view of the experimental physiologist. The considera-
tion of these subjects will be taken up in subsequent
papers.

CORRELATION BETWEEN EGG-LAYING ACTIV-
ITY AND YELLOW PIGMENT IN THE
` DOMESTIC FOWL!

Dr. A. F. BLAKESLEE anp D. E. WARNER

CONNECTICUT AGRICULTURAL COLLEGE

So far as the presence of visible yellow pigment is con-
cerned, there are two groups of domestic fowls. In the
first group, represented by the Orpington breed, yellow is
constantly absent from legs, beak and body fat. In the
second group, represented by the Leghorns and the so-
called American breeds, such as the Plymouth Rocks,
Wyandottes and Rhode Island Reds, yellow, in the form
of yellow fat,? is present in varying amounts in the parts
mentioned. In this latter group, individual birds may
undergo considerable change in the amount of the yellow
pigment visible. The standard of the show-room, how-
ever, demands yellow in the legs and beak in these breeds
and, in consequence, birds that have become pale in these
parts are liable to be scored down by the professional
poultry judge. The paling or yellowing of the legs in the
breeds mentioned has been attributed by poultrymen to
various environmental factors. Thus, good health and
vitality, abundance of range and exercise, proper food
such as meat, corn, gluten meal and ‘‘ green food ’’ are
said to increase the amount of yellow pigment, while poor
health, moulting, confinement with insufficient exercise,
running on sandy soil and in mud, as well as climate and
the mere aging of the bird, are held to be responsible for
the paling of the legs in these varieties.

Of recent years, some individual poultrymen have
claimed that paling of the legs is due to heavy laying.
This view has been maintained by J. E. Rice.” Mr. Tom
Barron,‘ one of the most successful of the English poul-
trymen, in an address before the Connecticut Poultry

1 Paper presented in tiga se the American Society of Naturalists,
Philadelphia, December 31,

2 Barrows, H R., oe st Basis of Shank Colors in Domestie Fowl,’
Bull. 232, akak Agric. Exper. Station, 1914.

3 Circular 54, N. Y. State Dept. of Agriculture, 1912.

4 Connecticut Farmer, September 12, 1914.

360

No. 582] EGG-LAYING ACTIVITY IN DOMESTIC FOWL 361

Convention, July, 1914, described his use of the color of
the legs in selecting high egg-producers. Moreover, the
Maine Experiment Station, in a cireular® which has come
to our notice since the data in the present paper were ob-
tained, advocates a similar use of the leg color in selecting
hens for breeding.

The requirements of the ‘‘standard of perfection,”
which controls judging in the show room, as well as the
common practice of poultry breeders, are opposed to a
belief in any connection between laying and leg color.
-= Woods, under the title, ‘‘Has Leg Color Value Indi-
cating Layers?’’, discusses the subject and concludes:

Personally we believe that, as a practical guide in the selection of
heavy layers or birds from which to breed heavy layers, the leg color, of
itself, has no real value.

He further supports this conclusion by quotations from
answers received from several prominent breeders to
whom he had addressed a questionnaire on the subject.

So far as the writers are aware, no published data are
available which show in how far the leg color may be of
any value in selecting the laying hen, and such suggestions
as have been made in this connection have confined them-
selves to a consideration of the legs alone. The results
tabulated in the present paper show conclusively, it is
believed, that a close connection does in fact exist between
the yellow pigmentation in a hen and her previous egg-
laying activity. They indicate further that the color of
the beak is at least as distinctive as that of the legs here-
tofore alone considered in this connection, and that, in the
Leghorns, the color of the ear-lobes is perhaps a better
criterion of laying activity than either legs or beak and is
more readily recorded.

The hens investigated were in the egg-laying contest
located at Storrs, Conn. Pullets enter the contest Novem-
ber 1 and remain for one year. They are housed in pens
of 10 birds each, are fed the same ration and so far as

5 Circular 499, Maine Agrie. Exper. Station. This is listed as an abstract
of Bulletin 232.

® Woods, P. T., Amer. Poultry Jour., p. 35, January, 1915.

362 THE AMERICAN NATURALIST — [Vou XLIX

possible are handled exactly alike.” The influence of
different environmental factors, therefore, can be largely
neglected.

A preliminary test was made the middle of last Septem-
ber by taking from each of a number of different pens a
pair of birds representing the extremes of yellow pig-
mentation and comparing their egg records. This test
indicated that the extremely pale birds were laying and
the extremely yellow ones were not. It indicated also that
the ear-lobes were much more easily graded as to color
and in addition were apparently more indicative of egg-
laying activity than the beaks and legs. The ear-lobes of
the American breeds are red like the comb and wattles and
do not show yellow pigment. The ear-lobes of certain
other breeds, like the Blue Andalusians, are white but
apparently remain without any appearance of yellow ever
taking place. The Leghorns, including Browns, Blacks,
Buffs and Whites, show marked changes in the amount of
yellow in their ear-lobes. White Leghorns, of which
there were over 300 in the contest, were accordingly
chosen for closer study.

Ear-lobe Color in White Leghorns.—Color can be con-
veniently measured quantitatively by means of the Milton
Bradley color top, which, when spinning, acts as a color
mixer. In matching ear-lobes, only yellow and white
sectors have been used. The matching is not perfect.
especially in the lower grades, since a certain amount of
bluish tinge is often present. The amount of yellow, how-
ever, has probably been more accurately measured than
if the other color components were considered. By the
method used, it appears possible under proper illumina-
tion for one to repeat readings with a change of seldom.
more than 5 per cent. yellow above or below the mean
observation. :

Top readings were taken of the White Leghorns listed

7 Four pens of White Leghorns and four of White Rocks, belonging to the
Experiment Station, had sour milk substituted for different ingredients of
the normal ration, but, since they showed no apparent differences in color
that could be attributed to the change in the feed, they were included in the
tabulations.

No. 582] EGG-LAYING ACTIVITY IN DOMESTIC FOWL 368

TABLE I
AVERAGE EGG RECORDS FOR DIFFERENT AMOUNTS OF YELLOW IN EAR-LOBES
OF 312 WHITE LEGHORNS

Per Cent, Son | July | Aug. | Sept. | Oct. | I Sept. |II Sept.| I Oct. | II Oct. Year
| l

5-10 | 7 | 23.1 | 21.3 | 19.7 | 15.3 | 9.9 | 9.9 | 9.29 | 6.00 | 197.1

11-15 | 36 | 21.8 | 22.1 | 182 | 142] 94 | 88 | 814 | 6.03 | 187.9
16-20 | 40 | 22.2 | 20.7 | 169 | 11.7} 88 | 82 | 7.50 | 4.17 | 184.3
21-25 | 19.8 | 214 | 164 | 81 | 81 | 8&3 | 5.56 | 2.50 | 164.3
26-30 | 20 | 195 | 189 | 103 | 3.2 | 55 | 4.8 | 2.75 | 0.45 148.5
31-35 | 180. 17.7 |} $ 0.5 | 3.5 | 2.1 | 0.45 | 0.00 | 139.1
338° | 197 |173 | 61.) 0232p 42 | 19 | O16 F 0.00 11906

41-45 | 41 | 182 | 16.2 | 4. 0.2 | 34 | 1.6 | 0.22 | 0.00 | 134.2
46-50 | 39 |. 180 | 15.6 | 4. 0.2 | 26 | 1.4. | 0.18 | 0.05 | 138.2
51-55 | 30 | 184 / 161] 3.6 | 0.1 | 29 | 0.7 | 0.00 | 0.07 | 137.8
56-60 | 13 | 148 | 10.7 | 24] 0.0 | 22 | 0.2 | 0.00 | 0.00 | 124.7
61-65; 4 |145)| 88| 13) 03] 03 0 | 0.25 | 0.00 8
66-70 | 1 | 30 00) 00] 0.0) 00 | 0.0 | 0.00 0.00 70.0
71-75 | 1.| 00! 00] 0.0] 00] 0.0 F 0.0 | 0.00 | 0.00 | 83.0

TABLE IT.
coe. OF HENS LAYING AND AVERAGE NUMBER OF pine SINCE LAYING
R DIFFERENT AMOUNTS OF YELLOW IN EAR-LOBES
White Se aaa: total number of records, 932; total number of birds, 317

eS S48 Pa ae 3 A meat chet ee
Per Cent. Yellow Zial Ki | Ai J 4 i | ap es Ge oe $i 3
| j |
No. records. .... 41 | 125; 80 | 67 | 62 92| 94| 94/108 84| 44| 28; 9 | 4
Av. days since ]
TS oe 0.4 1.6 7.3/17.1 26.2 37.9 41.5 44.0 45.1 51.3 55.9 61.4 50.3 71.0
No. records = |
He eee) 86| 98/ 44| 17/3 0|1}012]0/0]0]0]0
Per cent. records i
=laying. .. . .|87.8/78.4 55.0/25.4|04.8! 0 101.0! 0 019 0/01/0100

in Tables I and II at three different periods. The first
recording took from October 7 to October 14, the second
from October 19 to 21 and the last was completed in one
day, October 28. The top records were all made by the
same one of us (B), except for 197 records on October 28.
The men who took these records had already acquired
familiarity with the method, and while their readings are
not absolutely comparable with the others, they probably
are sufficiently so to be included in Table II. The three
top readings were taken on separate sheets and the egg
records were investigated after the readings were all
taken and the birds had left the contest. Personal bias
that might have influenced the readings was thereby
avoide

Table I shows the percentage of yellow in the ear-lobes

364 THE AMERICAN NATURALIST [Vou. XLIX

of 312 birds according to the records of October 19-21,
together with monthly and yearly egg records for the
different color groups. The months of October and Sep-
tember are each divided into halves. It will be seen that
in general as the percentage of yellow increases, the egg
production falls off, and that the correlation is most
marked during the periods nearest the time when the
records were taken. A distinct though slight correlation
seems to show as far back even as July and is strikingly
evident in the yearly averages. For months before Sep-
tember and October, the correlation with color is probably
an indirect one. Itis generally only the best birds—those
that make the large yearly records—that are laying in
October. Therefore, any method that selects the laying
birds at this season will select, at the same time, the birds
laying above average throughout the year and conse-
quently give high yearly totals. It will be observed that
30 per cent. seems to be a critical amount of yellow.
Above this amount comes the sudden drop in egg produc-
tion for the months of September and October and also
above 30 per cent. yellow the yearly totals fall to between
130 and 140, with but slight change thereafter.

In Table II, the records at the three different readings
have been used. A bird laying on the day of record or on
a later day within the month is considered to be laying and
credited with a zero. If she laid on the day before the
record but not later, she is credited with one ‘‘day since
laying,” and in a similar way a longer period of inactivity
in laying is indicated by a larger number of days since
laying. With the exception of a few cases where this was
not possible, three records were taken of each bird. Since
October is the season of decreasing egg production, the
majority of the birds increased their quantum of yellow
and consequently most birds are listed in more than a
single color grade. Beginning with the 41 records in the
5-10 per cent. color grade which show an average of only
0.4 day since laying, the number of days increases con-
sistently with the amount of yellow in the ear-lobes, the
irregularity at 70 per cent. being probably due to the

No. 582] EGG-LAYING ACTIVITY IN DOMESTIC FOWL 365

smallness of the numbers in this group. The percentage
of records that indicate actual laying drops rapidly from
87.8 per cent. for 5-10 per cent. yellow to zero for grades
of yellow above 30 per cent.’ The table shows that it is
practically certain that a bird with an ear-lobe showing
more than 30 per cent. yellow at the time of the records,
is not in a laying condition.

TABLE III

AVERAGE EGG RECORDS FOR DIFFERENT GRADES OF YELLOW IN BEAKS AND
LEGs OF 256 WHITE LEGHORNS
(P, M and Y are abbreviations for Pale, Medium and Yellow)

Phe Beak Legs Suly | Aug. | Sept. | Oct. 1 Sept. Pl ‘Toet Oct.| Year
51 P P 22.0 | 20.9 | 18.6 | 14.3 | 9.6 | 9.0 | 8.0 | 6.3 |186.4
17 | M M 18.5 | 17.8 | 11.4 4.8 | 6.5 | 4.9 | 3.4 | 1.4 | 146.4
ik X 16.6 |142| 29) 04| 2.2 | 0.7 | 0.3 | 0.1 |1293
P 51 |

S i P M2 $| 22.1 | 21.0 18.3 | 14.0 | 9.4 | 88 | 7.9 | 6.1 | 185.3

; Y 0

P 25 |

43| M M17 + 20.1 | 20.6 | 13.7| 67| 7.4 | 6.3 | 4.2 | 1.5 | 164.6
Yi |
P14 |

160 | Y M49 } | 17.7 | 16.1 | 4.8 | 0.8 | 3.2 | 1.7 | 0.6 | 0.2 |135.0
Y 97 |
P51 |

90 | M25 P 21.6 [213 117.0 | 10.4 | 8.9 | 8.2 | 6.3 | 4.1 | 179.9
Y14 | | |
P 2 | | | | |

68 |{M17}| M | 19.2 | 18.3 | 7.7 | 2.1 | 4.7 | 3.0 | 1.5 | 0.6 | 142.0
Y49 | | | | | |
P O | | | | |

98 |M 1 7 16.6 | 14.3 | 2.9 | 04) 2.1 | 0.7 | 0.3 | 0.1 | 129.2
lyo7 oe oe bel

256 |Averages of totals | 19.0 | 17.8 | 9.1 | 4.3 5.2! 3.9 | 2.7 | 1.6 1504

Beak and Leg Color—The beak and legs are more
dificult to grade quantitatively than the ear-lobes. The
color is less uniform in its distribution and has more of
an orange hue, requiring the manipulation of at least one

8 The three cases of laying, among the 557 records in the grades above 30
per cent. yellow were for sporadic layers. The one in the 40 per cent. group
laid October 18, but at no other time in October or September. This case

and 19th and had no eggs to her credit in the second half of September.

366 THE AMERICAN NATURALIST EVou. XLIX

extra color disk in taking the records. A rough grouping
by inspection into the three grades, pale, medium and
yellow, however, gives a striking corroboration of the re-
sults obtained by the more accurate records on the ear-
lobes and is applicable to breeds in which ear-lobe yellow
is not present. The grading was always done by the same
one of us (W.) who has had some familiarity in handling
poultry. Probably no two observers would entirely agree
in recording the colors but the difficulty comes in delimit-
ing the grade medium,and not in deciding between the
extremes, pale and yellow. The color records were taken
on October 31 and November 1 to 4, as the birds were
being packed for shipment and their egg records were
looked up for tabulation after they had left the contest.

Table III corresponds to Table I. In the first three
rows are listed the birds that agree in beak and leg color.
In the second three rows the birds are grouped according
to their beak colors without regard to their leg colors,
while in the last three rows they are grouped according to
leg color alone.

Table IV corresponds to Table II. Since egg records
for these birds stopped on October 31, a bird laying on
October 29 is counted among the layers even if she failed
to lay on the 31st—the day she left the contest.

It will be noted from Table III that, in the Leghorns at
least, where the numbers are large enough to make com-
parisons significant, the beaks, considered alone, seem to
form a slightly better criterion for picking out the hens
with high records, while the legs alone are better in select-
ing the poorest layers. In the great majority of cases in
all the breeds considered, if the beak and the legs fail to
agree in color it is the beak that is listed the yellower. In
October the hens are falling off in laying and in conse-
quence increasing in yellow pigment. Apparently the ear-
lobes and beak are more quickly responsive to this change.
In only 97 out of 160 Leghorns for which the beak was
listed as yellow had the legs reached a similar grading
in color. :

Of the 51 White Leghorns listed in Table III as pale in

No. 582] EGG-LAYING ACTIVITY IN DOMESTIC FOWL 367

both legs and beak, 31 had ear-lobe records of 20 per cent.
or less yellow on October 28. These averaged a yearly
total of 191.9 eggs. The 40 birds of those in Table III
that on this date had 20 per cent. or less yellow in ear-
lobes, irrespective of the color of other parts, averaged a
yearly total of 189.4 eggs. It appears therefore that hens
with a higher yearly average may be obtained by selecting
those that are pale in all parts—ear-lobes and beak as we'l
as in legs—than if only one of these parts is considered.

TABLE IV
PERCENTAGE OF BIRDS LAYING, AVERAGE NUMBER OF Days SINCE LAYING
AND YEARLY TOTALS FOR DIFFERENT COLOR GRADES OF BEAKS AND LEGS
(P, M and Y are guint for Pale, Medium and Yellow; the color of
beak is written first, followed by color of legs
White beakers (256 birds with yearly average of 150.4 eggs)

P.P."| M.M.| Y.Y.| P.M. | #.Y. | M. P Er Y.P. | Y.M.
Ro bies 60 5 17° ior Gos lil ila
Av. days since laying.. a 30.4 | 57.8! 30.5} — | 20.8 64.0 | 28.6 | 45.9
No. birds laying. :...... 2 1 a
Per cent. birds laying... . rt Si 18.8) 10 0 | 12.0 0.0 | 7.2 .0
Yearly averages........ 186.4 146.4 129.3 |150.5 | — 178.7 (122.0 158.4 139.9
White, Buff and Columbian Wyandottes (79 birds with yearly average of
144.8 eggs)
| p.p. | Mm. | Y.Y. | P.M. | P.Y. | M.P. | M.¥.| Y.P. | YM.
No. bras 7 los |13 ai rilo ee ooro.
Av. days sincs laying...| 6.5| 17.5| 48.9| 0 | — | 7 | | | 287
No. bi L agile ae | 16 5 0 1 — 2 — — 1
Per cent. birds laying 57.2| 38.5) 0.0/100.0} — | 50.0 — | — |111
Yearly averages........ 178.3 1130.7 1C8.4 1194.0! — 161.51 — | — 1145.6

P.P; | M.M. | Y.Y:

P.M.

M.P.

| MY.

No.

irds
cent. birds slaying.

Yeuty av

15 | 22
25.9 | 51.6
0

27
11.3 |
wo b]

55.5 20.0) 0.0
: 1153.7 149.7 123.0

3
1.0
2

66.6
163.7 |

100.
204.0 |

5
33.8

0.0
138.8

117.5

2
6.0

(114 birds with

yearly average of 128.8 eggs)

P.P. | M.M. | y.y.| P.M. | P.Y. | M.P. | M.Y.| Y.P. | Y.M.
To Ma E Ss 3 Tis +6 1 0o03 4-19
Av. days since rame.. 14.5. 155 50.4|.— | — | 19.0| 38.0] 183) 41.2
No. birds laying ........ 2 — — 0 1
Per cent. birds ig 53.9 46.7) 36,5 — — 00i 6.0! 96) 11
Yearly averages........ - 146.3 142.5 109.8! — | — 159.0 178.0 162.0 139.2

368 THE AMERICAN NATURALIST [ Vou. XLIX

The method of grading beak and leg color may appear
crude, but that it is capable of giving valuable evidence of
previous laying activity is further shown by data kindly
turned over to us by Professor C. A. Wheeler. On Octo-
ber 26, 1912, under his direction a series of measurements
of 132 White Leghorns from the contest was taken by Mr.
R. E. Jones. Among other records, the ear-lobes were
graded as white, cream or yellow and the legs as pale or
yellow, but no connection was worked out between the
color and the egg records. These 132 birds we find to
have a yearly average of 155.1 eggs. The 34 birds with
pale legs averaged 188.9 eggs; the 98 with yellow legs,
143.5 eggs. The 33 birds with white lobes averaged 190.1,
while the 99 with cream or yellow lobes averaged 143.5.
The 21 birds that had both white ear-lobes and pale legs
averaged exactly 200 eggs.

The data presented in the foregoing pages indicate a
connection between the amount of yellow pigment showing
in a hen and her previous laying activity. The most nat-
ural assumption is that laying removes yellow pigment
with the yolks more rapidly than it can be replaced by
the normal metabolism, and in consequence the ear-
lobes, the beak and the legs become pale by this subtrac-
tion of pigment.

Environmental factors, other than laying, may be of
more or less influence on yellow pigmentation. In fact,
birds obviously sick have been observed to be pale al-
though not in a laying condition. In the material investi-
gated, however, variation in the laying activity seems to
be the prime cause of the changes in yellow pigmentation
in the domestic fowl.

The data of the present paper have been summarized
in a preliminary report in Science, March 19, 1915. Pho-
tographs showing differences in yellow pigmentation in
fowls are given in an article in the Journal of Heredity,
April, 1915.

The change in yellow pigmentation is being further
studied by a twice weekly top record of a flock of birds
throughout the year.

SOME RECENT STUDIES ON FOSSIL AMPHIBIA

Dr. ROY L. MOODIE

DEPARTMENT OF ANATOMY, UNIVERSITY OF ILLINOIS, CHICAGO

THe anatomy and relationships of the earliest air-
breathing vertebrates have interested students of fossil
animals so greatly since Georg Jaeger described the first
Labyrinthodont in 1828, that the result to-day is a biblio-
graphic list of over 600 titles, varying in importance from
the magnificent work of Fritsch (‘‘ Fauna der Gaskohle’’)
issued in four folio volumes with scores of lithographic
plates, to short notices of a few lines. Many of the mem-
oirs are handsomely illustrated and beautifully printed.
The material so far described has been extremely frag-
mentary and the greater number of the contributions is-
sued have been dedicated to the description of species
based on incomplete material. The fauna was exceed-
ingly diverse like the plesiosaurs of a later period, and
new discoveries tend to confuse rather than to unify our
ideas of amphibian morphology. The few papers re-
viewed below form no exception to the statement made
above. Many new and importants facts are brought forth
in the contributions made during the past few months and
these are well worthy of consideration. Attention in
these reviews will be paid especially to new facts of struc-
tural importance.

Broili (1) in a short paper has added to our knowledge
of the Permian fauna of Texas by the description of two
new species of Amphibia based on incomplete skulls.
One of the species is very small, the skull measuring
scarcely half an inch in length. The same writer (2) ina
more extensive paper has given a popular review of the
chief work done during the past ten years on the early air-
breathing vertebrates and has listed the important papers
from which he has used illustrations to elucidate his re-
marks. This paper should be consulted by any one who
wishes a convenient and accurate survey of the earliest
land vertebrates. Doctor Broili refers to Micrerpeton,
the first branchiosaur known from the western hemi-

369

370 THE AMERICAN NATURALIST [ Vou. XLIX

sphere, as a microsaur. The distinction between these
two groups is clear, the former undoubtedly being .ances-
tral to the modern Caudata and the latter having reptilian
affinities. Likewise the author refers to Lysorophus'
as a reptile, while the majority of paleontologists regard
the form as Amphibian; Williston! even going so far as
to locate it in the suborder Ichthyoidea of the Caudata.
In conclusion Doctor Broili says:

Im übrigen haben wir im Laufe der letzten 10 Jahre über die ältesten
Tetrapoden so viel neues und wichtiges kennen gelernt, wie wohl in
relativ keinem anderen Zweige der Wirbeltierpaliontologie. .. .

Broom (3) has given the results of his studies on Per-
mian vertebrates in the American Museum. His reason
for again describing and studying this much described
and much studied material is that structural characters
are difficult to determine in these forms on account of the
very closely adherent matrix which has in many cases ob-
secured all sutures in the skull. His discussion is accom-
panied by restorations of the skulls of the chief Permian
- genera, indicating most of the sutures, something which
Cope was unable to do. He discusses some elements in
the mandible not previously observed among Amphibia
and suggests homologies between them and elements of
the reptilian mandible. Unfortunately, Broom has paid
no attention to the occurrence of lateral line canals on the
skulls of these forms. It is highly important that this
system of sense organs be distinctly understood. In view
of Herrick’s studies? on this structure in the catfish it is
certain that this system of sensory organs has a distinct
influence on the location of the peripheral osseous ele-
ments of the skull and mandible. I do not recall that
Herrick’s result have been noted by any paleontologist,
but they should be taken into consideration. Broom says
in regard to Eryops, the large Permian stegocephalian:

. Every detail of the cranial structure can be clearly made out.
He criticizes Huene’s (1913 b) work on the brain-case,
however, and makes no statement concerning the lateral

Aegae Bull., Vol. XV, No. 5, p. 229, 1908.
2 Jou ri. Comp: Neurol., Vol. 11, p. 224, 1901.

No.582] RECENT STUDIES ON FOSSIL AMPHIBIA 371

line organs which were imperfectly studied some years
ago? by the reviewer; so it is yet too soon to say that
every detail of structure is known. The palate is very
completely known and is figured by Broom. He has fig-
ured also very imperfectly, but for the first time, sections
through the ear and brain-case showing the probable size
of the dural cavity. He says that the portion of the par-
occipital which lodges the labyrinth was cartilaginous, but
does not give his reasons for this statement. In view of
the almost perfect preservation of the semicircular canals
in fishes, cotylosaurs and pterodactyls we should expect
a favorably preserved specimen of an amphibian to show
this structure also. He describes a pit in the basisphe-
noid for the reception of the hypophysis. He also figures
for the first time the complete osteology of the mandible
of Eryops. The author likewise describes and briefly
figures two new species of stegocephalians. The same
author (4) has given considerable attention to the study
of the osteology of the mandible in Trimeorohachis, the
discussion being very similar to that given in the above
paper. The discussion has especially in view the problem
of the derivation of the Amphibia from the Crossopte-
rygia, and he figures the mandible, shoulder girdle and
pectoral fin of Sauripteris taylori on account: `

. of (the) extreme interest from having the pectoral fin more closely
resembling the tetrapod limb than in any other known form

Case (5) reviewed before the American Paleontological
Society the recent trend of studies on the air-breathing
vertebrates of the Paleozoic. He states there are two
general conclusions which have been reached by students
of these early vertebrates. First, Baur initiated the idea
of the crossopterygian ancestry of the Amphibia, and
later workers have so far confirmed his suggestion as to
make it extremely probably that the land vertebrates arose
from these fishes. The intermediate stages are unknown.
The second conelusion is
that the primitive reptiles—the Cotylosauria—were derived directly
from the Stegocephalia.
So we are thus in possession of partial proof at least of

3 Journ. Morphol., 1908, Vol. XIX, p. 511.

372 . THE AMERICAN NATURALIST (Vou. XLIX

the origin of reptiles from fishes through the Amphibia.

We owe to Doctor Fraas of Stuttgart many important
contributions to the knowledge of the early air-breathing
vertebrates and he has recently (6) issued another memoir
on the labyrinthodonts of the Trias, the first study of
these animals since the appearance of his large memoir in
1889.4 The present contribution is devoted to discussions
of new species and new facts concerning previously de-
scribed species. The Plagiosternum granulosum is found
to be the most peculiar labyrinthodont yet described, in
that it is extremely frog-like in appearance, especially in
the huge size of the orbits and the expanded occiput. It
is interesting, furthermore, in the apparent absence or
indistinct preservation of the lateral line canals. The
photograph (Plate XVI, Fig. 1) of the dorsum shows
portions of the supra- and infraorbital canals. The re-
mainder of the cephalic system of sense organs was prob-
ably contained in pits, which, in the fossilized skull, are
not to be distinguished from the ornamental scrobicula-
tions of the membrane bones of the face. The auditory
meatus is on the posterior edge of the skull and is quite
large for the size of the skull. Doctor Fraas has given in
a drawing (Plate XVI, Fig. 3) the complete osteology of
the occiput of this unusual labyrinthodont. The re-
mainder of the memoir is devoted to a discussion of new
or disputed points in the osteology of various genera and
species of Triassic labyrinthodonts.

Gregory (7) has reviewed the studies which have thrown
light on the crossopterygian ancestry of the Amphibia,
dealing especially with Watson’s (11) recent paper on the
Larger Coal Measures Amphibia, and giving a list of thir-
teen contributions which deal directly with this derivation
of the Amphibia.

Huene (8) has again described the mandible of the pe-
culiar Permian genus Diplocaulus although it has been
many times studied, described and figured. He states, in
his introductory paragraph:

Gattungen, wie Diceratosaurus, Eoserpeton, Stegops, Amphibamus,
vieleicht auch Tuditanus zeigen Verwandtschaft mit Diplocaulus.

4‘‘Paleontographica,’’ Bd. XXXVI.

No.582] RECENT STUDIES ON FOSSIL AMPHIBIA ore

Just what the basis of this relationship is he does not
state. The reviewer® has previously stated that these
above-mentioned microsaurian genera exhibit no struc-
tural features which would ally them, except remotely,
with Diplocaulus. This Permian genus has no relatives
among the Coal Measures Microsauria, the reasons for
this statement being given in the above-mentioned essay®
and need not be repeated here. The material on which
von Huene bases his paper was collected in Baylor County,
Texas, and formed a part of a collection purchased by
Doctor von Huene from Charles Sternberg at Lawrence,
Kansas. The same writer (9) has again studied the Per-
mian Lysorophus, which is regarded by Williston as
closely akin to the salamanders.’ Huene bases his dis-
cussion on twenty-four skulls in the collection of the Uni-
versity of Tübingen. He describes and figures some mi-
nute limb bones, thus partially confirming Miss Finney’s
results.” He agrees with Williston that Lysorophus is
related to the Urodeles though suggesting : 4

Mit den Temnospondylia hat der permische Urodele Lysorophus:
noch grössere Ahnlichkeit als die jetzigen Urodelen. Sie liegen in der:
Schiidelbasis und der grosseren Anzahl der hinteren Schiideldeckknochen.
The same author (10) gives the results of his studies of
Permian vertebrates at the American Museum. The
paper is illustrated by sketches of various skulls and parts.
of skulls made by the author and showing his interpreta-.
tion of the elements composing the cranium of American
Permian amphibians and reptiles. He describes and fig-
ures a stapes in a skull of Eryops and gives the results
of his study of the brain-case of this genus. The stapes
has a length of 4 em. and in shape is not unlike a human
clavicle. His studies of Lysorophus, Gymnarthrus, Di-
plocaulus and other genera confirm the results of pre-
vious students of these forms. He concludes his paper
with a discussion of morphological results, and appends
a bibliography of twenty papers.

5 Journ. Morphol., Vol. 23, p. 31, 1912.

< 229,

€ Biol. Bull., XV, 1908,
7 Journ. Morphol., 23, p. 664, 1912.

374 THE AMERICAN. NATURALIST [ Vou. XLIX

Watson (11) has restudied the skulls of some of the
European Carboniferous labyrinthodonts, Loxomma, Pter-
oplax, and Anthracosaurus, and compared them with the
Coal Measures fish, Megalichthys. His results have
already been reviewed by Gregory (7), so that it will only
be necessary here to say that these genera approach the
crossopterygian type of structure in various features.
The same author (12) has redescribed an interesting mi-
crosaur in which he is able to give a very complete account
of the structure of the dorsal and ventral surfaces of the
skull and pectoral girdle. He compares the newly recon-
strueted microsaurian with Diplocaulus and Ceraterpeton.
It is very important that these little-known species from
Europe be restudied and redescribed, so that former ob-
servations may be corrected, corroborated and extended.
The status of

The classification of the smaller stegocephalian Amphibia, so abun-

dant in the Coal Measures and Permian Rocks of Europe and North
America, is in such confusion, to which some recent work has added,
that it is at present only possible to proceed by reference to individual
specimens which have been well described.
The reviewer finds himself in hearty accord with these
statements, although he must plead guilty of having
thrown some confusion into the classification of these ani-
mals in the hope that thereby order might ensue.

Doctor Williston (14) has determined the complete os-
teology of the mandible in the early reptiles and amphib-
ians, working especially with the material from the Per-
mian of America. He says:

In the structure of the mandible the amphibians are remarkably in-

termediate between the early reptiles and the contemporary cross-
opterygian fishes, differing from the latter chiefly in the reduced num-
ber of coronoids, and from the former chiefly in the possession of two
additional coronoids and a splenial.
These results are corroborated by the studies of Doctor
Broom on similar material, so that any doubts as to the
real structure of the stegocephalian mandible are placed
at rest by the results arrived at by these ee investi-
gations.

No. 582] RECENT STUDIES ON FOSSIL AMPHIBIA 375

The mandible of the primitive amphibians differs chiefly from that
of the early reptiles in the division of the coronoid into three elements,
or possibly four, and in the division of the splenial into two.

Wiman (15) within the past three years has become
much interested in the amphibian fauna of the Trias of
Spitzbergen. In the present paper he reviews the work
which has been done on the structures of the occiput of
seven genera of Permian and Triassic stegocephalians,
figuring the anatomy of this region of a new laby-
rinthodont from Spitzbergen. He describes this new
genus in a later contribution. In this latter paper (16)
Wiman discusses the occurrence of amphibian remains in
the deposits of Spitzbergen, accompanying his remarks
by photographs of the bone-bearing horizons. His paper
deals largely with new forms from Spitzbergen, which are
illustrated in four text figures and nine photographic
plates. One is at once struck, in the examination of
Wiman’s plates, by the clearness of preservation of the
cephalic lateral line canals. The author refers to these
structures as ‘‘ Schleimkaniile ’’ and gives a very careful
description of their occurrence; the only writer of recent
date who has done so. The term Lyrocephalus euri is
proposed for the new genus and species.

Der Gattungsname bezieht sich auf die ausserordentlich kräftig
entwickelten Schleimkanile des Kopfes. .. .

He refers to the various canals as ‘‘ Tremalkanile,’’ ‘‘ Na-
sofrontalkaniile,’’? ‘‘Temporalkanal’’ and ‘‘Maxillarka-
nal,’’ but makes no attempt to homologize them on the þa-
sis of the work of Allis’ (1889) and the reviewer (1908).
The lateral line canals are so unusually well preserved in
Lyrocephalus that it is thought worth while to give an
outline figure in another place of their occurrence and to
homologize them on the basis of previous work. The
columella auris is described and figured (Plate II, Figs.
4-5) in this species. It is unusually large. Other new
forms are described from these interesting deposits, many
of the specimens showing much of interest in a structural
way. The material described is chiefly cranial, although
a few thoracic plates (interclavicles), of the typical laby-
8 Journ. Morphol., II, 1889, p. 463; 1908, p. 511.

376 THE AMERICAN NATURALIST [VoL. XLIX

rinthodont form, are described and figured. Doctor Wi-
man is to be congratulated on his contributions to our
knowledge of these early vertebrates. His future papers
will be looked for with much interest.

Po

=

oan

On

S

~]

|

_
Oo

=
ar

a
=

BIBLIOGRAPHY

Broili, F. 1913 a. Uber zwei Stegocephalenreste aus dem texanischen
Perm. Neues Jahrbuch f. Mineral., Bd. I, Jahrgang 1913, pp. 96-1
Taf. IX.

1913. Unser Wissen ia die ältesten Tetrapoden. Fortschr.
d. Naturwissenschaftl. Forschung, herausgegeben v. Prof. Dr. E. Abder-
halden-Hall. ae Bd. VIII, pp. 51-93, figs. 14-62.
room, R. 1913. a. Studies on the Permian Temnospondylous Stego-
cephalians of Ra th America. Bull. Amer. Mus. Nat. Hist., XXXII,
II, pp. 563-595, with 21 ia
On the Structure of the Mandible in the Stegocephalia.
Anat. Anz., Bd. 45, No. 2/3, pp. 73-78, with 4 figs
Case, E. C. 1912. Paleozoic apaiia and gh, Bull. Geol. Soc.
aige a 23, pp. 200-
raa 913. Neue Labyrinthodonted aus der schwibischen Trias.
Sai ran Bd. LX, pp. 275-294, pls. XVI-XXII, with text-

1-5
. Gregory, wm. K. 1913. Crossopterygian Ancestry of the Amphibia.
VII, No. 960, pp. 806-808.

Science, N.S., Vol. XX
Huene, Fr. von. 1912. Die Unterkiefer von Diplocaulus. Anat. Anz.,
Bd. “a No. È p- 472-475, Figs. 1-3.

13 a. Über Lysorophus aus dem Perm von Texas. Anat. Anz.,
Ba. a No. 14/15, pp. 389-396, Figs. 1-5, with bibliography.

. —— 1913 b. The Skull Elements of the Permian pagite in the
. Mus.

American Museum of Natural History, New York. Bull. Am
Nat. Hist., XXXII, Fork XVIII, pp. 315-386, Figs. 1-57.

. Watson, D. M. 8, The Larger Coal Measure ee Mem.

and Proc. loa ii and Philos. Soc., Vol. 57, ; Dee.
1913. Batrachiderpeton lineolatum, a Coal Measure Btegocóplia-

figures in th See als :
tion of the ae Rept. 83 Meet. British Assoc. Advance. Se. Bir-
mingham, 1913, p. 532.

1914. The Cheirotherium. Geol. Mag., Dec. VI, Vol. I, No. 603,
pp. 395-398.

. Williston, S. W. 1913. The Primitive Structure of the Mandible in

Amphibians and Reptiles. Journ. Geol., Vol. 21, No. 7, pp. 625-627, 1
See also same EE further notes: eit: Geol., Vol. XXII, No.

Wiman, Carl. 1913. ier das Hinterhaupt der Labyiinthodonten,
Bull. of the Geol. Instit. of Upsala, Vol, XII, pp. 1-7, Figs. 1
Uber die Stegocephalen aus der Trias Spitzbergens. ‘Bull,
of the Geol. Instit. of Upsala, Vol. XIII, pp. 1-30, Pls. I-IX, Figs.
1- e Villines aphy.

SHORTER ARTICLES AND DISCUSSION

THE RESEMBLANCE OF YOUNG TWINS IN
HANDWRITING

By each of 144 children 7 to 15 years old, forming 72 twin
pairs, the first name (and usually also a word or so like ‘‘years
old’? or March or grade) was written. These were pasted on
cards identified by chance numbering. Twelve men and women
of good general education, but of no special experience in identi-
fying handwritings, were shown the 72 specimens belonging to
72 first members of twin pairs and asked to match each by the
specimen of the remaining 72 which most resembled it.

There was thus one chance in 72 of a correct match by chance,
or 12 chances for all judges combined. There were in fact 6, 4,
8, 4, 6, 6, 7, 1, 3, 6, 3 and 4 correct pairings made by the twelve
judges, respectively, or 58 in all.

It would be possible by the same method to derive a scale for
unintentional resemblance in specimens of handwriting as shown
roughly below. Such a scale might indirectly be of use in the
study of questioned documents, since the resemblance of one speci-
men of an individual’s writing to another specimen by himself
may be regarded as the limiting case of the unintentional re-
semblance found amongst different individuals. A scale for re-
semblances produced intentionally would presumably form a
Series in which the resemblances would, upon analysis, be found
characteristically different from the unintentional or natural re-
semblances, The genuineness of a questioned specimen of writing
might thus be determined in part by measuring its resemblance
to the unquestioned specimen in the different elements character-
istic of the two scales. Resemblances of certain sorts might thus
be used as actual evidence of forgery, and differences of certain
sorts as evidences of genuineness, more systematically and ob-
jectively than is now the ease.

Specimens 145 and 147 have a curious interest as a result of
possessing nearly as close resemblance between two ‘‘natural’’
writings by two different persons as is likely ever to be found.
It is probable that if the individuals in question had each written

378 THE AMERICAN NATURALIST (Vou. XLIX

ts

Fic. 1. Rough Scale of Resemblances in Handwriting. 145 and 147 were

regarded as matches by 11 of the 12 judges; 19 and 65 were so regarded by 5
judges; 38 and 90, by 2 judges; 4 and 96 by none.

a hundred natural specimens of the same two words, and if judges
of the training of tellers in banks had been given the task of sepa-
rating the two hundred specimens into the hundred by individual
A and the hundred by individual B, the percentage of failures
would have been considerable. These specimens, that is, may
illustrate the possibility of successful forgery without artifice.
In general, of course, the experiment shows how very, very
rare the case of substantially perfect resemblance of two natural
signatures by different individuals will be. One case amongst 72
pairs of twins probably signifies less than one-case in a thousand
amongst the general population of as close resemblance as 145

No. 582]. SHORTER ARTICLES AND DISCUSSION 319

and 147. Twins are probably distinguishable by their hand-
writing oftener than by their physical appearance; for I am con-
fident that the bodies of at least five and probably more of these -
72 twins would have been as hard to tell apart from a minute’s
visual inspection as specimens 145 and 147. Of people in general
this would probably not hold true, but the distinguishing value
of a specimen of natural writing is very high even for them.
EDWARD L. THORNDIKE
TEACHERS COLLEGE,
COLUMBIA UNIVERSITY

ALLELOMORPHS AND MICE

In the February number of this journal, C. C. Little points out
that Cuénot (1903) recognized certain factors in mice as allelo-
morphic,* and that in my paper of 1914 I not only failed to men-
tion that Cuénot had treated these factors in this way, but that I
claimed to have brought forward for the first time a demonstra-
tion of the allelomorphism in question. In fact, I did overlook or
had forgotten that Cuénot interpreted these types in this way;
and curiously enough, my work was undertaken because Little
on the alleged results of some of his own earlier experiments
denied that the factors for yellow and gray are completely linked,
despite Cuénot’s evidence, then published, which Little now says
established from the ratios obtained that the factors in question
are allelomorphic.? Little wrote as late as 1913:

“Yellow” in mice is no more alehenirniee to gray than is gray
allelomorphie to black.

If this is the conclusion at which he arrived after his elaborate
series of experiments and after Cuénot’s work had been done,
the need of further work would seem to be obvious.

The failure of several of us to fully appreciate the significance
of Cuénot’s statements and evidence in regard to allelomorphism
may in part be due to the fact that in his second paper Cuénot had
used the symbols G (gray) and N (black) as allelomorphs, and had
besides used the symbols A (albino) and G (gray) as allelomorphs
without, however, intending to mean that there was here a set of

1 Note 1903, Archiv. Zool. Exp. et Gen. (4), I.

2 The numerical results are the same for complete linkage and for multiple
allelomorphs. The evidence that would ites the one would also disprove
the other.

380 THE AMERICAN NATURALIST [ Von. XLIX

three allelomorphs, but using G in each of the two cases to rep-
resent a different factor of the gray mouse.’ This method is in
itself quite legitimate, but as a result when Cuénot later spoke of
the factor for gray, gray white belly, yellow, and black, all as
allelomorphic, some of us, it appears, failed to appreciate that in
doing so Cuénot was treating this set of terms in an entirely differ-
ent way from the way in which he treated the other cases, where
he had represented factors as allelomorphiec to each other. In the
second place, the full significance of multiple allelomorphs in mice
was not, I think, fully appreciated until its relation to complete
linkage became apparent, and in fact even now this relation is not
sufficiently understood by many geneticists themselves. Even,
however, had I taken fully into account all that Cuénot had done,
the somewhat extensive experiments that I undertook in order to
prove that the factors in question are allelomorphie would have
seemed to me necessary, as they still do, to establish that this
series of factors bears this relation to each other. Let us ex-
amine, therefore, the evidence, which, according to Mr. Little,
rendered my experiments a work of supererogation.

1. Little says:

As early as 1903 Cuénot recognized that albinos, potentially yellows,

when crossed with black gave besides yellow offspring either black or
agouti young but not both. This is, of course, evidence that yellow,
agouti and black‘ are all allelomorphiec to one another.
But the evidence proves nothing of the sort, unless Cuénot had
shown that his albinos should have been expected from their
history to contain all three factors in question. However likely
it may seem, to one so inclined, that such triple forms must sooner
or later have been met with by chance, the fact remains that
Cuénot had, as the offspring showed, used only double types, and
such a fact in the absence of explicit evidence as to the history
of the forms used can not be said to demonstrate anything in
particular.

2. Little continues:

At the same time he gives the ratios produced by crossing an albino

3 The ‘‘examples’’ given on page vii of the second memoir are also in-

structive in the present case.
4 Probably Little means here by ‘‘black’’ what he later calls the non-
f so he refers to a different AE from that which he

agouti factor.
. is, moreover, clearly

called black, when, in 1913, he wrote:—‘ Blac
a positive character which is dominant over its pecans

No. 582] SHORTER ARTICLES AND DISCUSSION 3X1

potentially a heterozygous gray (agouti) with a yellow carrying black
but no agouti and albinism. . . . Cuénot recognized that the ratio ex-
pected from the cross was 2 valine: 1 black and 1 agouti (gray).
If one turns to Cuénot’s experiment® he finds that Cuénot
crossed an albino carrying black (AN) to an albino carrying gray
(AG) in order to obtain a ‘‘dozen’’ white mice with the formula
AGAN, and similarly he crossed a black mouse (CN) to a white
mouse carrying yellow (AJ) in order to obtain another ‘‘dozen’’
mice with the formula CNAJ. We are not told whether each
dozen came from the same parents, or from several similar com-
binations. It will be observed that the yellow (J) and black (N)
were brought together to make one F,, and that gray and black
were brought together to make the other F,, hence since gray was
in one F, and yellow in the other F, it is not possible to tell
whether they behave as allelomorphs to each other. There is no
reason, then, for making gray and yellow both allelomorphie to the
same factor (N), black; for, in the first cross the gametes (omitting
A and C) might have been Gj (gray) and gj (black) and in the
other cross gJ (yellow) and gj (black). The numerical results
would then be those obtained by Cuénot, which would prove noth-
ing in regard to the allelomorphism of gray, yellow and black.
In other words, the letter N stands in this cross simply as a sym-
bol for anything in the black mouse that could be treated as
allelomorphie to G in the one ease, and to J in the other; just as
at first when rose comb in fowls was found to give a 3 to 1 ratio
with single comb it was treated as an allelomorph to single; and
likewise when pea was found to give 3 to 1 with single it too was
regarded as allelomorphie to single. Later it was found that S
(single) stood for two factors (‘‘absences’’), small r and small p.
3. Little thinks that both Cuénot and I have fallen into the
Same error in regard to black; but he fails to see that from our
points of view in regard to the other colors it was inevitable that
we should come independently to the same conclusion. Little
says that the ‘‘true’’ series of allelomorphs is yellow, white-
bellied gray, gray-bellied gray, and non-agouti (not ‘‘black’’).
The factor that Little prefers to call non-agouti, I call the black
factor. He regards a non-factor as a member of an allelomorphie
series, while I regard the black mouse as the result of the action
of a factor for black. By the same criterion as that by which
Little speaks of a non-agouti factor, he might equally well claim
5 Third note, p. xlix.

382 THE AMERICAN NATURALIST [Von. XLIX

that the ‘‘true’’ series is black, gray gray belly, gray white belly
and non-agouti (instead of yellow).

The race of white-bellied mice that I have kept for several
years does not correspond in all respects to Cuénot’s description
of those that he has studied. His account of them in 1908? is as
follows:

La Souris reste grise sur le dos, mais le ventre prend une teinte blane
roussatre, avec un bouquet de poils plus roux entre les deux pattes de
devant, et une bordure un peu plus rousse sur les flanes; elle resemble

`

d’une façon frappante ‘4 Mus sylvaticus, L.
Again in 19117 Cuénot says:

La première diffère de la Souris grise sauvage par la teinte du ventre,
qui, au lieu d’étre gris-clair, est blane roussâtre, avec un bouquet de
poils roux entre les deux pattes de devant et une bordure un peu plus
rousse sur les flanes; cette Souris a souvent de gros yeux saillants, de
sorte qu’elle ressemble d’une façon frappante au Mulot (Mus sylvat-
icus, L

In my race of white-bellied mice there is not a bouquet of rus-
set hairs between the front legs, and I have not observed that the
eyes are large and protruding more than occurs at times in other
mice. At present, however, I have two old mice that were re-
cently found that have a tuft of faint russet hairs between the
forelegs. Whether we have‘here still another allelomorph, or
whether a particular genetic constitution makes apparent the
bouquet in conjunction with the white-bellied factor, remains to
be worked out. While it seems probable that Cuénot’s type of
white-bellied mouse and that which I have studied are the same,
it is not certain that such is the case until further work has been
done.

Cuénot has not published as yet any conclusive evidence to
show that the gray mice with white belly belong to the series of
allelomorphs, although it is true he states that this type is allelo-
morphic to the three other types. Finally, I should like to add
that I am far from wishing to appear to minimize the importance
of Cuénot’s work, and it is now evident that he should have
received full credit for his recognition of the allelomorphic
nature of the four factors in question.. I still think, nevertheless,

6 Sixth note, II, p. xv.

7 Seventh note, p. xlvii.

No.582] SHORTER ARTICLES AND DISCUSSION 383

that there was room, as matters stood, for the analysis that Stur-
tevant published and for the work that I carried out.
T. H. MORGAN

COLUMBIA UNIVERSITY

A METHOD OF CALCULATING THE PERCENTAGE OF
RECESSIVES FROM INCOMPLETE DATA

In the very interesting article on ‘‘The Inheritance of Left-
handedness’’ by Professor Ramaley in the December number of
the NATURALIST, a table is given on page 736 showing the propor-
tion of right- and left-handed children in families where both
parents are presumably heterozygous for right- and left-handed-
ness, including only families with left-handed children. In 93
such families there are 282 right-handed and 116 left-handed
children. This gives 29.13 per cent. of left-handedness in these
families. It is clear, however, that this does not represent the

TOTAL NUMBER OF CHILDREN FROM PARENTS HETEROZYGOUS FOR RIGHT- AND
LEFT-HANDEDNESS, BASED ON THE NUMBER OF FAMILIES OMITTED
BECAUSE OF ABSENCE OF LEFT-HANDED CHILDREN IN PROF
RAMALEY’S TABLE 4

Children per Actual Number Actual Number} Corrected Number | Corrected ee
Family of Families of Children of Families’ | of Childre
1 4 4 16 16
2 14 28 32 64

3 if 51 294

4 23 92 33.63 134.5

5 18 90 23.6 118.0

6 5 30 .08 36.45

7 1 7 1.154 8.

8 6 48 6.666 53.35

9 4 36 4.325 38.
12 1 12 1.032 12.39

Total 569.87

true Mendelian proportions if right-handedness is a simple Men-
delian dominant over left-handedness. For instance, in families
where both parents are heterozygous and in which there is only
one offspring, the probabilities are that only one family in four
will show the recessive character. From the total population re-

8 Archiv. de Zool. Exp. et Gen. (4), IX; (5), VIIL

1 By oversight in Professor Ramaley’s additions one column of right-
handed children was omitted, so that the numbers given in the table are
incorrect,

384 THE AMERICAN NATURALIST [ Von. XLIX

sulting from such matings, we therefore leave out three fourths
of the families when we include only those showing the recessive
character in the offspring. In families of two children nine
sixteenths of the families are omitted. In general, the number of
families omitted in such a study is 3"/4", where n is the number
of children per family. In order to get the true Mendelian pro-
portions we must take account of these omitted families. The
accompanying table shows the most probable results in Professor
Ramaley’s study had he been able to include the proper propor-
tions of families in which left-handed children might have oc-
curred.

Thus if Professor Ramaley had had at his disposal the full
number of families of this character there should have been
about 570 children in them, 116 of which were left-handed, or
20.37 per cent.

This is somewhat lower than the theoretical 25 per cent., and I
would suggest as a possible cause of this the fact that so many
children who are naturally left-handed are from early infancy
trained to be right-handed. Hence the number of left-handed
children reported is probably less than the true number of
recessives in these families.

W. J. SPILLMAN

U. S. DEPARTMENT OF AGRICULTURE

VOL. XLIX, NO. 583 JULY, 1915

THE
AMERICAN
NATURALIST

A MONTHLY JOURNAL
Devoted to the Advancement of the Biological Sciences with
Special Reference to the Factors of Evolution

CONTENTS

Page
I. The Role of the Environment in the Realization ns a Sex-linked Mendelian
aracter in Drosophila. Professor T. H. M -3

Ii. On a Criterion of Substratum Homogeneity (or ee in Field Ex-
eriments. Dr. J. ARTHUR HARRIS 430

Ill. Shorter Articles and Discussions: A Note on a Gonads of ETEEN
of Drosophila ampelophila. F.N. DUNC 455

THE SCIENCE PRESS
GANCASTER, PA. GARRISON, N. Ë.

NEW YORK: SUB-STATION 84

The American

Naturalist

intended for peepee and books, etc., intended for review should be

MSS.
sent to eye Editor of THE A
S

ERICAN NATURALIST, Garrison-on- Hudson, N
cles containing summaries of

ork.
research work bearing on “the

problems of Freie evolution are especially welcome, and will be given preference

in publication

ndrea Ay es of = pe bos pet are supplied to authors free of charge,

One hu
F chee reprints will be

ions ane gd santei should be sent to the publishers.
ear. Foreign postage is
Canadian postage twenty-five cents additional.

jubéorigtion price is four dollars a

forty cents.

The
cents and
The —— for single copies is

The advertising rates are Four Dollars for

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Entered as second-class matter, are 2, 1908, at the Post Office at Lancaster, Pa., under the Act of
ngress of March 8, 1879.

Lancaster, Pa.

Garrison, N. Y.

FOR SALE
ARCTIC, ICELAND and GREENLAND
BIRDS’ SKINS
Well Cig’ aes Prices
ula
G. DINESEN. bial per
Husavik, North Iceland, Via Leidle, Enana

JAPAN NATURAL HISTORY SPECIMENS
Perfect Condition and Lowest Prices.
Specialty: Bird Skins, Oology, Entomology, Marine
Animals and others. Catalogue free. Correspond-

ence solicited.
T. FUKAI, emia et
osu, Saitama, Japan

BIRDS OF AMERICA by J. J. Audubon, 7 volumes,
please report cash price, stating condition, binding and dates
of volumes.

F. C. HARRIS,
Box 2244 Boston, Massachusetts

The ot of Chicago |

Arts, Lite stare,
» Li vie ure,
Commerce Administra-

ss HS ong Sly =

Mitchell Tower

Marine Biological Laboratory
Woods Hole, Mass.
RSTO Teas poe in eee.

le beginners in
Sete <6 work comer the nengen
$50.00.
INSTRUCTION

July—August
Physiology

logical and B

aen on g pea Stato, whioh

. pasa. or pi rot
ienen

THE
AMERICAN NATURALIST

Vou. XLIX. July, 1915 No. 583

THE ROLE OF THE ENVIRONMENT IN THE
REALIZATION OF A SEX-LINKED ME
DELIAN CHARACTER IN DROSOPHILA

Proressor T. H. MORGAN

COLUMBIA UNIVERSITY

CONTENTS
i The Influence of the Environm
The Linkage of the Factor for Abnormal ar ole Sex-Linked Factors.
(a) The Linkage of Abnormal and White.
lack, Red, Abnormal by Gray, White, Normal.
Gray, Red, Abnormal by Black, White, Normal.
(6) The Linkage of Abnormal, Yellow, and White.
ray, Red, Abnormal by Yellow, White, Normal.
Gray, White, Abnormal by Yellow, White, Normal.
a White, Abnormal by Gray, Red, Normal.
low, S Abnormal by Gray, White, Normal.
2. Change ay aa as the Culture Grows Older.
ests of Changed over Classes
. Influence of the Factor for Black on the Realization of the Abnormal

P 2

aracter.
- Influence of the Factor for Yellow on the Realization of the Abnormal
haracter.
- The Relative Influence of the Egg and of the Sperm on the Condition
of the Heterozygote.
7. Presence and Absence.
8. Other Types of Abnormal Abdomen.
9. The Non-Inheritance of an Acquired Character.
10. The Non-Contamination of Genes.

Oo

a

INTRODUCTION
THE mutant, from which the stock with ‘‘abnormal ab-
domen” was derived, appeared in 1910. It is charac-
terized by a peculiar condition of the pigment bands and
segments of the abdomen as shown in Fig. 1. The range
of variation of the character is very great; in its most
385

386 THE AMERICAN NATURALIST [Vou. XLIX

Fig. 1.

extreme condition not only do the pigment bands totally
disappear, but even the lines between the metameres are
broken up, and the location of the external genitalia may
be shifted to a more terminal position. All stages exist
between this extreme modification and a condition that
can not be distinguished from the normal. Owing to
this wide range of variability the study of the inher-
itance was very difficult until it was found that the reali-
zation of the type is a function of the environment.

the more extreme types the abdomen is deformed to
such an extent that copulation is difficult or impossible.
The sterility caused in this way helped also to make the
work burdensome, especially when breeding was made
with pairs. Instead of pairs, cultures of ten to twenty
individuals of the more extreme type were resorted to,
as a rule insuring the successful mating of some indi-
viduals. Aside from this mechanical difficulty in mating,
the mutant race is quite vigorous and of good size.

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 387

Two principal obstacles delayed the formation of a
pure strain. The new character is a sex-linked domi-
nant,’ but both the heterozygous and the homozygous
condition overlap the normal type which makes the selec-
tion of pure females difficult. Any male, however, that
shows abnormal abdomen at all is pure, for the charac-
ter is borne by the X chromosomes of which he has but
one.

The other obstacle was what at first appeared to be a
perpetual reversion of stock, seemingly pure, to the nor-
mal. So constantly did this occur, that, for some time, I
thought that I had an ‘‘ever-sporting’’ variety—one
that reverted to the normal without apparent provoca-
tion. I found, however, that the first flies that hatched
in the best-fed cultures were entirely abnormal, while
those that emerged later were less abnormal, until finally
those that emerged when the cultures were nearly at an
end were invariably normal flies. It seemed at first pos-
sible that such stock might be impure, and that the ab-
normal flies hatched sooner than the normal, but this view
was negatived by the fact that normals hatch as soon as
do the abnormal flies.

The one remaining possibility seemed to be that de-
velopment of the abnormal abdomen depended on some
definite condition of the culture—one that was present
when the food was fresh and the bottle wet, but which
disappeared as the food was used up and the bottle got
dry. I tested this hypothesis in many ways. Stock was
used that had been pure for nine generations. As a
bottle dried up an ever increasing proportion of normal
flies appeared. At intervals lots of flies were taken out
and put into new bottles where they were abundantly fed.
Their first progeny, as recorded below, shows that under
the new conditions the offspring were sometimes ex-
tremely abnormal irrespective of the general condition
of the original stock when used.

1 Morgan, T. H., ‘‘A Dominant Sex-Linked Character,’’ Proceed. Soe.
Ezp. Biol. and Medicine, IX, October 18, 1911.

388 THE AMERICAN NATURALIST [Vou. XLIX

Condition of Parents Next Generation

Feb. 26. Most flies abnormal—a few were normal.. Very abnormal.

Feb. 27. More than half were normal........... Flies fairly normal.

Feb. 28. About half were normal............:.. Nearly all abnormal.
Feb. 29. Practically normal <2... 0.06. eee tse os Very abnormal

Mohd. Nanri Orai o is i. 8 se oo 88 Abnormal, a few normal.
Mok A DORA ss fale ee lig a eee ee ee Very abnormal.

Men. De OPI ep eck de bse ece ee Very abnormal.

The preceding case shows that there is no necessary
relation between the development of the abnormality in
the parent and that in the offspring. This is only a
sample of a large amount of similar data. But this evi-
dence does not show what special conditions make for ab-
normality. In order to study this problem I generally
used heterozygous females which were obtained either
by mating an abnormal male to a wild (virgin) female
(in which case the daughters will be abnormal under suit-
able conditions and the sons normal), or reciprocally by
mating a normal male to an abnormal female (when all
the daughters will be abnormal (heterozygous) and all
the sons pure abnormal). Many experiments had shown
that the heterozygous female changes over more promptly
to the normal character than does the homozygous male
and the latter sooner than the homozygous female.

The one outstanding fact for some time was that as a
bottle crowded with flies gets old there is always a change
from day to day from abnormal towards normal, but it
remained to be shown whether the change was due to
the drying out of the culture, or to any one of a dozen
other parallel changes that obviously are going on at the
same time. The more significant results of a prolonged
set of experiments may be summed up as follows:

_ 1. Starvation.—Lack of food does not bring about the
change from abnormal to normal. Flies that are so
starved as to be extremely small may be very abnormal.
: 2. Acid, Alkali or Neutral Condition of Food Stuff. —
Most cultures change in the course of the ten to twelve
days from an acid through a neutral to an alkaline con-
dition. Fresh fermenting banana (in the old and acid

No. 583] . ROLE OF ENVIRONMENT IN DROSOPHILA 389

medium) was made more acid (and liquid) by adding an
equal amount of a 5 per cent. solution of acetic acid.
Other food was made alkaline by adding dry sodium
bicarbonate, or a 1 per cent. solution of sodium hy-
droxide. The acid food gave very abnormal flies; the
alkaline food was difficult to control as the flies refused
in most cases to lay eggs on it, if it remained alkaline,
and the food often dried up, or putrified, or grew mouldy.
Moreover, the highly alkaline food often became acid
over night owing to fermentation changes taking place
within the pieces of fruit used for food. But several
times good results were obtained with cultures that had
been strictly neutral and often alkaline throughout the
time of the experiment and from these the flies were ab-
normal, Omitting all details it may be stated that an
acid or alkaline (neutral) condition as such is not the
cause that conditions the character.

3. Food of Parents.—At one time it seemed possible
that the kind of food that the female was supplied with
might for a time continue to affect her eggs, even al-
though the parent was transferred to a medium that
acted in the opposite direction. Careful tests showed
conclusively that such was not the case. Some of the
evidence for this statement will be given later.

4. Egg versus Sperm.—Heterozygous females may be
produced either by using a normal female and abnormal
male, or conversely an abnormal female and a normal
male. Certain cultures seemed, at one time, to show that
when the egg parent was abnormal the offspring were
more abnormal than when the egg parent was normal,
but careful tests disproved this view. The difference in
the cultures, that led to the suspicion mentioned, was due
to the large number of eggs laid by the normal females,
hence greater crowding and more rapid disappearance
of the moist food.

5. Influence of Genetic Factors.—Certain mutant stocks,
notably black, seemed at times to show the abnormality
less strongly than other stocks, but here, as in the last

390 THE AMERICAN NATURALIST [Vor. XLIX

ease, the results were found to be due, when carefully
tested, to the number of eggs laid and the promptitude
with which they are laid when the food is fresh. The
question will again come up in certain of the crosses.

6. Amount of Water in Food.—Normal cultures lose
much of their water as the brood of flies develops. It
was a fact noticed at the start that in ‘‘wet’’ bottles the
abnormal characters appeared to best advantage, and in
most of the work on linkage that knowledge was utilized.
But whether the wetness was only incidental to other
changes, or in itself the normal condition was not pre-
cisely determined. Under all conditions the air in the
bottles must be completely saturated with moisture so
that we must be dealing with the water taken in with the
food and not with the amount of water in the inspired
air. In three ways the effect of water was studied. (1)
Food that had been fermenting for two or three days in
the old acid medium was squeezed until freed of much
of its water. The solid part was then further dried su-
perficially by pressing between pieces of filter paper, and
finally put into a bottle with more dry filter paper. The
fluid squeezed out was diluted with an equal amount of
water, and put into another bottle. Virgin normal flies
and abnormal males of pure stock were set free in these
two bottles. The results at the end of nine days were
most striking. In the dry bottles the F, females were
all normal; in the wet bottles the F, females were ex-
tremely abnormal.

7. Changing the Adult from Wet to Dry and Vice
Versa.—In this same series the old (P,) flies that had
been in the wet bottle were transferred to dry food, and
conversely the ‘‘dry’’ flies to wet food. Their progeny
showed the influence of the food that they were reared
upon, and no effect of the feeding in the previous bottle.
Once more they were changed, the wet to dry, the dry to
wet, and the results were the same as before, i. e., the
actual conditions, not the preceding ones, ba acnounted
for the results that were obtained.

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 391

8. A culture that was giving F, normal females (that
were heterozygous for abnormal) was made extremely
wet; into a sort of swamp. The flies that emerged dur-
ing the next six days were normal, on the seventh day
the flies were slightly to fairly abnormal, on the eighth
and ninth days the flies that emerged were slightly to
quite abnormal. It is evident that the influence of the
wet conditions does not appear unless the flies are sub-
jected to it throughout most of the larval life, or else that
the first few days of larval life is the critical period.

9. Larve that were about ready to pupate were trans-
ferred to very wet new food, where they pupated, in
from 12 to 24 hours. The pupe remained in the same
bottles until the flies emerged. These flies were entirely
normal in appearance; the stock from which the larve
came were also giving rise to normal flies. The sojourn
of one or two or even three days in a wet environment
at the end of the larval life does not suffice to alter the
effects that have already been induced in an earlier stage.

Conclusions.—The preceding evidence makes clear
that the amount of water in the food, determines the
realization of the ‘‘abnormal’’ type. The water may
produce its effect either by being taken in with the food,
or by being directly absorbed; or it may determine the
nature of the bacterial or yeast flora that in turn deter-
mines the nature of the fermentative changes that take
place within or without the larve. It would be a very
difficult matter to find out in which one of these ways the
effects are brought about. However this may be, it is
possible for the experimenter to determine at will the
nature of the flies that will be produced in his cultures
by controlling the food supply.

THE LINKAGE OF THE FACTOR FOR ABNORMAL ABDOMEN
WITH OTHER SEX-LINKED FACTORS
Owing to the overlapping of the abnormal and normal
types, the study of the linkage has presented unusual
difficulties. The following experiments were made for
the most part during the second year when the influence

392 THE AMERICAN NATURALIST ~— [Vou. XLIX

of the environment was not fully under control. The
conditions under which the experiments were made were,
however, favorable for the appearance of the abnormal
condition, at least in the first counts of each brood, for
the bottles were supplied with an abundance of wet fer-
mented food.

The linkage of abnormal abdomen with white eyes and
yellow body color was studied in different combinations;
and since the factor for abnormal abdomen proved to be
quite near the other two factors the choice was a favor-
able one in certain respects. A special method by means
of which the error, due to the variability of the charac-
ter, can be largely eliminated will be given after the evi-
dence has been presented.

THE LINKAGE OF ABNORMAL AND WHITE
When red-eyed (R) abnormal (Ab) females were mated
to white-eyed (W) normal (N) males, red abnormal
males and females were produced.? When these were
mated the results recorded in the next table were ob-
tained
By means of the following diagram, I have tried to

WE PE

wW N

DIAGRAM I.

show what the expectation is for this combination. The
two parallel lines are intended to represent the two sex
chromosomes of the F, female. From her mother she

2 Throughout this paper I have used the letters R for red eyes, W for
white-eyes, N for normal abdomen, Ab for abnormal abdomen, Y for yellow,
B for black instead of using the allelomorphic system; because for present
purposes, where analyses are unnecessary, these letters suffice most simply
to indicate the operations that are involved. For comparison with other
papers the allelomorphie symbols for the same neti would be:

w= the factor for white. W =its normal allelomorph = red.
A'b = the factor for abn. abd. a’b—Zits normal oo = normal abd.

y= the factor for yellow. Y=~its normal allelomorph = gray.

b= the factor for black, B=its normal enara gray.

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 393

got the sex chromosome bearing the factors for red and ab-
normal (RAb), from her father the homologous sex
chromosome that carries the factors for white and nor-
mal (WN).
TABLE I
ParRENTS: RAD? BY WN g
F,: RAb9—RAb ¢

RAb WN | RN WAb |
seas | POV EREE] TEESE E RR gli EETA RIT EAR No. of Culture
A9 F | toe | 9 J | g
Fs deat ae.” ie or ee Ha 2 ee tea Ih
148) ee | ee AS | 6 1 eh eee Ils
106 | 3 is ee or. 3 ea, Ils
ee Gece R - ae peed vee MS sss aa a Il;s
B | Ob A081 EP E e Iza
a a eee a eee eer 4 Tiss
GO 6 BOs ieee 1 “hee pet Ti»
647 i 17 el o. kæde han n

If these chromosomes unite at synapsis without ex-
change of materials, half of the eggs that result (one
chromosome being eliminated in the polar bodies) will
contain the red normal combination, the other half the
white normal. These represent the ‘‘non-cross-over’’
gametes. If, however, these chromosomes should cross
and reunite, as in the diagram (the crossed lines indi-
cate where the crossing over may occur, not how it oc-
curs), the two resulting chromosomes will be red-normal
RN, and white abnormal, WAb, which represent the
other (the cross-over) kinds of gametes of the F, female.
The ratio in which they are produced is the gametic
ratio and is a measure of the linkage.

In the F, males there is but one X chromosome, hence
there is no opportunity for interchange here between
the X chromosomes. The mate of the X chromosome is,
in the male, the Y chromosome. Other experiments have
shown that the Y chromosome carries no factors; hence
interchange seems precluded; and, so far, no loss of X
chromosome factors to the Y chromosome has ever been
observed. The X chromosome passes into the female-

3 An unexpected individual that can be accounted for by equational non-
disjunction.

394 THE AMERICAN NATURALIST [Vou. XLIX

producing spermatozoon, which carries therefore an X
chromosome received from the mother of the F, males
and bears her character. In the present case the male
carries the chromosome bearing red abnormal.

Since red and abnormal dominate, all the F, females
should be red abnormal, except in so far as the conditions
suppress the abnormal and induce the normal type. The
experiment, Table I, shows that very few normal females
were present.

Four classes of males are expected—the large class of
non-cross-overs RAb and WN, and two small classes of
cross-overs RN and WAb. It will be observed (Table I)
that the linkage between R and Ab is very strong, since
nearly all of the males are either RAb (647) or NW (664).
Only a few crossovers RN (25) and WAb (13) males
were present. The percentage of crossing over is 1.97
per cent. when the abnormal males alone are used for
calculation.

In the reciprocal cross the RAb male was mated to
WN female, and gave in F, RAb females and WN males.
The F, record is given in Table IT. —

TABLE II
PARENTS: RAb g By WN?
F,: RAb9—WN ¢

RAb WN RN WAb

By 9 J 9 a g Ty

68 62 59 47 1 0 2 1 Il:

38 69 68 0 3 0 1 II20
115 170 130 147 2 2 2 2 IIa

03 103 97 103 3 7 4 2 II

75 96 94 49 2 1 5 0 Tis
399 486 449 194 8 13 i ro Gt

Since the same two pairs of factors enter as before,
the same chromosome diagram will suffice for the gametes
of the F, female. The F, male is, however, a double re-
cessive (WN); in consequence four classes of females
are expected as well as of males. The gametes of the
F, female are as before the following:

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 395

Non-cross-over RAb Crossover RN
gametes WN gametes WAb

The percentage of crossing over as calculated from
the abnormal classes (males and females) is 2.1.

In order to obtain further data for linkage the pre-
ceding experiment was repeated in the winter of 1914,
but the linked factors entered differently combined. The
experiment was begun by crossing white abnormal fe-
males to wild males which gave red abnormal females
and white abnormal males. These were inbred and gave
the following results in five different cultures (kept with
abundance of moist food).

TABLE III

PaRENTS: WAb? By RNG
F,: RAb 9—WAb g

RAb | WN RN | WAb
No. | RESCUERS US ARLENE reas
a | 9 Pee ce fhe ee ee ee
1 0 44 0 2 33 0 52 55
2 4 66 0 0 Sl s 57 50
3 1 38 0 0 62 0 37 48
4 1 lil 7 1 95 0 77 100
5 1 65 2 0 65 0 | 34
Total | 7 | 324 oe 312 4 | 251 | 287

The sum of the two non-cross-over males (251 + 312 =
563) plus the cross-overs (16) divided into the sum of
the cross-over males (7 +9—16) gives 2.7 as the percent-
age of crossing over. Since the white normal males
may receive contributions from the changed white ab-
normal, the result may be freer from error if the two
correlative abnormal male classes, viz., red abnormal (7)
and white abnormal (251), be utilized to calculate cross-
ing over. Dividing the former by the total (251 +7) —
gives 2.7 per cent. of crossing over which is the same as
the preceding estimate.

The reciprocal cross, RN 2 by WAb ¢, was also made
once and the results in F, combined with other similar
results are as follows:

WAb RAb WN

a ọ g g g g
1,220 (withRAb) 854 20 1,323 89

396 THE AMERICAN NATURALIST — [Vou. XLIX

The other results were obtained in the following way:
The abnormal red eyed F, females obtained from the
first experiment are heterozygous for abnormal (AbN)
and white (RW), except in so far as this class may
contain cross-over flies that are heterozygous in white but
homozygous in abnormal AbWAbR. Except for these
flies, these F, females are like the F, females, and if
mated to abnormal white males will continue in each
successive generation to give the same linkage data as
do the F, classes above. If bred in pairs exceptional
females homozygous for abnormal will be at once de-
tected, and can be thrown out; but even if bred in small
batches of four or five females the chance is small of in-
cluding homozygous abnormal females.

In these counts no separation of the normal red fe-
males (when they occurred) from the abnormal red females
was made but the red females were put into the latter
class. Since the females were not intended to be used
for comparison this grouping does not affect the prob-
lem involved. If we divide the cross-over red abnormal
males (20) by the abnormal white males (854) plus 20
abnormal red males, we get the per cent. of cross-overs
which is here 2.3. This is slightly lower than that ob-
tained for the preceding data.

Black, Red, Abnormal by Gray, White, Normal

Another series of experiments, carried on for a some-
what different purpose, may be utilized here for further
data. Gray, white, normal females were mated to black,
red, abnormal males. The daughters were gray, red,
normal (or slightly abnormal), and the sons gray, white,
normal. Inbred they give the results shown in Table
IV. Since the factor for black is not sex-linked, the
gray and the black classes may be added together as
shown in Table V.

The results differ from those of Table II in the follow-
ing points: There are relatively more red normals which
may be assumed to be due to the external condition pre-

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 397
TABLE IV
PARENTS: BRAb g By GWN 9?
F,: GRN or SLIGHTLY Ab9—GWN g
GRAb | BRAb | GWN | BWN GRN | BRN ama
ll SP Sat aes

PPL A] Ph Pe ta pigaa e|elelale|

35| 63) 7| 8| 160| 170] 48| 58|126 |146| 54| 43|...]...| IIs
50| 78} 6| 16| 111| 104| 34| 18| 48| 30| 34| 28 efes III29
65 | QO} 1] 221) 187| 13| 7/181}159| 11] 16; 1: Hi
187 |152| 27| 36| 254) 268| 58| 72|,102| 99| 45| 65 4/1) UR
33| 56| 12| 12| 37| 63| 15| 10 0] 0] o|...|...} Ie
103 |140| 34| 34| 146| 191) 52) 42| 54| 52| 21| 14| 6| 1| IIIs
50| 33| 14| 10] 93| 69| 22| 24| 30| 24| 7| 6j 1|...| Ils
8} 71| 0 2) 76 123) 20 40| 80| 66| 27| 61]...|...| IIs
91| 74| 17|- 16| 89| 89|-30| 45| 75 | 66| 79| 59| 1|...| Ie
92| 88|:27| 24| 135| 153| 33| 59| 62| 76| 19| 20|...| i | Tse
2 8 3 7] | 4 0) 0; 0)...)...) Ex.
_82| 95| 32| 23| 195) 191| 47| 65| 66 66 | 94| 34 29}... 1 | ihe

| | |

798 899 182 |185 1,524 1,625 375. PEE

vailing at the time, or else the black factor may have had
some influence that is favorable to the normal condition
in the heterozygous abnormal flies. If the latter were
the true explanation we can understand the large num-
ber (here) of the GRN class (for two thirds are hetero- -
zygous in black and intermediate in color) and the com-
plete absence of the BW Ab class which should be one third
as frequent as the GWAb class.
was made to test this possibility and will be described

TABLE V

PaRENTS: BRAb BY

A special examination

GWN?

F,: GRN or sLIGHTLY Ab9—GWN ¢

RAb WN RN WAb |
F Q F ọ F Q oo]
42 60 208 237 179 1960 E (oT
56 94 145 122 82 PE ean Saliva! Iie
65 53 234 194 142 175 ii Dhi
214 188 312 340 148 64 4 1 | Ile
45 68 52 } GE rr i ae II Tio
137 192 198 233 75 66 6 i | die
64 43 115 93 37 30 Looe Wa
8 73 96 163 107 iT peo | Ils
108 90 119 134 154 125 io) Die
119 112 168 212 81 E Í IIIso
8 11 10 cy Oe RE E © POC cee A Cana, Extra
105 118 242 256 100 cj to B
S41 | 1,102 {| 1,809 | 2065 { 1,105 | 1,168 | 13 | & |

398 THE AMERICAN NATURALIST [Von XLIX

later, but it may be stated beforehand that no certain
evidence could be found in favor of this view. The num-
ber of larvæ in a culture brings about a rapid alteration
in the condition of the food, so that it changes more
quicky from an acid to a neutral or alkaline condition.
If the black flies used gave vigorous F, offspring the
effect in question could be explained as due to numbers,
and not as connected with the black factor.

Gray, Red, Abnormal by Black, White, Normal

The results of this cross and of its reciprocal are given
in Tables VI and VII. The RN class (cross-over) is
relatively too large, but the increase is due to the transi-
tion from abnormal to normal.

TABLE. VI
PaRENTS: GRAb@? By BWN GC
F,: GRAb 9—GRAb J

Grab | BWN | GwN | BRAb | GRN | BWAb| GWAb | BRN _
glelalelalelalelalelalele|e lal 2
95 |194| 49 | 2 |163| 1 | 20| 32 | 56] 71 |...1...| 4 |..... | 24 | 45

PARENTS: BWN 9 By GRAb¢
F,: GRAb d—GWN 9°

215 |143| 91 | 91 |314 |276| 56 | 11 | 28 [azol...[...)9 | 2] a | 50 _

THE LINKAGE oF ABNORMAL, YELLOW, AND WHITE
, 2

In the following crosses three pairs of sex-linked fac-
tors characters are involved, viz., yellow, white, abnor-
mal and their normal allelomorphs whose location at one
end of the X chromosome is shown in Diagram II.

DIAGRAM II.

Gray, Red, Abnormal by Yellow, White, Normal

When a YWN @ is crossed to a GRAD ĝ the daughters are
GRAb and the sons YWN. The F, male is a triple re-

No.583] ROLE OF ENVIRONMENT IN DROSOPHILA 399

cessive, hence, neither his female-producing nor his male-
producing sperm affect the dominant characters that the
eggs carry, and in consequence the entire F, count, fe-
males as well as males, are indicators of the gametic
composition of the eggs of the F, female. The F, results
are given in Table VII.

TABLE VII
Parents: GRAb g By YWN 2?
F,: GRAb 9—YWN ¢}

YWN GRAb |YRAb| GWN | YWAb| GEN | YRN | GWAb |

gleleleleleleleleleleleleleiele!|
1; 45| 59| 44| 51 | Be ee ae a ee Se ae Iı
21 831 OO) a BB. T W iss diesters Is
3| 115| 107| 132) 150 1} 2|-..]...1 @| 6] 4} T I;
4| 174| 177| 152| 205 2| 5| 2] 4...) 6 9...) 1)...)... Iio
5| 100| 128| 108| 150 TiS Sra a o oa In
G| 42| -52| 58| 4 SR Ge Bee ie cee ne ee 3
7| 92| 94| 105| 114 WI r A r. 6
8| 105) 180| 123| 97 B41 die AB oe es) Basle s
9| 83 78| 92 eit Sore a> S ak bone 9
10; 83) 81) 90| 107 Lt Bt A e 10
11| 375| 374| 441| 443| 1| 6| 4| 3| 5| 7| 22| 14)...)...|...)... u
12| 103| 127] 125| 136| 1|...|...|...| 2]...] 7 81|- 12
13) 135) 116| 119| 169/ 1] 1}...) T T 9) Bo 13
14/ W01; 92] 116] 105) 1| Irop e 4} WA a 15
15| 29) 56 Pic th wie Sa 18
16; 45; 58, 50| 77 $d AB linet ales sie o 19
ee Se me a | ae aD pee Ge ees A | as A A a a
t8; 31) 45; 30 1 Pia ifi cae 1 | Iss
19| 236; 231| 283| 276| 6| 2| 4| 3| 9} 1) 6 6...) 1)...|... 27
20, 47) 31 eres ee ren ee eo 30
21; 66; 79; 101| 64/...|...) 2] Apin 17) 52)...).-.)--.)--- a
22| 325; 307| 209| 250| 1| 1| 3| 2| 1)|.../139/251) 2 | 1 |. | IVs
23| 286| 184| 233| 321| 6| 2| 5| 3| 2| 2j197/191| 1 | 1 |...| 1 |IVs

2,734 2,744 2,773 3,008 | 22 | 45 | 30 | 33 | 43 | 21 |478/644, 4 | 8/0 2
a | Ba khob a {oOo

The relation of the classes to each other is evident from
the following diagram (III) which represents (as before)

G Ab
EMD EE nas

DracRaM III.

the sex chromosomes of the F, female. The classes of
gametes of the F, females are the following:

400 THE AMERICAN NATURALIST [Von XLIX

Non-Crossovers Single Crossovers Double C
YWN YRAb YRN
RGAb GWN GWAb

YWN
GRN

In this and in the following tables the order of the cross-
over gametes is always given the same, viz.: the first
factor to the left above (Y) joins the two following below,
R and Ab, (taking the switch as it were at the first cross-
over). Then follows the cross-over that is the converse
of the preceding (the first factor to the left below switch-
ing over to join W and N). The second crossing is taken
in the same way, thus Y and W switch over to Ab, and
conversely G and R switch over to N. The double cross-
over takes the switch twice; thus Y to R and then to N;
and conversely G to W and then to Ab. The F, flies
should correspond to these gametic classes (since the F,
male was a triple recessive) except in so far as the ab-
normal classes change to phenotypic normal types. Thus
the non-cross-over class GRAb will, in this sense, con-
tribute to the single cross-over class GRN; and the single
cross-over class YRAb to the double cross-over class
YRN. The last-named class can not, therefore, be used as
a measure of the double crossing over, since it is more
probable that any flies of this kind that appear will be
only phenotypic YRN, than that they should belong to
the YRN class genetically. Only the GWAb class may
be used as a measure of double crossing over, and, as will
be shown below, much caution must be used even in this
case.

It will be seen in the table that only relatively few of
the GRAb type have changed to the normal type, because
the conditions were favorable for abnormal although the
cultures ran in most cases for ten days, but during this
~ time they still contained plenty of wet food. It will be
noticed that the changed class GRN corresponds to one
of the single cross-over classes, consequently GRN is a
mixed class, and can not be used to base any calculation
on. It is true, one may roughly determine how many

No.583] ROLE OF ENVIRONMENT IN DROSOPHILA 401

cross-overs are expected in this mixed class by compari-
son with the other single cross-over class (YWAb).
these are subtracted, the remainder shows how many of
this GRN class are due to a change from the abnormal to
normal. Another point to note is that one of the double
cross-over classes, viz., YRN, is likewise subject to addi-
tion from the single cross-over class, YRAb, and can not
itself be taken as a measure of double crossing over, while,
on the contrary, all cases in the other double cross-over
class, viz., GWAb, count for their full value. Only two
such double cross-overs occurred.

On the basis of the amount of single crossing over it is
possible to calculate, as Sturtevant has shown, the ex-
pected number of double cross-overs. The number of
the double cross-overs (two) in Table VIII is larger than
expected. I repeated (December, 1913) the last experi-
ment to test the question because abnormal arrangement
of the rings of the abdomen is not a very rare occurrence
and may sometimes be the result of injury to the larva or
to the pups, or in still other cases may be due to other
mutations, some of which will be described later. The
abnormal mutation itself occurs not infrequently under
conditions precluding contamination. In repeating the
experiment extreme care was taken not to classify any fly

TABLE VIIT
ParENtTs: GRAb g BY YWNQ
F,: GRAb 9—YWN ¢

| YWN GRAb YRAb GWN YWAb GRN YRN | GWAb
lelelelelalelale|elele|elielelale

A 58| 64| 57 6L O06 1067 8 1-oy i

D 60) O71 461 O81 oi 45 1 OT OTe rf i

E 79) l 601 7601-86 OB 1) 8 8

F O7| 921 SB) 7111] 91119] eo 3f 1] $

H Sti 40) 47) ilr oroo 0r eT a

J O71 SO) Set BOT it 72 te Or 1i eS

K 65 Bist OF11o} 21 OF: 1] 2

L 83 901116 [126° 9 6-04 1 Oa Liù 4i 3

M 74 765101 9100] 3| $| 6] 3

N 40; 51| 3881 6710] 01010] 0} 1} 3) 8

Totals .| 605 |675 |625 |683| 9 | 10| 5 |5 | 16 | 12 | 24/| 24,/0/0/0) 0

402 THE AMERICAN NATURALIST [ Vou. XLIX

in the double cross-over class as abnormal unless there
could be no reasonable doubt as to the nature of the char-
acter. In case of doubt the flies were tested by crossing
again.

As before, yellow white normal (abdomen) females were
crossed to gray red abnormal males. These gave in F,
YWN ¢ and GRAb 2 which inbred gave the results shown
in Table VIII.

The double cross-over class is GWAb. The combina-
tion did not appear once amongst the 2,690 flies that are
recorded in F,. The percentage of crossing over between
Y and W is 1.0; that between W and Ab was 2.1. The
expectation of double crossing over on this basis (without
interference) would be .02 per cent., or about 1 in 5,000.
But the expectation would be far smaller than this be-
cause of a principle that we call interference. We mean
by this term that should a cross-over occur at one point
the chance of another occurring near it is greatly dimin-
ished, because if crossing over is due to twists of the
chromosome the length of a twist would usually preclude
the occurrence of two cross-overs near one another. In
other words, if the loop that makes the twist is more likely
to be of a certain length then the likelihood of the occur-
rence of a small loop necessary for a double cross-over is
very small. In two cases, B and C, the F, counts (from
pairs of F, flies) gave no YWN males as shown in the
next counts.

| -YWN GRAb YRAb GWN YWAb sires
le@[efele|e| Ps ee
pei

a o| | | at |i pee fed due

49 | 47 | 47 | 0 t2 be 1S.
The absence of the YWN males, when the other classes
showed that no error in the experiment had been made,
was not understood until the occurrence of lethal factors
was worked out. Here clearly a lethal factor in the YWN
grandmother has been carried over into her GRAD
daughter. The lethal factor must have been closely

eo;
oo] +70
oe

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 403

linked with yellow and with white. The F, YWN son of
the original YWN female must have come from the other
sex chromosome of the YWN female—the one that did
not carry the lethal. The count of the males in the F,
gives both in B and in C a 2:1 ratio which is the charac-
teristic ratio for a sex-linked lethal. The reciprocal cross
was also made, but only twice; the F, counts are given
in Table IX.
TABLE IX
PARENTS: GRAb? By YWN g

F,: GRN@ (or suicHtLy Ab)—GRAb ¢

YWN | GRAb | YRAb | awn | ywab | GRN YRN | GWAb
gigialelalelale|alel a] jalelale
141 |---| 165 | 339 | we CA eee k 59 [10 AR
202 Joo: O86 1:86) [8s |. e be Fao Ie ME Se es Boe

The expected gametes of the F, female are the same, of
course, as before, but the male contains all three sex-
linked dominant factors, GRAb. Consequently in F, half
of the GRAb female class is pure and half is heterozygous
for abnormality. The GRAb F, males, on the other hand,
are all pure, in the sense that they have only one factor
for abnormal and no factor for normal. It is probable
that most of them here are phenotypically abnormal.

The relation of the non-cross-over and the cross-over
gametes is the same as in the reciprocal cross, since only
sex-linked factors are involved, but the cross-over classes -
given in Table IX are different in the female classes in so
far as the female producing sperms, that carry GRAb,
contain three dominants. In one of the two counts given
in the table the cross-over class that has changed to
phenotypic normal is relatively large; in the other count
it is small.

Gray, White, Abnormal by Yellow, Red, Normal
The next largest series of experiments involves the same
three pairs of characters but combined in a different way.
The results are shown in Table X. Diagram, IV shows
the relative positions of the factors in this combination.

404 THE AMERICAN NATURALIST — [Vou XLIX

Y R N

Pe (eS

G:. W Ab

DIAGRAM IV.

The gametes produced by the F, female are the following.

Non-Crossover Single Crossover Double Crossover
Gametes Gametes Gametes
R YWA WN
GWAb GRN GRAb
YRAb
GWN

The classes of special interest are non-cross-over GW Ab
males which change as the culture gets old into GWN
(which is a single cross-over class), and GRN which is
the corresponding female class (but heterozygous).

TABLE X
PARENTS: GWAb ¢ By YRNQ
F,: GRAb? (to N)—YRN ¢

YRN GWAb YWADb GRN YRAbDS; GWN YWN GRAb

Cre CPI eit Se Ss Se eee reer ee
Yee” 7g Dg ei thy ete 2145/1 Bh ea Oe ae
410) 404) 3201.. Tt. J101 185i Pie et Tr. Aa] 90i
G2) OTE WWE er e e Eeo e l 124 | IIIa
327) 123) 50 St p 1 56 T. 130 | IIIs
196| 181 1i eae) 8 Saol 239 | Iles
234| 286! 168 Te 166. bk 69 Ad 151 | IIs
108! 149 1k ae. 49 in er 79| III;
321| 381| 218 Lit St 90i 3) 208) ie 259 | IIIsz
158; 167 Pt Oe ee eee ie 96| III
178| 169 1-81 m6 3i 10 Pe 62) IIL
185| 181| 107 ele Ciera: igs ee 119 III
109 H ai oor a4 I 28 t 114 | IIIs
189} 185| 141 ie Tt 4a 27 De 172 | IIIs7
332| 322) 233 SiO) mi3 i 27l. a 309 | IITs
279| 19 .| 6| 324]. 1s FOS Fees: 14| IIe
147| 144) 98 .| 6} 169). 661... a Is7
236| 221| 244 Sl Iba: 56). Lia, 241 | Illes
30} 25) 19 ees ti. GSE Gee Be 19 IIIz7
230| 227| 97 bie | 16i... V6 3h 86 | III79
106} 80| 74 ip aii 43 $i; 30 | IIIo
130| 115 Ta iL: ERL 64 IIl»
111| 128| 82 2 e 42i ti 2 61 | III»s

83 Hai g ooo Peg 47

225| 202! 100 SoTa IB a: 94

4,205/4,409/2,677 |. ..| 13 |. . .| 85 (2,521 |61 | 8 |1,552| 5 |20|...| 2 |2,817

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 405

An examination of Table VI shows how extensively
changing took place in almost every one of the experi-
ments. The contrast with the result of Table V is very
striking.

w Ab
A
G R | N

Yellow, White, Abnormal by Gray, Red, Normal

The experiment was made once one way (Table XI) and
seven times reciprocally (Table XII). In the first case

TABLE XI
ParENTS: YWAb? By GRN J
F,: GRN 9—YWN ¢

_GRab | ywn | GwN |YRab| GRN | ywab | GWAb| YRN |
glelalelalelalele|ele|eialelale|
3 |105| 58 | 25 {11/1 |...1 2 |167/145| 741 47 |...131 5110 | Tes

nearly all of the GRAb females are of the normal type.

The only GR males that are abnormal are single cross-

overs (Diagram V). This means that the heterozygous

females are affected more easily than are the pure males;
TABLE XII

PARENTS: YWAb<o By GRNQ
F,: GRN 9—GRN g

GRAb YWN GWN YRAb GRN | YWAb GWAb YRN
a |elelelelelelela|elalelelelale
j f pe
re tb Si. cL e sw ck is a s
ost s] I i 4 OUR ais gi oe El i
idee. alila na er a e e ae
1 aiak do oed A AOS a a
3 69| 40|...| 5]. 488! 648| 59/...) 6 |...) 3 III;
a7 i. at 67| 113|.. aata IIs
1 OL sko Jr. ; 219| 249| 36 1 | 1|...| Uls
| j |
Total S 174 s08 1 isi a. 1,489/2,320| 169 !...| 9 Be a.

4 There is an exceptional case in the table, viz., two GWN 9.

406 THE AMERICAN NATURALIST [ Vou. XLIX

but even amongst the females a large percentage of gray
reds are normal. The yellow white abnormal class is
relatively much more abnormal, i. e., relatively fewer
have undergone the transition.

The results from the reciprocal cross are given in the
next table. Here the F, male contains two dominants
(GR) and one recessive factor (N). The females GRN
YWAbD carry only one dose of Ab, yet they are largely
abnormal. The GRAb are single cross-overs.

Yellow, Red, Abnormal 2 by Gray, White, Normal 3

Only one experiment of this kind was made, but as the
number of F, flies was rather large the results may be
given (Table XIII).

Y R Ab f
DiacramM VI.
TABLE XIII
PARENTS: seperated BY GWN 9?
: GRAb9—GWN ¢
O GWN y YRAb GRAb YWN GWAb YRN GRN YWAb
S/S Pi ePlaelegilelelalelaelelale|e| 9
166 | 171 135 |.. a A oe a ala a 1 | 36
taisi a hee ee a a | 31
31013151179... J. ee ee ie 2 8
217 (223| 165 |...-| 1 |1aO | Pee Bole. hes 1 | 101
791 | 824| 662 |....|...,1718] 8 |....1 7.18 lect...) 2 1268
In this cross the gametes are as follows:
Non-Cross-over Single Cross-over Double C er
GRAb GRN
YRAb YWN YWAb
GWAb
YRN

The F, male is @WN and contains, therefore, one domi-
nant sex-linked factor, viz., G. Therefore, all of the F:

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 407

females are gray. The F, male classes alone can be used
for testing the extent of crossing over.

CHANGE OF TYPE AS THE CULTURE GROWS OLDER

The preceding tables do not bring out the change that
takes place as the culture gets older—a change by which
the abnormal classes become replaced by the normal
classes. A few results will therefore be given in detail to
illustrate this relation.

In none of the relatively few counts in Tables I, and II,
involving two pairs of factors (RN and WAb), was any
change in type during the time of the experiment noticed;
but in other cases a very marked change was observed as
the cultures grew older.

In the two following tables consecutive counts of the F,
flies (from YRAb g by GWN @) emerging from day to day
from the same culture are given. The change of the
YRAb to the YRN and of GRAb to GRN is very striking.

TABLE XIV

PARENTS: YRAb¢ By GWNQ
F,: GRAb 9—GWN ¢

YRAb | GWN | YWN | GRAb | YRN | GWab |YWAb| GRN
essere | | | AD EEEE
giejalelalelalelale/a|ejaleja|e¢
eS ME Ob ae, a) OB) Fo Pa
AFA) Md i BO Be A ee othe
7 19| 34|. Ba ee haa ae ed
135| | 116/171 16618 PS
Tins t ak BLY sl.) | 1
38|...| 39| 39| 1 43
33| 38 64. 1
40 36| 27 36 +
63| 51 6| 74 81
179|...1216|210| 2 |...|...1992| 74 |...| 2 | E A 88

In the next table two F, counts are given derived from
GWN È? by BRAb ¢ grandparents. The GRAb changes
to GRN and BRAb to BRN.

408 THE AMERICAN NATURALIST [Vou. XLIX

TABLE XV

PARENTS: BRAb¢ By GWN@
F,: GRN@ (or sLIGHTLY Ab)—GWN g

GRAb GRN BRAb | BRN | GWAb | GWN | BWAb | BWN |
Counts ' r i |
g|alelalelale|alelaleialelae.
| I
E T T a a Ea TrA
e oie) Oli! Bet ar at.. 0t. | 2]
3 | 1| 0| 17/24] 0| ot 6| 7|...|...|15|24 21 /6| HIG
4 | O| 0} 36|40| 0| of15|12)...]...)41)30]...|...]15 | 22
5 | 6| 3/76 \71 o| ol31 20/...|.../67)81)...|.../28 21
1 | 19} 24] 1] o/11| 6| 2| o.. 23 |20 .| 10 | al)
2 | 26|28| 0| 0| 7\10| 0| ol.. 60|20| 1 |...| 912|
3 |29|30| 0| o| 8| 2| 3| 0l.. 28 |37 lth Titans
4 | 8/13] 2] 0] 6| 5 Shet 11 |14 | 4| 4|
5 | o| o| 27 |43| 0} opn.. 32 42 Telr
6 | o! olis oO! of17/16|...'... 51/49 [10 21 |

In the next case eight consecutive F, counts are given. The
GRAb changes to GRN. In the first three counts there
are 221 GRAb to 76 GRN, or 3 to 1. In the last three
counts there are 148 GRAb to 296 GRN or 1 to 2.

TABLE XVI
arrr GRAb ¢ By YWN 9
RAb 9—YWN
YWN GRAb | YRAb GWN | ywab | GRN | YRN | GWAb
e SAIL O1AIL O14] 9 9 | FIO AIL APIA] G
106 | 90}. 011100! 0 | 6.1 4.1 Eroro 1-167 BE) Ett
301-37) -17)-190)-0-1.6 1616 1-04-06 o 397 0 4:0
O21 36 281-611-076 i eT 81 6 aj 400
PID e Hororo 00o or aloro
5-4- 1i 1 04.6 | oloo] ó 37414 E,
6zi GB) 14) l0 loiil 1i 0i olizo
47| 28| 23| 27 0 | 0 | 0 | 0 | 0 | 0 | 18| 36] 0 | 0
42| 63| 35 39 | liit Oio 0o] 4160 01-0)
} * f —
325 | 307 | 219 250| 1 iiaia Lo oam of 0 1 8

Finally six cultures are given in the following table of Fz
flies from YRN 2 by GWAb ¢ grandparents. The GWAb
males change to GWN males and the GRAb females to
GRN females. In both cases the increase in the normal
flies in the last two or three counts is marked.

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA

F,: GRAb 9—YRN ¢

TABLE XVII
PARENTS: GWAb od By YRN?

409

YRN | GWAb |YWAb| GRN | yrab| GWN | YWN | GRab
ue alelalele|elele|alelelela|e¢
ics} 4) 48) a ak i a Da Bae ee ee S | 46
w 88) SS aes Pap ee Nee ar hat ee | 73
ee ai aes Nr ae oe rp eta ES Gage eS Oe ae A 57
me err oe es Be Re ade ae 35
| sej OB) SBP Lee see 31}. | 1 | 3
25} 88) OO a 1| 10). EME Ree OS oo 3
60} 49] 49|...]...|... ‘eee oe age Meeks Sah EH 11
87) AO) BBL daia: Sea ee ie Wit a
38; 43/2...) 1)... seen ae rd te a ek a re
T E et Be ST ae We es te ee oe ee sais
: | |
Wisc.) a r nie r. a a | 80
36| 43] 28). Ee E a 44
50| 46| 20). ...| 29 es ye en 24
29| 48] 6l. xag a. z
75| 98} 53l. 14] 88 ro eae 3
His... 28 33 oii; he e ee pera a | Dn RN E 46
art a a ee as mt pee | eas coe 48
arte See LS Gand lol 2
42| 31 Shei 4a. a ee e Meee Sir mcr) Sean
10} skad hash as 26 30 2).
Mia. 36) w aklo 6| 2 re oes bed Sar 39
96) 994 gels Bed. 47 |. a ree aS i
Syl dil al did i. 1g RO a O82 4
63} 68| ol. whe ae a Og T aN ee ae
g a pk 8| 23 pale
Tw) s AL wk da eh a
144| 181, 126 oo E Pe Be 8 41
97 78 2. 16l VEEL AE 3 Ji. 187
43] 2l l. ad ds al eat SE 81
IIIz9... 46 39 47)|. OAS BE || Bt Be) Saray came A 65
23| 24| 28 ila eek Ba 19
18| 12] 17) CAL el aod: 1
il B 8 ke 1| 12 iat 1
$2) w o. bes ee 56 i. 2
62 erie oe Oa, ae 46 . 63 i
a a o l 1| 9.. 40l il
1,431 |1,608 |1,101 |...| 6 |...137/687| 7 | 2 1631! 0 |11' 0 | 0 | 938

TESTS OF CHANGED-OVER CLASSES

In a number of cases in which some members of an ab-
normal class changed over to become phenotypically
members of a normal class; some of these apparently
normal flies were tested under conditions favorable for
the appearance in the next generation of abnormality.

410 THE AMERICAN NATURALIST [Vou. XLIX

These cases may be given. Two kinds of crosses are ex-
pected. In a few cases the normal will be found to be a
true normal (single cross-overs) and give therefore only
normal offspring when bred to normal (recessive). In
other cases the expectation is for abnormal offspring,
and where change of type has been extensive, these kinds
will be in the majority.

In experiment II,, one gray red normal female when
tested gave GRAb J and 9.

In experiment I, seven GRN 2 were bred to BWN g.
Four gave some abnormal offspring and three gave only
normal offspring.

In experiment I, five GRN 2 were tested. Three gave
some abnormal, two gave only normal offspring.

In experiment II; seven GRN 2 were tested. Four
gave some abnormal offspring, and three gave only nor-
mal offspring.

In experiment IIT, one GRN ¢ tested gave some ab-
normals,

In III, one GRN 2 bred to YRN g gave some GRAb gg
and 99.

In III; some BRN ¢ were bred to their BRN sisters.
All BR offspring were abnormal.

In IIT,, GRN 2 paired to GRN ¢ gave GRAb ¢ and &.

In III, GRN Ẹ bred to GWN brothers gave GRAb 2
and GWAb ¢ and 9.

In III,,; GRN ¢ to GRN ¢ gave GRAb ¢ and 9.

In III,, GRN 2? by GWN ¢ gave GRAb 2 and GWAb 3
and 9.

In III, GRN 2? by GWN ¢ gave GRAb 2 and GWAb J
and 9.

In III,;, GRN 2 (17) by GWN ¢ (4) gave the same re-
sults as the last. .

In III,;, GRN 2? X YWAb ¢ gave GRAb ¢ and 8.

In I.s, GRN 2? to GWN ¢ gave GRAb ¢ and 2 and
YWAb ¢ and 9.

In II, GRN 2 to GWN ¢ gave GRAb ? and GWAb 3
and Ẹ in three different tests.

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 411

In IIT,, one GRN ¢ (the only one present) when tested
gave GRAb J and 2? and GWAb g and 9.

In ITI,,, GRN 2? to GWN ¢ gave GWAb ¢ and @ and
GRAb 2.

These results show without any question that in the
great majority of cases the phenotypic normal class (when
abnormality is expected) is in reality made up largely
(entirely, except for cross-overs according to expecta-
tion) of genotypically abnormal individuals. Their ab-
normality is shown by suitable breeding tests such as
those here recorded.

INFLUENCE OF THE FACTOR FOR BLACK ON THE REALIZATION
OF THE ARNORMAL CONDITION

Some of the evidence seemed at times to indicate that
flies heterozygous in black are less likely to show the ab-
normal abdomen, but even if this is true it is still uncer-
tain whether this might not be due to other conditions
than those caused directly by the heterozygosity for black.
It might be that the black stock contained other factors
that influence the cross. Moreover since the number of
eggs laid by a given kind of female determines how many
larve will appear in a given time, and since the relation
of the larve to the food is an important factor in the
results, it seemed hazardous to put any emphasis on such
results.

In order that the heterozygous flies might be reared
under conditions that the control showed were favorable
for development of the abnormal condition in homozy-
gous forms, some black, red-eyed normal females were
mated to gray, white-eyed abnormal males. After the fe-
males were fertilized they were put into a new bottle with
some of the stock white-eyed females (fertilized). Some
‘of the daughters were red- and some white-eyed; all of
the latter were very abnormal, but the red-eyed females
(heterozygous) were all normal through five counts. At
the fifth count the white-eyed males that had been ab-
normal up to this time became normal. The result is in
accord with many similar observations; for as conditions

412 THE AMERICAN NATURALIST [Vou. XLIX

alter, the abnormal males first change to normal, then the
heterozygous females, and lastly the homozygous females.
Several attempts were made to find out if, when the
F, female, heterozygous for abnormal abdomen, is her-
self abnormal, her offspring are more likely to be abnor-
mal than when she is normal. There is evidence every-
where throughout the tables to show that the condition
of the mother has absolutely no effect on her offspring.
In December—January, 1914, the following experiments
were made which are the converse, in one respect, of
some of the preceding experiments since black abnormal
females were used. The crosses are indicated below.

(1) Black, white, abnormal 2 by gray, white, normal g.
(2) Gray, red, abnormal 2 by gray, white, normal ¢.
(3) Black, white, abnormal ? by black, white, normal JZ.
(4) Gray, white, abnormal 2 by black, white, normal g.

The F, females from (1) compared with (2) should
show whether females heterozygous for black (and ab-
normal) are less abnormal than those pure for gray;
provided, white and red eye make no difference in the
development of abnormality. The F, female from (3)
compared with (4) should reveal whether pure black
heterozygous for abnormal are less abnormal than flies
heterozygous for gray as well as abnormal.

The results need not be given in detail. It was found
that the (F) daughters from (1) show the same degree
of abnormality as those in (2). Hence heterozygosis in
black need not have any influence on the realization of ab-
normality. The mothers were not, however, in the same
bottles, but in different cultures kept as much alike as
possible. To this extent the experiments are unsatisfac-
tory. It was found that F, females from (3) were like
those from (4), hence no evidence was found that the
heterozygous type is more affected than the homozygous
black. But here also the flies were reared in different
bottles. In order to overcome this difference, some ab-
normal females that were heterozygous for black were
bred to black normal males (both having white eyes).

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 413

The daughters were either heterozygous for black or pure
black, likewise the sons. Hence, direct comparison could
be made. The following protocol gives the results for
four successive counts:

1. Black 9 quite abnormal. Intermed. 2 quite abnormal.
Black ¢ quite abnormal. Intermed. ¢ quite abnormal.

2. Black Q quite abnormal. Intermed. ? quite abnormal
Black ¢ quite abnormal. Intermed. jf quite abnormal.

3. Black Ọ fairly abnormal. Intermed. ? fairl a
Black ¢ very abnormal. Intermed. i very abnormal.

4. Black 2 none present. Intermed. 2 fairly abnormal.

Black g fone quite, Intermed. ¢ quite abnormal.
one fairly abn.

The evidence shows no difference between the extent of
development of abnormality in the homozygous black and
heterozygous black females and males.

In another way an attempt was made to get light on the
same question. Red black females were mated to white
abnormal males; and, simultaneously, red gray females
were mated to white abnormal males. The females were
later put into the same bottle and their offspring reared
together. All the daughters for four counts were normal.
At the fifth count an attempt was made to separate the two
classes of daughters, which is possible, because the off-
spring heterozygous for black are darker than the grays.
The heterozygous pairs were normal or slightly abnormal
while the pure grays appeared a little more abnormal ;
but the difference is hardly to be relied upon, since the
abnormality is less striking in the black flies.

To test this possibilty some of the preceding experi-
ments were carried to F,, when pure abnormal grays, in-
termediate and pure abnormal blacks appear. The most
abnormal grays were no more abnormal than the most ab-
normal blacks, which so far as it goes shows that the
homozygous black flies themselves may be as abnormal as
the grays under the same conditions and with the same
ancestry.

414 THE AMERICAN NATURALIST [Vou. XLIX

INFLUENCE OF THE FACTOR FOR YELLOW IN THE REALIZA-
TION OF THE ABNORMAL CONDITION

Experiments similar to those with black were made
with yellow. Yellow, white-eyed normal flies were bred
to gray, white-eyed abnormal. In ten tests, the F, fe-
males were abnormal in eight cases, and normal in two.
It is not apparent that the yellow factor has any decided
influence on the results. :

In order to compare the females heterozygous in black
and in yellow, and others also homozygous in both, the
four following tests were made. By utilizing the red and
the white eye colors it was possible to distinguish be-
tween the different classes of females. Previous experi-
ments, described above, had made it highly probable, that
no effects are produced by red and by white, but by making
reciprocal crosses here this possible effect was more cer-
tainly eliminated. In all cases the females were mated
separately for a few days to gray, white-eyed abnormal
males to better ensure fertilization and were then brought
together in one bottle.

(1) (2)
Han 2 by GWAb g. ena 2 by GWAD J.
BRN 2 by GWAb Z. YRN@ by GWAb J.
1.{GR 9 Fairly Ab. 1f 3 GRN 9.
GW 9 Fairly Ab. 1 GWN 9.
o (GR Ọ Like last 3 12 GR slightly Ab 9.
GW @ Like last. ‘| 3 GW slightly more Ab 9.

3 {GRQ Like last. 1 GRN 9.
‘GW @ Like last. 3.) 8 GW slightly Ab 9.

| 4 GRNQ.

4) 8 GWNQ

_§ 1 GRN

9-110 GWN 9

{ 2 GRNỌ

6.114 GWNQ

In (2) the females heterozygous in yellow were slightly
more abnormal than those heterozygous in black or at
least they gave such an impression.

No.583] ROLE OF ENVIRONMENT IN DROSOPHILA 415

(3) (4)

em 9 by GWAb g. ee 2 by GWAb g.
GWN? by GWAb g. GRN 9 by GWAbD &.
1.{GR slightly Ab 9. 1 5 GR slightly Ab 9.
ə N or nearly NỌ ‘] 2 GW nearly N 9Q.

"(GW fairly Ab 9. o |14 GR fairly to slightly Ab 9.
3) 5 GRN and 4 GR quite Ab 9.) 6 GW nearly N 9.

L On T quite to fairly Ab. 3.{GR slightly Ab to N9
ra 13 GR i 28 GRN 9.

"F16 ae quite to slightly Ab 9. 3 GWN Q.

5 15 GRN |17 GRN 9.

‘1 9 GW fairly Ab to NQ. ") 2 GWN 9.

K 6 GRN 6.{ 6 GRN 9?

‘110 GWN Ọ.

In (3) the females heterozygous in black were slightly
more normal than the grays. In (4) there is hardly any
difference, but so far as difference is noticeable the hetero-
zygous type (GW) is again more nearly normal. This
difference was even more apparent in a second culture
from the same parents.

THE RELATIVE INFLUENCE OF THE EGG AND OF THE SPERM
ON THE CONDITION OF THE HETEROZYGOTE

At the time when the F, generation began to hatch
the extent of the abnormality in the females was noted.
This was at the time when the flies were taken out to be-
come the parents of the F, generation. The terms used
were necessarily somewhat vague, but give a fairly ac-
curate idea of the condition of the cultures as a whole.
If most of the flies were distinctly abnormal this was indi-
cated by Ab to N, if more of the flies were normal or nearly
so but some were abnormal this was indicated by N to Ab.
If the flies were normal in appearance this was indicated
by N. The results for many of the cases recorded in the
preceding tables are brought together in the next table.
he results are far from uniform, as was to be expected,
but in most cases it will be noted that when the female
was normal and the male abnormal, the daughters were
frequently normal or nearly so, while in the reciprocal
cross the tendency was in the opposite direction, i. e., the

416 THE AMERICAN NATURALIST [ Vou. XLIX

daughters were more likely to be abnormal. These
records were made at a time when no suspicion of such a
relation was present in my mind. If these observations
are to be trusted they mean that when abnormality comes

| abn. | Abn.toN. | N.toAbn. | N.

YWAbg fo CSS Ee eis | 1 7
GRN foc [Gogooonoodns Coen ccc econ 2 (few)

GRAbg fcc 14 2 6 1
YWNo fic rer ease 1 1

BRAbg foc | 3 3 4

YWN9? | >
anisat R A Se iO,

in with the egg the heterozygous female is more likely to
show abnormality than when the abnormality comes in
with the sperm. Conversely the result may be stated in
this way—when normality comes in from the egg the
daughters are more likely to be normal than when the
normality comes in from the sperm. In other words, we
might extend this conclusion and state that the cytoplasm
of the egg has an influence on the soma of the individual
which arises from it, or the cytoplasm plus the nucleus
of the egg has more influence on the next generation than
the nucleus of the sperm.

When this possibility was realized it was evident that
some of the experiments must be repeated under condi-

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 417

tions where a more exact comparison between a cross and
its reciprocal could be made. In the autumn of 1913
I went over the ground again with this object in view. It
was found that the F, females heterozygous for abnor-
mality are just as likely to be abnormal when their ab-
normal factor comes from the father as when it comes
from the mother. The extent to which the abnormality
is realized depends on the condition of the food. This in
turn will depend in part not only on its amount but to
what extent it is worked over by the larva which again
depends, in large part, on the number of eggs laid by the
female. To this extent and only in this sense does the
condition of the mother affect the condition of her daugh-
ters. If the females lay too many eggs for the amount
of food thatis present, crowding results and the daughters
show abnormality to a less degree than when fewer eggs
are laid (that hatch) and little competition takes place.
Now the normal female is more likely to lay more fertile
eggs than the abnormal female. Hence other things be-
ing equal the heterozygous daughter of a normal mother
is more likely to be normal than the heterozygous daugh-
ters of an abnormal female (which are therefore again
more like their mother—very abnormal in this case, be-
cause the former mother is more likely to lay more eggs
than the latter). The relation between the two cases is
therefore not owing to the egg transmitting abnormality
to the daughters better than the sperm, but to the number
of eggs likely to be laid by the mother in question.

In order to examine further whether when abnormality
comes in with the egg it is more likely to be shown in the
F, heterozygote, a number of parallel experiments were
made, of which the following are samples:

Gray Rep ABN. ? py Wip g. Gray Rep ABN. ji By Wi 9.

(1) Very abn. g and 9. (1) Fairly abn. 9.

(2) eons (a few slightly ab- (2) Most fairly, a few very abn.

rmal). (3) Most fairly, a few very abn.

(3) Mohini (a few slightly ab- (4) Slightly abn

normal), (5) Slightly abn.
(4) Normal (a few slightly ab- (6) NỌ (40) 1 slight abn. 9.
normal). (7) NQ.

418 : THE AMERICAN NATURALIST [Vou. XLIX

While it is true in the first count above that when ab-
normality entered through the egg there was greater ab-
normality in the offspring, yet this is offset by the counter
evidence in this set that the change to the usual pheno-
type took place sooner in this set than in the others.
This point will be taken up again in connection with other
data.

In order to compare, under changing conditions, hetero-
zygous and homozygous females, some white abnormal
females were mated to red abnormal males, and, inde-
pendently, some other white abnormal females were
mated to red normal males. After several days both
kinds of females were separated from their respective
males and put together into a single new bottle. All of
the daughters had red eyes. In the first count two types
of females could readily be distinguished. Some were
quite abnormal, others were slightly abnormal or normal.
In the second count (next day) again two types appeared,
one quite abnormal and the other slightly abnormal fe-
males. In the third count some females were fairly ab-
normal, the rest normal and this held for the fourth
count. The result leaves little doubt that under these con-
ditions, the homozygous were abnormal and the hetero-
zygous less abnormal or quite normal.

In order to see if the factors for red and for white
affect the condition of the zygote, homozygous for ab-
normal; white abnormal females were mated to red ab-
normal males, and, separately, other white abnormal fe-
males to white abnormal males. After several days the
females were put together in a new bottle and the males
removed. Through five consecutive counts the red and
the white daughters were alike, at first quite abnormal,
later nearly normal. Red and white abnormal females
therefore behave alike.

PRESENCE AND ABSENCE

It is not without interest to examine the bearing of
these results from the point of view of the ‘‘presence

No.583] ROLE OF ENVIRONMENT IN DROSOPHILA 419

and absence’’ hypothesis, even although I myself prefer
a more non-committal form of factorial interpretation
than that offered by the ‘‘presence and absence” theory.

The abnormal male (Ab) has one dose of abnormality
and the degree of his abnormality is the same as that of
the female (Ab, Ab) with two doses. But the hetero-
zygous female, AbN, has only one dose (or factor) for
abnormality. The degree of abnormality that she shows
is very variable ; she is less abnormal on an average, than
the abnormal male.

Which condition is to be interpreted as absence—the
real absence of one Ab in the male, or the absence of one
Ab in the other (normal) chromosome of the female? A
moment’s thought will show, however, that nothing of any
value can come from a discussion of this question, be-
cause the heterozygous female (AbN) differs from the
male not simply by the factor N, but by a whole chromo-
some including amongst other factors a factor which in
duplex produces a female. Moreover, an advocate of
presence and absence might maintain that the relation
of a dominant to the normal allelomorph is not the same
as the relation of a normal allelomorph to a recessive for
it is the latter that is ‘‘absent.’’ In other words, he might
conceivably accept the hypothesis of absence for a reces-
„sive, but reject it for a dominant mutation.

I have pointed out elsewhere that it seems to me un-
warrantable to interpret the absence of a character to
mean necessarily an absence of a factor in the germ
plasm.” Yet this literal interpretation of the presence
and absence hypothesis has often been made. If the
linear arrangement of factors in the chromosomes be ad-
mitted as a plausible hypothesis the absence of a factor in
this literal sense would mean a hole in the chromosome,
and a corresponding displacement of the linear sequence
of factors. The evidence does not support this hypothe-
sis. On the other hand, if the locus of a factor be con-

5 Although of course a changed factor might cause the failure of some
substance to develop that is necessary for a given reaction.

420 THE AMERICAN NATURALIST [Vou XLIX

ceived as a particular chemical body at a given level in
the chromosome then any change in this body would be
expected to affect one, or more, or even, at times, all
characters of the complex that gives rise to the body
character or characters. The particular change might
involve no more than a rearrangement of the materials
of the locus or the addition of a chemical element (or com-
pound) or the loss of a chemical element (or compound)
—any one of these changes might lead to the loss of a
character in the soma. As to what happens in the locus
we can form no idea, and so far as the mechanism of
heredity is concerned it is a matter of no immediate im-
portance. If, however, any one finds a greater satisfac-
tion in the view that a loss of something from the locus
(an atom or a molecule) leads to a recessive character,
there is not the slightest objection to his doing so, unless
by loss he means the loss of the entire locus. He may do
this if he rejects the linear arrangement of different
material in the chromosomes, but if he accepts the latter
view the assumption of a literal absence involves him in
unnecessary difficulties. It is not as generally under-
stood as it should be that the facts which the presence
and absence theory was constructed to account for do
not require the assumption that the absence of a char-
acter means the absence of a factor in the germ-plasm. °
It is entirely gratuitous to involve the theory of Mende-
lian heredity in such an interpretation which adds noth-
ing to the theory and by bringing in a new hypothesis may
involve the Mendelian theory in further difficulties. An
example may make this clear. It is known that when a
chocolate mouse is bred to gray and the F, grays that
result are bred together there appear in F, grays (9),
cinnamons (3), blacks (3) and chocolate (1). Gray was
written GBCh and chocolate gbC, which gave in F, @BCh
(9) GbCh (3) gBCh (3) and gbCh (1). The occurrence
of the black class of gBCh is accounted for through re-
combination. But the same end is accomplished if we

No.583] ROLE OF ENVIRONMENT IN DROSOPHILA 421

suppose that a factor in the wild or agouti mouse mu-
tated so that the recessive black was produced as a result
of the activity of the new gene. Then bl—black, and Bl
= gray with respect to black. Likewise cinnamon agouti
may be represented by ci, and gray, with respect to cin-
namon, by Ci. Chocolate is then the double recessive
blei and the symbol Ch for ‘‘chocolate’’ becomes super-
fluous. All the experimental results may be explained
on this basis.

It is not necessary to try to state what kind of a change
in the germ-plasm led to these two mutations. The fac-
torial hypothesis should be entirely non-committal as to
the kind of change that took place, for we can know noth-
ing about the nature of the change, yet the results are
predictable as well on one view as on the other.

There is another way to interpret a dominant factor
like this one that gives abnormality, namely, that there is
present in the normal fly a factor that restricts the yellow
of the abdomen to the bands. When this restrictor, ab,
changes (Ab) the yellow is dispersed over the abdomen
and the black bands fail in part or entirely to appear.
The new factor, acting with the rest of the cell, gives ab-
normality, just as the normal restrictor or inhibitor (ab)
acting with the rest of the cell gives normality or band-
ing. The interpretation is non-committal in regard to
the nature of the change, which is an advantage in the
direction of simplification. In contrast to this view, a
different interpretation of the meaning of a restrictor
might be entertained on the presence and absence view.
It might be said that a restrictor factor has been ‘‘lost’’
from the normal fly, which failing to restrict the color
has given rise to abnormality. The first objection to this
hypothesis is that it postulates (as above) the nature of
the change in the germ plasm, because it says something
has been lost. The second objection is that the facts
show that a restrictor has not been lost sensu strictu
because there is a wide range of variation in regard to

422 THE AMERICAN NATURALIST [ Vou. XLIX

the loss of banding and in certain environments there ts
a return to the normal banding to the extent that the fly
can not be distinguished somatically from a normal
banded fly. My contention is that since we know noth-
ing of the nature of the change in the germ-plasm that
leads to the appearance of a new or the loss of an old
character, any assumption that is based on the nature of
that change involves the Mendelian interpretation in un-
necessary implications. We need only assume that some
change has occurred, as the result indicates; my formulas
give the same results as do those of presence and ab-
sence and serve the purpose of briefly indicating a change,
the machinery involved, and the necessary consequences.

OTHER TYPES OF ABNORMAL ABDOMEN

Irregularities in the arrangement of the rings of the
abdomen are not uncommon in Drosophila. Sometimes
they appear to have been caused by injury to the
larve or pups, but still other abnormalities are inherited
in the sense that they occur in certain stocks in more or
less definite percentages. Several times I have bred ab-
normal types: some of them have failed to reappear;
others have reappeared in a certain percentage of cases.
Two stocks of the latter kind may be referred to here.
My main purpose in describing them is to anticipate the
possible confusion that might arise if some one finding
these or similar ones should suppose them to be the same
types as those described as abnormal abdomen in this
paper.

The six drawings in Fig. 2, a-f, represent some of the
characteristic types of a certain stock. The failure of the
third abdominal ring to extend across the middle line, as

e It is not an objection to this hypothesis that an absence (loss of re-
strictor) appears to dominate presence. This interpretation rests on a com-
plete misunderstanding of the nature of the factorial hypothesis; for, ab-
sence here means only that the rest of the cell fails to produce banding

when a certain factor is lost, or, when as in the female, one of the inhibitors
is lost.

No.583] ROLE OF ENVIRONMENT IN DROSOPHILA 423

Fig, 2.

seen in the first two figures, is the more usual form of
abnormality in this stock; but modifications of other
rings shown in the other figures are probably due to the
Same cause or causes. Two consecutive rings may form
a spiral as shown in c or half of a ring may be absent
as in e, or an entire ring may be lacking as in f. Indi-
viduals with abnormalities like those shown in the figures
were bred to each other usually three or four together.
Their progeny was examined and the normal and the ab-
normal types recorded. The latter were again used to
breed from for three or four generations. As no increase
in the proportion of abnormal offspring appeared, the
breeding was abandoned. The results given below are
in the order in which they were obtained without regard
to the generation in which they appeared.

424 THE AMERICAN NATURALIST [Vou. XLIX

In these counts there were normal to abnormal flies both
of whose parents were abnormal. Since the normals also
throw some abnormals it is probable that there is here a
case of multiple factors like that of beaded and truncate.
Special tests will therefore be necessary to work out the
case.

N a a ae N Ab N Ab

85 E ee oa et 40 0 73 0
6 1 Woe B p il

28 5 33 2 52 7 32 5

32 10 20 2 15 5 105 3

15 3 13 4 w l i 37 10

40). [8 ee G eee

yin ey | | | |

The abnormal abdomens shown in Fig. 3, a-f, are from
another stock, discovered by Mr. Bridges. While some of
the types are not unlike those of the last series, they are
more extreme and there can be no doubt but that the two
stocks have a different composition.

In the last drawing the entire fly is figured (the one
wing present has been cut off at the base), the upper half
of the thorax is absent. This same condition appears in
rather high proportions in certain other stocks, notably
in vestigial stocks. Even both sides of the thorax may
be absent so that the head rests above on the abdomen.
Although I have tried a number of times to obtain pure
stocks of this thoracic abnormality, I have never suc-
ceeded in getting a stock that did not throw a high per-
centage of normal individuals.

This type of abnormal abdomen appeared in a cross
between a cream male and an eosin female as a single
female, Fig. 1, a, which had only three instead of five
bands in the abdomen. She was mated to one of her
brothers, and produced offspring all of which as far as
known had normal bands. A pair of these offspring gave
in the next generation abnormal bands in about half of
the flies. The abnormal band acted as a recessive. In
subsequent generations the character behaved in an irreg-

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 425

Fie. 3.

ular manner though no serious attempt was made to dis-
cover the cause of the irregularity. A stock of cream eye
color was made up from this strain and selection against
the abnormal was carried out in a rough way for several
generations, but this selection failed to eliminate the ab-
normal condition, and a recent examination of the stock
showed that for a year the abnormal abdomen had main-
tained itself and was still present in about half of the
ies.

A male was again crossed to a wild type female and
gave normal F, flies. In F, there were 128 red normal
females, 29 red normal males, and 28 eosin normal males.
No abnormals appeared. Crossed to eosin the F, were

426. THE AMERICAN NATURALIST [Vou. XLIX

PARENTS: AbE By NE

F,: NE

EN ọ EN ¢ | EAb ọ | EAb ¢
78 56 1 1
13 14 0 2
46 40 0 2
6 pi 0 0
35 32 0 1
39 46 7 1
217 195 | 8 | 7

all normal; these inbred gave in F, the classes given above:
Two eosin females heterozygous for white were crossed
each to an abnormal male. The normal F, daughters
were bred to those sons that had white eyes, and gave the
following kinds of offspring:

PARENTS: E-WNQ By EAb ĝ
F,: EN? (By WN ĝg)

ENọ | ENa | wabo | BAbo | WN 9 WNg | wabe | wabe
18 26 1 1 20 15 1 1
34 28 1 $s 25 2 1
52 | 54 2 2 | 8 woa 2

Abnormal males were bred to eosin females and gave, as
before, normal F, sons and daughters. Some of the
daughters were backcrossed to eosin cream abnormal
males and gave the following results:

PARENTS: AbE BY NE
F,: NE (BY AbE ĝ)

EN 9 ENG EAb ọ | EAbg
35 32 0 1
21 4 1 2
48 40 0 1
72 72 8 6

176 149 9 10

These tables show that the abnormal condition rarely ap-
pears in F,. Its realization must be due therefore either

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA . 427

to multiple factors or to environmental effects. That
the former rather than the latter is the main explanation
is shown in the frequency with which the abnormal flies
appear in the inbred stock (where the conditions are the
Same as in the experiments) and the rarity with which
the character appears when the stock is outcrossed.

THE NON_INHERITANCE OF AN ACQUIRED CHARACTER

The acquirement of a new character by a pure stock
implies by definition the capacity of this stock to respond
to the imposed conditions. Conversely if an animal does
not acquire a new character in a changed environment it
does not come within the scope of the definition of an
acquired character, and even should its offspring show
new characters as a result of the new environment in
which the parents have been placed the result is still
excluded by definition from being a case of the inheri-
tance of an acquired character. At least this is my un-
derstanding of the use of the term and the way in which
I shall use it in the following statement.

The mutant stock of abnormal abdomen offers an ex-
ceptional opportunity to examine the possible influence
of an acquired character on the offspring. As the experi-
ments have shown this stock is very susceptible to en-
vironmental influence, and the effects produced pro-
foundly affect the structure of the organism. Moreover
it is possible to carry the stock through several genera-
tions in either of the phenotypic conditions, and then, at
will, to cause the other type to appear at once in its com-
pletest form, by regulating the external conditions in
which the young are reared.

No better material could be found for studying the
possible influence of the environment through its effects
on the soma of the individual. The evidence shows in the
clearest manner that the condition of the parent, whether
normal or abnormal in type, has no effect on the charac-
ter of the offspring. The evidence is so clear and so
positive that it seems unnecessary to elaborate the point.

428 THE AMERICAN NATURALIST [ Vou. XLIX

THE NON-CONTAMINATION OF GENES

Recently the question of the possible contamination of
genes (or factors) has been under discussion. Were
such contamination possible one might expect to find
some evidence of it in a case like this one of abnormal
abdomen, if one is justified, at all, in drawing infer-
ences from the nature of the character to the nature of the
gene that stands for that character. I do not myself
think that there is the slightest justification in drawing
such conclusions, but let us assume for the moment that
such an inference is justifiable in order to examine the
bearing of the evidence furnished by this mutant type.

The heterozygous female carries a factor for normal
and one for abnormal. She herself may be either normal
or abnormal according to the environment in which she
was reared. It might be supposed, since she is abnor-
mal, that her normal gene might be more predisposed
to contamination by the abnormal gene. The evidence
shows that this does not occur; for, by means of the link-
age we can identify the normal flies that should carry the
normal, or the abnormal genes, and we find that the re-
sults conform completely to expectation; 7. e., they are in
full accord with all other linkage results where there is
no reason to suppose that contamination takes place.

Conversely it might be supposed that if the hetero-
zygous female were normal in type her abnormal gene
might be predisposed to contamination by the normal gene,
but again the evidence contradicts the assumption.

If, on the other hand, it is not supposed that the pheno-
typic condition of the female has any part in bringing
about contamination (or in serving as an indicator, that
conditions are favorable for contamination) but that con-
tamination is due merely to juxtaposition of genes in the
same cell, then in refutation of the contamination of genes
I may cite the evidence cited above, where in several suc-
cessive generations the breeding took place from hetero-
zygous females bred to recessive males and the gametic

No. 583] ROLE OF ENVIRONMENT IN DROSOPHILA 429

ratios were the same in the late as in the earlier gen-
erations.

Lastly the tests made of individuals that were pheno-
typically normal, but genetically abnormal, showed in all
eases the validity of the genetic evidence, which would
not have been the case if the apparent exceptions had been
due to contamination of the genes. I may also cite the
two peculiar matings, B and C, recorded on page 402,
where an expected class did not appear. It might have
appeared that here actual contamination had occurred.
In reality, the result turned out to be due to a lethal
factor. Our study of these lethals, that give verifiable
results, fully under control, made it possible to interpret
this case that otherwise would have been inexplicable, and
might have been cited in favor of the view of contamina-
tion of genes. Taken all together the results obtained
with this mutant type make out a strong case against the
supposition that genes become contaminated through jux-
taposition. I shall not discuss here, therefore, the un-
pragmatic character of such a supposition, but rest the
case on the evidence from the experiments.

ON A CRITERION OF SUBSTRATUM HOMOGE-
NEITY (OR HETEROGENEITY) IN FIELD
EXPERIMENTS

Dr. J. ARTHUR HARRIS

CARNEGIE INSTITUTION OF WASHINGTON
I. INTRODUCTORY REMARKS

Every one who has had practical experience in variety
or fertilizer tests or in any other experiments involving
the comparison of field plots must have been impressed
by the great difficulty of securing tracts with uniform soil
for their cultures.

A careful examination of the agricultural literature
bearing on the question of variety tests will reveal many
cases in which the experimenters have noted the difficulty
of securing a uniform substratum, or in which there is
internal evidence for the influence of mubeatum hetero-
geneity upon the result.

For example, in 1894-1895 tests of varieties of wheat
were made on 77 plots at the University of Illinois.! As
a check on the other strains, the variety known as Valley
was sown on nine different plots ‘‘well distributed over
the area sown.’’

the yields of this variety varied from 11.7 bushels to 24.1
feahelac an average of 19 bushels which is remarkably close to the
average of all the varieties, It is again remarkable that but eight yields
were above the highest of the Valley, and but three below the lowest
of the same variety, . .

The only reasonable explanations that can be given for
such results are either (a) that the plots were so small
that the results are due purely and simply to the errors of
random sampling, or (b) that the wide divergences in the

1 Bull. Univ. TU. Agr. Exp. Sta., 41, 1896.

430

No. 583] ON SUBSTRATUM HOMOGENEITY 431

results for the individual variety are due to substratum
heterogeneity.

In either case, the results secured are obviously worth-
less as indicating differences in the value of the individual
varieties.

Seventeen years ago, Larsen? reached the conclusion
that the results of experimental tests were much more
exact when a given area is divided into a large number of
small plots upon which the tests are made than when it is
divided into a few larger plots.

Hall’ has laid great emphasis upon irregularities of ex-
perimental fields. Mercer and Hall in their interesting
paper on ‘‘The Experimental Error of Field Trials’
discuss at considerable length various phases of the influ-
ence of soil heterogeneity upon field results. In an ap-
pendix to their paper, Student® takes up the problem of
the method of arranging plots so as to utilize to the best
advantage a given area of land in testing two varieties.®

The influence of substratum heterogeneity is also read-
ily seen in Montgomery’s interesting experimental data
for wheat.”

Indeed, it is quite possible that without special precau-
tions irregularities in the substratum may have greater
influence upon the numerical results of an experiment
than the factors which the investigator is seeking to com-
pare. Elsewhere® I have shown that the sac aaeadan

ring B. R., ‘‘Andra nordska Landbrakskongressen i Stockholm,’
1897, I, Ý. 72; fide G. Holtermark and B. R. Larsen, Lanwirtschaftl. e ;

such- Stationen, p a, 190.

3 Hall, A. D., ‘‘The Experimental Error of Field Trials,’’ Journ. Board
Agr. Great Britain, 16, 365-370, 1909

4 Journ. Agr. Sci., 4, 107-127, 1911.

5 Student, Journ. yas Sci., 4, 128-132, 1911.

6 For several years, I have a eareful tests labelled each seed individually
and scattered them at random over the field to eliminate the influence of soil
een

7 Montgomery, E. G., ‘‘ Variation in Yield and Method of Arranging
Plots to Secure Comparative Results,’? Ann. Rep. Neb. Agr. Sta., 25, 164-180,
1912,

8 Harris, J. Arthur, ‘‘An Illustration of the Influence of Substratum
Heterogeneity upon Experimental Results,’’ Science, N. S., 38, 345-346.
1913.

432 THE AMERICAN NATURALIST [ Vou. XLIX

in an apparently uniform garden plot may be sufficient to
mask entirely the influence of the weight of the seed
(Phaseolus vulgaris) planted upon the size of the plant
(as measured by the number of pods) produced. It is
very probable that certain pure-line experiments and con-
clusions are entirely invalidated by the fact that the in-
fluence of irregularities in the substratum were not suff-
ciently guarded against.®

Several authors have tried to obtain some measure of,
or some corrective term for, substratum heterogeneity.
For example, Mercer and Hall (loc. cit.) have plotted the
yields across the field in both directions. Such methods,
however, give but a very imperfect idea of irregularities
in the soil. Heterogeneity is perhaps more likely to occur
as a spotting of the field than as a relatively uniform
change from one side to the other. This is clearly indi-
cated in the diagrams published by Montgomery. The
mere plotting of yields in any line across the field can not
adequately take into account such irregularities. Fur-
thermore, some quantitative measure (and some probable
error of this measure) of the amount of irregularity, not
merely a demonstration of its existence, is required.

Some generally applicable measure of the degree of
homogeneity of the soil of a field (as shown by actual ca-
pacity for crop production) seems highly desirable. Such
a criterion should be universally applicable, comparable
from species to species, character to character or experi-
ment to experiment, and easy to calculate.

I believe we may proceed as follows. Suppose a field
divided into N small plots all planted to the same variety
of plants. Let p be the yield of an individual plot. The
variability of p may be due purely and simply to chance,
since the individuals of any variety are variable and the
size of the plots is small, or it may be due in part to differ-
entiation in the substratum. If the irregularities in the
experimental field are so large as to influence the yield of

9 See ‘‘The Distribution of Pure Line Means,’? AMER. NAT., 45, 686-
700, 1911.

No. 583] ON SUBSTRATUM HOMOGENEITY 433

areas larger than single plots’? they will tend to bring
about a similarity of adjoining plots, some groups tend-
ing to yield higher than the average, others lower.

Now let the yields of these units be grouped into m
larger plots, Cp, each of n contiguous ultimate units, p.
The correlation between the p’s of the same combination
plot, Cp, will furnish a measure (on the scale of 0 to 1)
of the differentiation of the substratum as expressed in
capacity for crop production. If this correlation be sen-
sibly 0, the irregularities of the field are not so great as
to influence in the same direction the yields of neighboring
small plots. As substratum heterogeneity becomes greater,
the correlation will also increase. The size of the co-
efficient obtained will depend somewhat upon the nature
of the characters measured, somewhat upon the species
grown, and somewhat upon the size of the ultimate and
combination plots. A knowledge of the values of the cor-
relation to be expected must be determined empirically.

Fortunately, very simple formule are now available for
calculating such coefficients."

Let S indicate a summation for all the ultimate or com-
bination plots of the field under consideration, as may be
indicated by the capital Cp or lower case p. Then in our
present notation which is as much simplified as possible
for the special purposes of this discussion

{LS(C,?) — S(p?)]/m[n(n — 1) 3} — 7
oS

T pipa Er

where P is the average yield of the ultimate plots and op
their variability, and n is constant throughout the m com-
bination plots.!?

10 Trregularities of soil influencing the plants of only a single small plot
may in most work be left out of account, since they are of the kind to
which differences between individual plants are to a considerable extent
due, and are common to all the plots of a field.

11 Harris, J. Arthur, ‘‘On the Calculation of Intra-class and Inter-class
Coefficients of Correlation from Class Moments when the Number of Pos-
sible Combinations is Large,’’ Biometrika, 9, 446-472, 1913.

12 For the benefit of those who are frightened by formule, it may be
paraphrased as follows: One merely adds together the yields of a chosen

434 THE AMERICAN NATURALIST [Vou. XLIX

Ultimately, I hope to present experimental data of my
own bearing on this problem. For the present, the method
is admirably illustrated by a number of published records.

IJ. ILLUSTRATIONS or METHOD
A. Cases in which the Combination Plots are Equal in
ize
Illustration 1. Influence of substratum heterogeneity
on yield of experimental plots of mangolds.
TABLE I

YIELD OF COMBINATION PLOTS FOR MANGOLDS, OBTAINED BY COMBINING THE
ENTRIES OF MAP A BY FOURS AS INDICATED BY THE HEAVIER LINES

1,209 1,175 1,215 1,239 1,276
172 183 171 175 205
1,250 1,321 1,274 1,293 1,310
185 191 187 184 207
1,204 1,333 1,268 1,290 1,268
159 188 172 185 200
1,300 . 1,272 1,222 1,272 1,388
172 177 167 173 215
1,385 1,375 1,314 1,260 1,373
ies 194 193 180
1,380 1,387 1,309 1,314 1,380
204 202 177 188 229
1,320 1,295 1,304 1,332 1,397
180 188 187 194 226
1,331 1,264 1,310 1,325 1,337
183 183 188 203
1,404 1,325 1,334 1,335 1,312
194 190 190 ' 1 2
1,418 1,373 1,339 1,403 1,401
193 196 189 198 226

number of contiguous p plots to form a number m of Cp plots. The sum
of the squares of p is subtracted from the sum of the squares of C. d
he result divided by m[n(n—l)] where n is the number of ultimate
plots in each of the m combination plots. The quotient is reduced by sub-

pine required. Thus the calculation m the eriterion is very simple

435

No. 583] ON SUBSTRATUM HOMOGENEITY
NAG J IZ be Sd Beh ee e ea yg
yp | 3/0 | 302 | 288 | 325|32/ | 29/) 306 | 306 | 306 | 330
Yl 46 | ¥/| ¥8| ¥S| ¥/| ¥/| #5 | 4B | 57
q |290 | 307 1267 | 275) 308 | 275) 3/7 | 3/0 | 3/6 | 324
Y| 43 | ¥0| 54| YS | 40) ¥4¥| #5] 8 | 52
3 [322 | 307 |322 | 324) 330 | 286 | 300 | 325 | 302 | 278
49 | ¥S5 | ¥3| 47| 32| 40| 46) ¥7 | 47 | #97
y [397 \310 | 324 | 357 | 342 | 3/6 | 324/344) 34/ | 367
44147 | 46) 53 | 5/7) 4¥| 4¥| 471 57] 60
g | 2 78\320 |335| 350 |342 |307 | 3/0 | 322 |327 | 300
33 | 42146 | 5/1) 47 | 40) 45| 48E | 50| $2
6 (992 |30F | 3/0 | 338 |3/6 | 30/ |3 28 |330 | 328 | 3/4
37| 42 | 45| 46 | 3 | #0 | ¥ Figg 32
q | 396\3/8 |302|332|297|277)33/ | 322 | 384| 337
HL | #42 13921 50| ¥3| 37) 44| 461 6/ | 52
g [333/343 |3/8 |320| 335|3// | 276 | 323/327 | 338
2 4 | 44| 44| ¥7 | 40| 38| ¥5) 47 | 59
q | 337/336 | 324 347 |337 | 3/3 |372 300 |346 | 343
Y4) y6 | ¥6| 47| 5/ | #2) ¥8| 4/ | 52) 87
Jo |362 |350 |354| 350) 348 | 322) 325) 323/335 347
$/|52 | 57| 50| 54| #6 | #3| 48 | 32] 60
jp | 346 |362 |372 | 3#7|3#3 308 227 |328 |354 |307
$/| 55 | s7 | S¥| 46 | ¥O| 9156) oS
j2 | 327 | 34S | 33/ | 335) 342 3/6|339|350| 365 |354
47| 57 | ¥8 | 472| 47 | 44| 47| $7 |34| OF
j3 |370 |36% | 300 | 337 |32/ | 327 |3 #/ 349 | 363 | 347 |
ygi 50| 46| ¥5| #6] 76] 51| #7 | 58 | 57
zy | 3/7 |329 | 32/ | 3371340 | 3/4) 32/ | 32/ 346 | 3#
4/ | 44| 45| 52| 50 | 45| 48 | #6 | 52 | 57
jg |323|326 | 290) 328 1348 | 32513958) 332 349 | 335
44) y| Y/ |) 48| 49 | 45| 43| 44| S| TO
1G, |353 |327 |3// | 335 )33/ |306 | 3/8 3/7 |332|32/
49| 46 | ¥7| ¥7\ 46 | 48| 47| £¥) 46 | SO
17 | 397| 348 |30/ |335 |3#0 |336| 327 330|343|3/7
52| 46) 44|-47| 5/7 | 45| #6| 50 | 54 | SF
zg |392| 337 | 337 |350 |328 | 330 | 343 335 |326 | 326
46|. $0| ¥7| 52 | 47 | 47| ¥7| 47 | 47 | 56
19 \349| 365° | 357 | 337340 |332 356|336 |338 | 3/6
yé] $2| 4g | ¥7|4% | 471 0| #82 | 57 | $6
20 |352| 352 | 340/335) 332/335 | 356 355 |37/ | 376
49| 46 | 50| 51|## | 48| 50| 50) S2| 67
Map A. Pounds per Plot of of Mangolds. Data of Mercer

Roots and Leaves

and

Hal

436 THE AMERICAN NATURALIST [ Vou. XLIX

Map A represents the Rothamsted field of mangolds
grown by Mercer and Hall (loc. cit.). The upper entries
are for pounds of roots, the lower for pounds of leaves.

I now reduce the 200 areas to 50 by combining the ad-
joining plots by fours, as indicated by the heavier lines on
the map. Thus for leaves the Southwest combination
plot, Cp, is 67 +52 +56 + 51—226. Table I gives the
result.

This gives for roots:

s p) = 65715, S(p”) = 21674871, N = 200,
== $28.575, op? = 4132.8243
S (Cry = Sa, m[n(n — 1)] = 50 X 4 XxX 3 = 600,
[S (Cp?) — S(p?)]/m[n(n — 1) ] = 108104.280,
and

_ 108104.280 — (328.575)?
Pips = 412.824

= .346 + .042."

The results for yield of leaves are

S(p)==9541, S(p?) —45941, N—200,
p==47.705, cp? == 23.938,
S (C?) = 1832095, m[n(n —1)]—50 x 4 X 3= 600,
[S(C,?) — S(p?) ]/m[n(n — 1)] = 2286.923,
whence |
= 2286.923 — (47.705)?
IPE, B
Illustration 2. Influence of Substratum Heterogeneity

upon the Yield of Straw and Grain in Experimental Plots
of Wheat.
18 The standard deviation is most conveniently calculated in eases like the

present, in which one requires the summed squares of actual values for other
purposes from

= .466 + .037.

Op? = = (p*)/N — [E (p)/N F.
14 The probable errors have in all cases been calculated upon the actual,
not the weighted, number of ultimate plots as N.

No. 583] ON SUBSTRATUM HOMOGENEITY 437

The wheat field of Mercer and Hall is divided into
25 X 20—500 plots, Map B. Combining the plots by
fives from east to west and by fours from north to south,
I have condensed this into 5 X 525 Cp plots, each of
20 ultimate plots as shown in Table II.

TABLE II
YIELDS OF COMBINATION PLOTS OF ROTHAMSTED WHEAT, 4 X 5 GROUPING.
ORIGINAL AREAS SEPARATED BY DOUBLE LINES IN MAP B

82.89 | 83.05 | 78.63 78.76 74.70

139.36 13241 | D2% 120.53 114.58
7i | $4.34 | 75.61 | 80.32 74.87
130.60 | 140.31 alo ear eat
79.80 | $4.70 | 74.94 81.50 77.34
13331 | ls | X 133.28 | p
84.36 | 82.42 7300 | 71.35 75.81
142.79 | it iwo o PiB ? 0
85.19 | 84.56 | 82.25 | 68.52 | 76.69
147.95 | 146.78 13842 | 12009 | 124.88

Summing the actual yields and the squares of yields
for the ultimate plots and the squares for the combination
plots, I find the following values:

For wheat grain

S(p) = 1974.32, S(p) = 7900.6790, N=500,
p=3.949, op? = .209600,
S (C2) = 156419.3106, m[n(n — 1)] = 25 X 20 X 19
= 9,500,
LS(0:2) — S(p?)]/m[n(n — 1) ] = 15.633540,
which leads to

15.633540 — (3.949)? _
R Boe = .186 + .029.

a 20
For wheat straw
S(p) = 3257.40, S(p?) = 21623.9802, N= 500,
p=6.515, 7 = .805341,
S'(C,? — 427479.9920, m[n(n—1)] 9500,
[S(Cp?2) — S(p?)]/m[n(n — 1) ] = 42.721685,

THE AMERICAN

2 3

4.

$

6

NATURALIST

7

8

7

/0

[Vou. XLIX

Il | 12

4/5 | £06
"68S | 7/7

S/a
777

320%
47

¢F8
6.08

PIS
734

40y
6.08

4.14
6.98

400
5.87

437| 402
6.75 | 6.10

H| £15
749| TH

464
7.80

403
6.34

3.74
6.63

y. 56
788

4.27
6.35

403
6.4/

450
6.50

3.97 | 4/7
6.07 | 6.¥3

429| #40
77/ | 7.35

4.69
750

IIT
6.17

446
6.78

4%
8./8

3.76
5.73

3.30
SIS

3.67
6.40

3.74 | 4.07
6./8 | 6.37

LOF | 40S
8.23| 7.87

£04
6.66

347
370

ERA
6.46

454
760

452
729

3,05
3.2

LST
§.4/

4£0/\3.34
3-77 | 5:60

427) 492
7.73 | 8.58

464
736

3.76
6.05

410
6.77

#40
7.4

4/7
720

3.67
A383

5.07
805

3.83 13.63
6.36 | 6.43

3.53 | £08
6.45 | 704

473
7.98

3.6/
587

3.66
6.1/5

437
7.36

3.84
6.28

4.26
76/

436
5.58

3.79 | 4.07
5 46| 6./0

3.50 | £23
5:87| 7.02

437
6.98

328
4.97

3.56
6.06

414
8.06

406
6.8/

¥32
737

4.86
7.5/

3.796 | 3.74
6.23} 6,38

338| 407
37 | 705

$66
7.48

372
5°78

3.84
6./0

yyy
750

3.40
S97

407
6.97

493
757

3.73 | 3.04
6./3 | 4.96

3.6/ | 467
6.0/| 764

449
6.95

ITS
$F

File.
6.83

46y
7.92

3.97
5°07

¥37
VEER

502
9.23

356|357
5:75 6.03

3.7/ | 427
6.27 7/7

yyt
6.9

4/3
7.3/

4.20
6.86

#66
739

36/
6.33

3:97
7.26

4 YF
ATS

3.36 |377
6.14 |6.26

SbF | 384
6.30 | 6.60

4S/
786

40/
718

42/
8.23

77
8.23

375°
ZH

4.17
754

437
7:73

4ST \#17
720\708

3.70 | 3.82
6/7 | 6.87

ZES
717

3.37
6.53

437
875

3.77 | E4¥/
6.33| 703

457
7.93

ie
706

447
8.53

4ps.
8.7%

444
G02

Kos
Hf
324

6-70

372
748
3.86

7.20

¥S6
7.73
4.77
7.67

4.4/0 |307
6.90 6/2
4.99\3.7/
7.82\737

409|4.39
7221773

4.3/
7.3/

#29
708

4-47
EIs

4.37
7.67

344
6.62

382

463

4.36|3.77

708

1.87

737 |6.33

H4 |Y?
320| 7.852

408
6.67

3.976
6S4

3.76
7.10

3.87
6.36

$.//
758

3728
6.87

4.03
7.16

4.07\3.82
7.03| 730

3.89 | 4246
7051699

452
6.93

3.78
6.72

3.54
6.46

Z7
eg

4/2
7.32

4/3
Zat

HUT
784

341 |355 |
5-26 | 6.70

3871723
682| 7.14

4.58
773

3/7
6.06

3.47
6.63

3.7/,
7.34

441
7S3

42/
Z#/

4of
REI

427|#06
7/17\700

4/2 7.39
138 LSS

3.92
6.70

484
8.85

394
6.75

438
743

414
732

3.76
70#

4.27
6.76

4S2\4/7
7.73 | 730

rt HSE | GT

7.03 | 8.06

5/6
£78

4 #6
754

44i
8/5

468
7.57

437
ZiT

4/5
fab

44/
7.76

£68 | 5-/3
£07 | 8.3/

422| 442

S09

7.63 |\8.4¢S

8.72

366
709

Fiz
7.72

406
706

3.97
753

387
736

446
6.9/

EME #52
6.87 | 8./7

A

Map B. Wheat Yields, Upper Figures Grains, Lower F

/

N

/

2.

7

N

/

‘
igures Straw,

No. 583] ON SUBSTRATUM HOMOGENEITY

13 | HIE \| 46177 | 18 119% | 20 || 24/ | 22123 | 24

Y¥S8NF72 |3.64||3.66 13.57 | 3.57 | #27 |3.72 || 3.36 |3.17 |2.77 | #23
7.23) 6.33 |S// 1| $96 | 5:12 | S05 | 6SY¥ | S¥7|| 476 | 475 | #853 | 6.08

408|3.97 |3.6/ || 3.82 |3.4# |3.72|426| £36 || 3.67 |3.53|3/F | £0713,
6.57 | 6.03 | 5:58 || 580 | 500 | 5:83 | 8.61 | 6# || $856 5-07 | S17 |57 |5

A^

3.73 | 459 | 3.64 || £07 |344 |3.53| #28 | ¥3/ || #33 |366 |3.57 |3.77
6.02 |743 |586 || 674| 556| 49/ | 65S | OF#|| 617| 6.15 |E H |628

40G |31? |375 || 4854| 3.77 |3.77 | #30 | 4/0 ||3.8/ 3.87 |3.32 |3.76

3.74 |414 13.70 || 3.72 |3.77 | 429 | #22 |3. 7# ||3.55 | 3.67 (3.857 | 5.76 |%
6.13 | 598 | Z67 || 6/4|833 |588| 6/5 |576 || £87 |5453 24 |362

| 619 | 6.56 |462 ||708 | 6.03 |577 |595 | £96 || 6.13 | £72 | #62 |5% |6.

3.72 |376 | 3.37 || #01 | 3.87 | 435 | #44 |358 || #20 | 3.74 | #24 | 3.75 | #
6.03 | IHF | E00 || 597 |8857] 6.09 | S88 |361 || 5.92 |887 |582 | 5.50

43313.77 |3.7/ || 45713.77 | #38 |38/ | 406 ||3.42 |3.05 | 344 | 2.78
6.77 |548 | 566 || 728 |603 | 624| 85467 | 64S || 54557] ESC | #48

IS

372 1323 |37/ | 476 13.83 |37/ |3S¢| 3.66 ||3.75 | 3.8F|3.76 |3.47
397 | 6.07 | 5-77 || O47 | 6.29 | £91852 | 5°78 || 572 | F06 324 | 5ST
40S |326 |378 ||473 | #24) 42/| 385 | #41 || 42/ | 3.63 |4/7 |347
6.82 | 6.35 | 5/2 || 8.64) 64S | 6.29 | 6/5. | 61S || 604| 58/ | S58 | 48/

3.37 (3.47 |3.09 || £20| £09 | 407| 407 | 3.9S"|| #08 | 403 | 3.97 | 2.84
6.25|5.78 | S¥7\| 6.47 | 6.16 | 6.18 |5847 | 6. || 700 | 3-72) 565 | $10

4093.491337 ||37# |3.4/ |386 | 436 | 454 || 424 | 408 3.87 13.47 |3.
7.28g |8 7/ | 644 || 8.63 | 578 | 6.74 | 739 | ZZE ||720 |684| 5-79 |387

3.99 | 3.14 | 486 || 436 13.57 |Z47|3.IF | F447 4H 13.97 | 407 |356

£.09\3.05|3.39||3.60| 4/3 |389 |367 | #54 || 441 | FSB | £02 13.73 | 4
774 |570 |886 || 6.27| 687| 623 | 6.20 | 733 || 664) 6.77| 635 | 567

7 13.| $05"| 639|| 226 | GU (£70 668 |784 || 695| 687| 3.80 |638 |6

A

3.56 13.47 13.64113.60|3./7 |380 (372 |3.91 13.35 | Z U | 4397 |3.#7
6.69? | 3°77 1636 || 584| 887 | 6.14 | CFF | 696 || 6.47 | 66| 6M |878

357 |3.43 | 3.73 ||3.37 |3.08 |348 |308 |368 || 3- |325 |36? | J43
GSS 838 | 8.58 || 6.42 |S L2 | 552 |520 | 6.60 || 627 | 6.37|S/8 |584

3.76 |3.47 (3.30 || 3.39 | 2.22 |343 |3.45-|3.26 ||3.42 | 3.6? |3.82 | 3.77

3.75 |327 |3,57/ || 345 | 7.05 |368 |352 |3.7/ || 3.87 |387 | #4/ | 3.08
6.31 | 6.21 | 599 || 40S -764| 584| 585| 6.7/ || 6/3 |750| 548 | 6.0/

SI4| 584 |870 || 580 | 49S |533 | T25 | 664 || £40 |593 | £72 | 627 | -

G49| 3.821360 || B/¥ | 2.73 | 507 13.66 |377 || 3.48 13.76 | 3.67 | 3.87
757) 6.37 |634 || 548| 477 |841 | SPF | O98 || OF | 6M |543 | 6IS

3

6
417 | 44 |3.54||3.0/ | 295°| 3.36 |385| 4/5 || 3.93 |3.9/ | #33 | #27
6.23 6.78 |558 || 5769 | 4.96 | O14 | CLS | 68E || 6.87) 6.09 | 694 | 678

370 | 428 |324||3.49 | 348 |3.47 |368 |3.36 || 3.7/ |3.5# |357 | 3.76
ó

6.80 | 6.97 | 895| SSB | 552 |582 | 6.76 |608 || 635 62/1466 | 6.36
4 A N N

W

on Rothamsted Acre.

whence

THE AMERICAN NATURALIST

_ 42.721685 — (6.515)?
805341

[ Vou. XLIX

- = 343 + .027.

Illustration 3. Influence of Substratum Heterogeneity
upon Yield of Grain and Nitrogen Content in Experi-
mental Plots of Wheat.

Table III is condensed from Map C of Montgomery”

832| S/F
2./3 | 2.06

525°
2.10

50/
176

TIF
2/0

486
2.03

¢ 83
2/7

¥S/
206

375:
2 08

440
2.43

432
2.03

441
2.1

4/15

4/0"

6/2 | 5/0
2./6 | 2.05

2.10 | 240

575| 480| ZSS

#60
206

530
2.00

538
2.09

“Ti
198

¥37
2/7

446
2.14

yay
4/3

#30
196

FLY
197

GOS
2.0/

436
19g

2.08

we L74 443
YoY IT

2.06

2.05

yaj
Tal

2M

446
21S

393
2.04

4/4
207

4al
197

422
195-

#23
2.03

380
196

TS3| THZ
2.03 | 2.00

STS
2d

A

HSA

Hr

AT

2.4/3

Fal

2.08

7v0

2.05

4g
2/8

#00
4/7

393
2.08

434
12g

39S
EKA

42/
2.06

#90

175

478732
2.12 \2.06

350
2097

ERA
4.05

¥80
2.0#

432
2.00

420
2.04

#51
2.9/

460
212

447

192

Pha
24

#43
2.02

YS
174

333

1 12.07

530
2.2/

576
2.05

gI
2.07

S66

Ass

5/7
2.04

530
2.43

S#6
197

YIS
197

474
188

434
1&6

AN

533
4.05

SIT
2.14

503
2.04

380
2.04

S/F
2.03

64392
2.25

740
486

64/
“a7

S06
2.0%

#95
2.02

$40

197 |.

550
2.08

S40
2.//

¥77
2.00

my
2/8

3/7
19S

606
2/4

756
LS a

656
200

s87
2/0

623
193

597
2.00

487
2.0/

343
ve?

628
2.24

729
243

6/6
42.05

620
2./6

724
4-43

67S
49S”

G47
2.09

7/0
496

7/1
198

$83
199

v/s
2.06

o/s
2.27

s35
2.43

467
2.4/

577
4.26

5 8/
Alt

648
2.2/

707
4.424

738
ad

oy
af

769
2.1/4

s87
1.93

528
2.07

SSE
4/8

575

S3/

M | 2.05

686.

4/0

656
497

7/6
4/9

739
2.02

730
2,08

597
2.00

644
4.0/

64H
2.06

S4/
1.08

673
194

676
2.04

7/2
4.34

666
2.4

G88
2.04

639
2,04

6/3
2.07

693
2./0

666
4.40

G43
203

674
2.16

66/
2.00

742
2/2

02
221

634
242

634
2.06

570
4.03

560
207

oes
2.1/4

376
207

3#
2.17

$57
40s

537
ass

590
247

58s
2./6

650
2.08

586
2.22

S33
LG

6/7
204

#96
227

527.

4.23

657
2.17

385
a. Uf

58S
2.09

875| 495
199 | 2.04

502
2.08

584
2/6

4.26

7/6 °

725
2/0

477
2/4

3/3
4.44

647
2,03

547
2.06

#88
2.06

Map C. Yields

15 Montgomery, E. G.,

of

164-177, 1912.

7

N

Grain in Grams
Wheat Plots. The e Entries a

N

ay perg of Nitrogen

e Yield
Percentage glastag Cont

A^

in M r GOAL!
n Grams of Grain, the Low
ent.

‘Variation in Yields and Method of Arranging
Plots to Secure Comparative Results,’? Ann. Rep. Neb. Agr. Exp. Sta., 25,

No. 583] ON SUBSTRATUM HOMOGENEITY 441

by combining adjoining plots 22. The following are
the numerical values.
For grains produced,
S(p) =123429, S(p?) = 70112319, N= 224,
p=—551.022, op? = 9375.826,
S (C?) = 277945243, m[n(n — 1)] = 642, |
[S (Cr?) —S(p?)]/m[n(n — 1)] = 309275.184,
whence
fipa = .603 + .029.
For percentage nitrogen, .
S(p) = 465.29, S(p?) = 968.3721, N = 224,
p=2.077187, op? = .008383,
S (C?) = 3868.5047, m[n(n—1)] = 672,
[S(Cp?2) — S(p?)]/m[n(n — 1) ] = 4.315673,
and
p43 == AL i 044,

TABLE III

COMBINATION PLOTS OF MONTGOMERY’S WHEAT, 2 X 2-FoLD GROUPING AS
INDICATED BY HEAVY LINES IN MAP

2,168 2,016 2,029 1,819 | 1,689 1,702 1,788

8.22 8.20 8.56 8.30 8.12 8.31

2,090 2,126 1,700 1,667 1,652 1,661 1,769
8.33 8.38 8.29 8.52 8.12 7.98 7.93
2,242 1,981 2,071 1,955 1,785 1,886 1,985
8.45 8.42 8.16 8.24 8.00 7.98 7.95
2,074 2,140 2,004 2,271 | 2,793 2,208 2,429
8.34 8. 8.26 8.37 8.27 8.09 7.88
2,043 1,928 2,406 2,280 2,628 2,802 2,682
8.23 .32 8.87 8.68 | 8.38 8.38
2,339 2,528 2,271 2,363 2,730 2,809 2,582
8.3 .0 8.39 8.04 8.36 8.21
2,573 2,456 2,470 2,286 2,498 2,524 2,540
1 8.27 8.46 8.43 8.32 8.88 8.45
2,450 2,322 2,591 2,097 2,326 2 2,200
8.26 8.54 8.37 8.67 8.32 7.98

442 THE AMERICAN NATURALIST [ Vou. XLIX

Illustration 4. Influence of Substratum Heterogeneity
upon the Yield of Experimental Plots of Timothy Hay.

I take as a final illustration of the application of the
criterion of substratum heterogeneity here proposed, the
plot data for timothy hay published by Holtermarck and
Larsen, loc. cit. By combining their plots into groups of
4 Table IV is secured,

S(p) = 4268.8, S(p?)=77968.50, N = 240,
p= 17.787, o,’ = 8.503,
S(C:2) = 309491.48, m[n(n—1)] =720,

whence
1p192 = -609 + .027.

TABLE IV
COMBINATION PLOTS 2 X 2, SHOWING YIELDS OF TIMOTHY HAY SECURED IN
| THE EXPERIMENT oF LARSON
The original field is not mapped here

a 87.4 99.0 | 78.5 65.8 67.2 63.3
76.4 72 | 750 73.1 67.7 59.7
76.9 65.2 64.2 89.7 72.1 64.3
65.1 54.1 66.4 98.9 83.3 64.3
57.9 64.7 61.1 88.6 72.2 64.8
73.0 55.6 62.2 75.6 82.8 71.1
71.7 “s © M 64.8 81.6 75.2
68.8 70.4 | 617 81.2 72.8 61.4
77.5 ne | 669 | s 73.9 | 685 _

B. Cases in which the Combination Plots Vary in Size

In the foregoing illustration the combination plots have
been of uniform size, i. e., have contained each the same
number of ultimate plots. It may be desirable or neces-
sary to have some of the combination plots smaller than
the others. Thus the wheat field of Mercer and Hall is

No. 583] ON SUBSTRATUM HOMOGENEITY 443

laid out in a 20 X 25 manner. This permits only 2 X 5,
45 or 5X5 combinations of the same size throughout.
Montgomery’s experiment comprises an area of 16 X 14
plots which may be combined in only 2 X 2 or 4 X 2 equal
areas suitable for calculation. In each of these cases
other groupings are desirable.

The formule are quite applicable to such cases: the
arithmetical routine is merely a little longer. The for-
mula is again

{[S(C,*) — S(p*)]/SIn(n — 1)]} -

Tryp, cee

but p and cp are obtained by a (n —1)-fold weighting of
the plots,1® where n is the number of ultimate plots in the
combination plot to which any p may be assigned, i. e.,

p=S[(n—1)p]/S[n(n—1)],

Me = Da (Aeru,
T Sina- 1)] ~ Sha- 1))

The point may be illustrated in detail on the wheat data
of Mercer and Hall. I adopt a combination by twos from
north to south, 7. e., arrange the data in 10 rows of com-
bination plots instead of 20 rows of ultimate plots. From
east to west there are 25 rows of ultimate plots; these can
be only reduced to a 2 X 2-fold grouping for the first 22
rows. The lines of division are indicated on Map B by
marginal arrows.

Row 23-25 must be thrown into combination plots each
of 6 units. The possible permutations within a combina-
tion plot are 1/2 n(n — 1), but since the surfaces are ren-
dered symmetrical, the total permutations for the whole
field is S[n(m—1)]. There are only two sizes of combi-
nation plots, of which 110 have 4 and 10 have 6 ultimate
plots each. Thus the weighted population N is

16 That is, each ultimate plot is multiplied by the number less one of the
plots in the combination plot to which it is assigne

444 THE AMERICAN NATURALIST (Vou. XLIX

S[n(m—1)] = (110 x 4 x 3) + (10 x6 X-5)=1620;
In the calculation of the weighted means and standard
deviations each entry, and the square of each entry, in the
first 22 rows must be weighted in an (n — 1)-fold=3-fold
manner, while those for the last three rows must be
weighted in a 5-fold manner.

The numerical values are:

For grain,

Sl(n— dae = 6378.72, S[(m—1) p*] = 254524154,
==3.937, dp? = 207610,

S o = 33129.7080, Ə (7) = 1900.6790,
whence

Tssa — .354 + .026.

Note that S(p?) is constant for all groupings.
For straw,

S[(n — 1)p] = 10474.52, S[(n — 1)p?] = 69042.7194,
p = 0.466, op*== 813000,
S (C2) = 89985.8976, S(p?) = 21623.9802,
whence
Taro ™= ALD + 028.

Weighting has not materially changed the physical con-
stants from the values given under illustration 2 above.
The reasons for the conspicuous differences in the corre-
lations will be taken up presently.

Montgomery’s wheat data have been grouped into 2 X 2-
fold combination plots in the illustration above. If we
again combine the entries of Table III by twos, beginning
at the upper left-hand corner, we have 12 combination
plots each 4 X 4, or of 16 ultimate plots, and 4 combina-

17 Since each individual ultimate plot is compared once as a first (or as
a second) number of a pair with every plot classed with it, the weighting

of the ae plots for means and standard deviations is an (n—1)-
fold o

No. 583] ON SUBSTRATUM HOMOGENEITY 445

tion plots each of 2 X 4=8 ultimate plots. The method
of dividing up the field is indicated by the marginal ar-
rows on Map C.

S[n(n —1)] = (12 X 16 15) + (4&8 X 7) = 3104.
For grain,
S[(n — 1)p] = 1707635, S[(n — 1)p?] —9683408.57
p=550.140, op? = 9311.307,
S (C?) — 1023184887, S(p?) = 70112319,
whence
rae == 412 036:

For nitrogen,

S[(n —1)p] = 6458.63, S[(n — 1)p?] = 13464.6031,
p=2.080744, op? = .008327,
S (C2) = 14409.6095, S(p?) —968.3721,
and
1192 = .096 + .045.

Again the weighted means and standard deviations do
not differ widely from those used above. The differ-
ences in the correlations will be discussed below.

In concluding this section it may be pointed out that
all of the foregoing values are surprisingly high. They
indicate clearly that the irregularities of an apparently
uniform field may influence profoundly the yield of a
series of experimental plots. They also bring out an-
other interesting point. In the three cases in which two
different characters were measured on the same species
they show very different susceptibilities to environmental
influence. Thus, for example, the correlation of man-
gold roots is r = .346 + .042 as compared with r==.466 +
-037 for leaves. For grain on the Rothamsted field with a
4 X 5-fold grouping the correlation is r=.186 + .029 as
compared with r= .343 + .027 for straw. For Montgom-
ery’s data for yield and composition the differences are

446 THE AMERICAN NATURALIST [Vou. XLIX

even more conspicuous. The correlation for per cent. ni-
trogen is r=.115-++.044 as compared with r= .603 + .029
for weight of grain produced.

This point will not be discussed in greater detail here,
since the problem of the relative susceptibility of various
characteristics of the individual to environmental influ-
ence has been the subject of experimental and statistical
studies which have been under way for several years and
will probably eventually be published.

III. ON THE Nature OF THE REGRESSION or ASSOCIATED
PuLots

The correlation coefficient is strictly valid as a measure
of interdependence only when regression is linear, i. e.,
when the means of the second variable associated with
successive grades of the first lie in a sensibly straight line.
The equation for the regression straight line

a
poe Op para Op,

ps Pes (P: — pps >) F pipe pı
Tp, Tp,

for the second on the first ultimate plot of the same com- —
bination plot reduces to

Pp — 18) +10,
when symmetrical tables in which p, = Po, op; = Fpa are
used.

The testing of the linearity of regression in any indi-
vidual case is rendered somewhat difficult by the necessity

TABLE V
YIELD OF GRAIN IN ROTHAMSTED WHEAT EXPERIMENT
Mean Yield Mean Yield
Yield of First | Weighted pone Yield of First |` Weighted
Plot Frequency ot —— Plot a Frequency " —
2.75-2.99 133 3.76 4.00-4.24 1786 3.99
3 3.24 75 3.78 4.25-4.49 1444 4.07
3.25-3.49 1026 3.81 4.50-4.7 03 04
1634 3.89 4.75—4, 247 4.05
3.75-3.99 1919 3.93 5.00-5.24 BIG

No. 583] ON SUBSTRATUM HOMOGENEITY 447

of actually forming a correlation table from which to com-
pute the means of arrays. The labor is greatly lessened
by the use of some such scheme as that described for the
formation of condensed correlation tables.1®

TABLE VI
YIELD OF STRAW IN ROTHAMSTED WHEAT EXPERIMENT
Yield of First| Weighted | jp een wield | vield of First | Weighted | paan Yield,

Plot Frequency Plots Plot {Frequency Plots
4.00-4.24 19 6.11 6.50-6.74 608 6.56
4.25-4.49 19 5.68 6.75-6.99 817 6.69
4.50-4.7 133 6.08 re 779 6.86
4.75-4.9 171 6.07 7.25-7.49 65 6.84
5.00-5.24 304 6.19 7.50-7.74 627 04
5.25-5.49 418 6.13 7.75-7.99 323 6.96
5.50-5.74 6.18 8.00-8.24 247 7.14
5.75-5.99 1121 6.20 8.25-8.49 09
6.00-6.24 1273 6.31 8.50-8.74 152 6.75
6.25-6.49 969 6.38 8.75-8.99 76 7.28

eS
~
N
™~
~
~
5

20 | 27 |22 |23 |24 |25 |26 |27 | 28 |29 30 |37 [32 |33 [a+ |35 |36

w
%

Bi
Ioi

Wha T T U T U T T T T T T
oR BoE Boe TE a ee a
= & a
= $ eee.
-20 z
L L L L L L L L 1
Figure A Figure 2

Figs. 1 AND 2. Mean Yields of Grain and Straw on Ultimate Plots Asso-
ciated in the Same Combination Plots of a Given Yield. Rothamsted Wheat.
Empirical Means and Fitted Straight Line. Units are Quarters of a Pound.

18 Harris, J. Arthur, ‘‘On the Formation of Condensed Correlation Tables
when the Number of Combinations is Large,’ AMER. NAT., 46, 477-486,
1912,

448 THE AMERICAN NATURALIST (Vou. XLE

For the 5 X 4 grouping of the 500 wheat plot of Mercer
and Hall I find the values given in Tables V—VI.

For the regression of the second on the first plot the
equations are:

For grain, g, g =3.214 + 186 9;.
For straw, s, Sy = 4.280 + .343 S1-

Figs. 1 and 2 exhibit the usual irregularities of sam-
pling in the means, but show no certain departure from
linearity.

TABLE VII

YIELD OF GRAIN IN MONTGOMERY’S WHEAT EXPERIMENT

rield of First | Weightea | MeO Yield | vita ot rirst | Weighted | SMU ee
hes Plot g | Fisaneine | at gh $ Plot pet Pes eney_ | * = reas
325-374 | 9 | 516.88 575-624 liL | 579.82
375-42 | 63 440.22 625-674 90 | 616.21
425-474 | 93 471.23 675-724 45 656.37
475-524 | 108 | 540.24 725-774 30 | 628.80
525-574 | 120 l 548.24 775-824 3 | 574.00

350 | #00 | +50. | 500 sso | 600 | 650 | wo | 750 | 800

Figure s

Fic. 3. Grain Yields in Nebraska Wheat. See Figs. 1-2 for Explanation.

No. 583] ON SUBSTRATUM HOMOGENEITY 449

Table VII gives the first plot character, weighted fre-
quencies and empirical means for associated plots for
2 X 2-fold combinations from Montgomery’s grain yield
data in wheat.!®

The equation is

For grain, g, 9, = 218.993 + .603 g4.
The graph figures indicate sensible linearity.

IV. INFLUENCE or NUMBER or ULTIMATE PLOTS COMBINED

If an experimental field exhibit irregularities of condi-
tions which influence in a measurable degree the yield of

TABLE VIII
5 X 2-FoLD COMBINATION OF PLOTS OF ROTHAMSTED WHEAT
divisions of the field are indicated by the double vertical lines and
the arrows along the right margin in map

41.11 42.51 40.32 38.53 36.65
68.19 66.59 62.22 59.52 54.45
41.78 40.54 38.31 40.23 38.05
71.17 65.82 60.62 61.01 60.13
40.35 41.92 37.77 40.01 39.48
69.10 70.37 60.65 58.10 58.00
37.80 42.42 37.84 40.31 35.39
61.50 69.94 59.46 61.17 54.21
40.42 42.03 36.69 41.84 38.83
65.95 71.09 59.97 64.09 56.99
39.38 42.67 38.25 39.66 38.51
67.36 78.49 65.30 69.19 63.10
42.77 42.17 38.07 38.05 40.22
71.95 75.20 66.92 64.05 63.45
41.59 40.25 | 35.53 33.30 35.59
70.84 72.24 | 64.88 57.13 58.57
41.75 41.44 40.12 34.00 38.13
71.84 71.92 67.99 60.55 62.36
43.44 43.12 42.13 34.52 38.53
ae tT 70.43 59.54 62.52

19 Because of the many differences in the two experiments it is inadvisable
to attempt drawing the regressions lines in a strictly comparable form.

450 THE AMERICAN NATURALIST [ Vou. XLIX

neighboring small experimental plots, this heterogeneity
should become apparently less when expressed on a scale
of correlation between plots as the number of ultimate
plots combined increases. The reason for this condition
is quite simple. If the irregularities are very local in
nature they will influence in the same direction the yield
of only a very few neighboring plots. If too many ulti-
mate plots be combined the correlation will tend to vanish
because of the increased frequency of association of un-
like conditions due to the fact that the combination plots
have been made so large that they themselves have become
heterogeneous.

That these conditions have been observed in actual ex-
perimentation is shown by the following constants based
on different groupings of the data used above. .

Consider first the Rothamsted wheat. For a 4X5
grouping of the plots the results were found to be

For grain, 1 9199 = 186 + .029,
For straw, Tpyp9 == 043 + .027.

If the plots be grouped by fives from east to west and
by twos from north to south, Table VIII is obtained. The
values S(p), p and op are the same as in the preceding
case,

m[n(n —1)] =50 X 10 x 9 = 4500.
For grain, S$(C,?) =78265.2822, rpp = .214 + .029.

For straw, S(C>?) = 213939.8774, ryp: =.365 + .026.

If the combination plots be made even smaller by group-
ing in a 2 X 2-fold manner for all but the last three north
and south rows, where a 2 X 3-fold combination must be
adopted, the results are, as illustrated above,

For grain, Tpi = -394 + .026,
For straw, 9199 = A19 + .023.
For Montgomery’s wheat data the results for a 4 X 4

fold grouping (in as far as the nature of the records will
permit) have been shown to be

No. 583] ON SUBSTRATUM HOMOGENEITY 451

For grain, fpa = 472 + .035,
For nitrogen, p19 = -096 + .045,

as compared with the following values for a 2 Xx 2-fold
grouping

For grain, 492 = -603 + .029,

For nitrogen, Y'pypq = -115 + .044.

Finally consider the constants deduced from the hay
yields published by Holtermark and Larsen.

For a 2 x 2-fold grouping, tars = .609 + .027,
For a 2x 4-fold grouping, — fma = .471 + .034,
For a 2 X 8-fold grouping, fpina = .278 + .040.

Thus for every species of plant and every character con-
sidered the correlation between associated ultimate plots
decreases as the number of plots grouped increases.”°

TABLES IX AND X

2 X 4-Forp AND 2 X 8-FoLD COMBINATION OF THE DATA FOR PLOT YIELD
IN TIMOTHY HAY, TABLES DERIVED FROM TABLE IV

163.8 169.2 153.5 138.9 134.9 123.0 p
305.8 288.5 284.1 352.8 310.4 264.5

20 Of course, the same effect would be produced if comparisons were
drawn between tests for substratum heterogeneity on fields Sree
in every regard except for the size of the ultimate plots. Possibly, t
plains in part, at least, the striking differences in the correlations for grain
yield found from the records of Montgomery and of Mercer and Hall.

The Rothamsted plots were 1/500th acre in area or 87.12 square feet.
Montgomery’s plots were 5.5 X 5.5==30.25 square feet, or only about 1/3
of the area of the Rothamsted plots.

452 THE AMERICAN NATURALIST [ Von. XLIX

V. RECAPITULATION AND DISCUSSION

If the methodical production of new varieties of animals
and plants to be made possible by the laws discovered
in experimental breeding is to be of material practical
value, more attention must be given to the development
of a standardized scientific system of variety testing.
From the practical standpoint, nothing is to be gained by
the formation of varieties of plants differing in discern-
ible features of any kind unless some of these varieties
ean by rigorous scientific tests be shown to be of superior
economic value.

It is equally true that if tests of fertilizers or of dif-
ferent methods of irrigation carried out on an experi-
mental scale are to have any real value as a guide to a com-
mercial practise, the differences in the experimental re-
sults must certainly be significant in comparison with
their probable errors.

The problem of plot tests has several different phases,
all of which must ultimately receive careful investigation.
The purpose of this paper has been to consider one of the
problems only. To what extent do the irregularities of
an apparently homogeneous field selected for comparative
plot tests influence the yield of the plots?

The question has been far too generally neglected,
although indispensable to trustworthy results. It is ob-
viously idle to conclude from a given experiment that va-
riety A yields higher than variety B, or that fertilizer X
is more effective than fertilizer Y, unless the differences
found are greater than those which might be expected
from differences in the productive capacity of the plots
of soils upon which they were grown.

The first problem has been to secure some suitable
mathematical criterion of substratum homogeneity (or
heterogeneity). Such a criterion should be expressed on
a relative scale ranging from 0 to 1, in order that com-

The 2 X 2-fold grouping of Montgomery’s plots gives a correlation of
.603 + .029 as compared with r = .354 + .026 for as nearly a perfect 2 X 2-
fold grouping as the Rothamsted records permit.

No. 583] ON SUBSTRATUM HOMOGENEITY 453

parisons from field to field, variety to variety or character
to character, may be directly made. It should also, if pos-
sible, offer no difficulties of calculation.

The criterion proposed is the coefficient of correlation
between neighboring plots of the field. An exceedingly
simple formula for the calculation of such coefficients has
been deduced.

The method of application of this coefficient is here il-
lustrated by four distinct series of experimental data.

The remarkable thing about the results of these tests
is that in every case the coefficient of correlation has the
positive sign and that in some instances it is of even more
than a medium value. In short, in every one of these ex-
perimental series the irregularities of the substratum have
been sufficient to influence, and often profoundly, the ex-
perimental results.

It might be objected that by chance, or otherwise, the
illustrations are not typical of what ordinarily occurs in
plot cultures. But they have been purposely drawn from
the writings of those who are recognized authorities in
agricultural experimentation, and who have given their
assurance of the suitability of the fields upon which the
tests were made.

For example, Mercer and Hall state the purpose of
their research to be, ‘‘to estimate the variations in the
yield of various sized plots of ordinary field crops which
had been subjected to no special treatment and appealed
to the eye sensibly uniform.’’ Their mangolds ‘‘looked
a uniform and fairly heavy crop for the season and soil,’’
while in their wheat field ‘‘a very uniform area was se-
lected, one acre of which was harvested in separate plots,
each one five hundredth of an acre in area.” The data
of Larsen were drawn from an experiment ‘‘auf einer
dem Auge sehr gleichmissig erscheinenden, 3 Jahre alten
Timotheegraswiese.’’ Montgomery’s data were secured
from a plot of land only 77 X 88 feet in size, which had
been sown continuously to Turkey Red wheat for three

454 THE AMERICAN NATURALIST [Von XLIX

years, ‘‘and was of about average uniformity and fer-
tility.’’

Nothing could, it seems to me, emphasize more emphat-
ically the need of a scientific criterion for substratum homo-
geneity than the facts that correlations between the yields
of adjacent plots ranging from r= .115 to r= .609 can be
deduced from the data of fields which have passed the
trained eyes of agricultural experimenters as satisfac-
torily uniform.

December 12, 1914

SHORTER ARTICLES AND DISCUSSION

A NOTE ON THE GONADS OF GYNANDROMORPHS OF
DROSOPHILA AMPELOPHILA

Five gynandromorphs of Drosophila amelophila were sec-
tioned and their gonads studied in order to determine whether
the gonads corresponded to the secondary sex characters ex-
pressed by the somatoplasm. The specimens were either lateral
or fore and aft gynandromorphs.

I. This gynandromorph arose from a cross between a white
eyed fly and a fly of the wild type. On one side of the body the
eye was red, the wing long, the sex comb lacking, and the ab-
domen characteristically female. The other side had a white
eye, short wing, sex comb and male abdomen. The external
genitalia were abnormal.

The fly would not mate, not only because of the abnormality
of the genitalia, but because its mating instincts were indifferent.
It was courted by males but, in turn, it itself did not court
females.

Since the fly was externally a bilateral gynandromorph one
would expect to find that the gonads on one side were male and
on the other side female. This, however, was not the case. The
gonads on both sides were male and the testes were filled with
ripe spermatozoa.

II. This fly arose from a cross of cherry club vermilion with
the wild type. The left side had a cherry eye, sex comb, long
wing and an abdomen of the female type. The sex comb is
characteristic of the male and the long wing of the female. The
right side had a red eye, no sex comb, short wing, and abdomen
of the male type. The absence of the sex comb is characteristic
of the female while the short wing and dark abdomen on this
side were male. This is lateral and, at the same time, a fore and
aft gynandromorph. The left side was male anteriorly and
female posteriorly while the right side was female anteriorly
and male posteriorly. The gonads were male but immature.
No ripe spermatozoa were seen.

III. The origin of this fly was the same as the last. Both

456 THE AMERICAN NATURALIST [ Vou. XLIX

eyes were red, sex combs lacking, left wing long, and the ab-
domen characteristically male. The external genitalia were
apparently half male and half female. This is not a fore and
aft gynandromorph but a lateral one in which the parts involved
are restricted to the abdomen and the posterior part of the
thorax.

The fiy was courted assiduously by males but it would not
mate. The gonads on both sides were female and ripe eggs were
present. It is probably true that the eggs could not be deposited ©
on account of some defect in the oviducts.

IV. The origin of this fly was the same as the last two, i. e., it
came from a cross of cherry club vermilion with the wild type.
The eyes were red, sex combs lacking ; the wings were of the same
length; the abdomen was divided into a female and a male side
and the external genitalia were apparently half female and half
male. Anteriorly the fly was female, and posteriorly it was half
male and half female.

A male courted this gynandromorph as long as the male re-
mained in front of it. When the male with one wing vibrating
made a half circle to the tip of the abdomen, it immediately
dropped its wing and turned and ran. Sections showed mature
spermatozoa in both testes.

. This gynandromorph arose from the cross of an abnormal
form, a possible mutant, with the wild type. The eyes were
red, but on one side there was a sex comb and a short wing,
while on the other side the sex comb was lacking and the wing
was long. The abdomen was characteristically female. The
gonads were of the female type on both sides.

The conclusion, if one is justified in drawing a conclusion
from so few data, is that the gonads of lateral gynandromorphs
do not follow the separation of the somatic cells into a male and
a female side, but are always the same on both sides, either male
or female. Since the cells of an early embryo must be either
male or female producing, we can understand why the gonads
of a gynandromorph should be alike on both sides, regardless of
the somatic condition, if we suppose that the gonads are derived
from a single cell of the embryo.

F. N. DUNCAN

COLUMBIA UNIVERSITY

VOL. XLIX, NO. 584 AUGUST, 1915

THE
AMERICAN
NATURALIST

A MONTHLY JOURNAL
Devoted to the Advancement of the Biological Sciences with
Special Reference to the Factors of Evolution

CONTENTS
Page
I. The Chromosome View of Heredity and its Meaning to Plant Breeders. Pro-
am a a a me a ea ae
- — -495

Regeneration Posteriorly in Enchytræus albidus. H. R. HUNT

The Origin of Bilaterality in Vertebrates. Professor A. C. EYCLESHEIMER - 504

Shorter Articles and Discussion: The Tortoiseshell Cat. Dr. PHtngas W.
ee a a ke a o

THE SOIENCE PRESS

LANCASTER, PA. GARRISON, N. Y.
NEW YORK: SUB-STATION 84

The American Naturalist

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

sent to the rage of THE AMERICAN NATURALIST, Garrison-on-Hudson, New York.

Short articles containing summaries of research work bearing on the

În publicati of organic evolution are especially welcome, and will be given preference
n pu

ry ndrea reprints of oe are supplied to authors free of charge.

F arte reprints lek be supplie

THE SCIENCE PRESS

_ NEW YORK: iesi Station 84

tt,

Lancaster, Pa. Garrison, N. Y.

elasna

April 2, 1908, Post Office at Lancaster, Pa., under the Act of
Congress va? ate 8, 1879

JAPAN APRN HISTORY ee

FOR S
ARCTIC, ICELAND and GREENLAND
DS’ SKINS
Well Prepared Low Prices
Particulars of
. DINESEN, Bird Collector
Husavik, North Iceland, Via Leidie, England

Perfeot Condition and Lowest Prio
Specialty: a Skins, Oology, asad: Marine

Animals and others. Catalogue free. Correspond-
ence solicit
T. FUKAI, Naturalist,
Konosu, Saitama, Japan

WANTED

Marine Biological Laboratory

|
BIRDS OF AMERICA by J. J. Audubon, 7 volumes, | Woods Hole, Mass.

please report cash price, stating condition, binding and dates INVESTIG ATION Facilities for research in and Bet
of volumes. Entire Year pomp enka: = logy a labors ml
ie for not pipe
k G HARRIS, tories, $ ne prin tables ar sighs
Box 2244 Boston, Massachusetts able ne ee ae tea
ay members of the oe The fee
for such a table is
ee
INSTRUCTION —— of laboratory in spente aa
logs Physi-
July—August brate Zoology, Embryology, ay
The University of Chicago] z oley, Morpaiogy and fazonom
ers instruction during the Sum- logy of ne Strand and Bos
ae Coren on the same basis as Vegetation. Each course es
during p. s ee aes of the the full time of the student. Pras
The undergraduat Philosophical Aspects of Biolo
e colleges, the 0:30) s
graduate ools, and Phage Lage s also of
sional schools e courses DRT
Arts, Fiteratere, ien SUPPLY ste, ras
bog La roe Mach Administ tra DEPARTMENT pan y het ‘al of all of
tion, end Divinity. T foarte cate Open the Entire Year anger and of Alm, Fungi, A
: pi daki ystaf = hewmen = bg ioc pay or ie ‘i museum.
WwW. f: 2 3
f in the summer by appointment of Living material furnished Bnp
profess i as ordered, Price lista of Zor
met l and Botanical materini
sent on application pistes w and sl
mmer Quarter, 1915 is desired. - For price l sta a
on Term. June 21~July 28 informat regarding
2d Term July ear cana
Detailed ann iouncements will
gent upon application, - GEO. M. GRAY, Curator, Woods Hole, Mass
tion to
i The University of Chicago The annual announcement will be sent on Seok
Mitchell Tower Chicago, Illinois The Director iological Laboratory, Woods Hole

THE
AMERICAN NATURALIST

VoL. XLIX. August, 1915 No. 584

THE CHROMOSOME VIEW OF HEREDITY AND
ITS MEANING TO PLANT BREEDERS!

E. M. EAST

BUSSEY Institution, HARVARD. UNIVERSITY

DEFINITE advice as to practical procedure must be based
on a firm foundation of fact if the leaders in the applied
science are to retain any confidence in those who lay the
first stones in the pure science. At the same time, if it is
clearly understood that science only approximates truth,
that so-called ‘‘established laws’’ are only highly prob-
able and never absolute, it can hardly be said to be unwise
if an inventory of fact is taken at any time. The hand-
writing on the wall is never finished; some words are dim
and the erasures and omissions are many, but that is no
reason why one should not try to read it and to see what
it directs if he has translated aright.

This preliminary justification of the title of this article
is made because our present stock of facts regarding
heredity points clearly to the chromosomes as vital parts
of the mechanism, and I wish to emphasize some impor-
tant practical deductions in case this position continues to
become more firmly established.

A just and complete dissertation upon the rôle of the

1 This paper is based upon two lectures delivered at Harvard University
in 1914. I hope that any cytologists who may have their attention called
to it will overlook the repetition of some well-known facts in the first few
pages, as it is intended to be merely a general statement of a particular
point of view with certain deductions that follow if it be accepted. I wish
to thank Doctors O. E. White, T. H. Morgan and R, Goldschmidt for their
kindness in giving me many suggestions, but in justice to them I should state
that they are not responsible for the conclusions drawn.

457

458 THE AMERICAN NATURALIST [Vor. XLIX

chromosomes in heredity not only would fill many pages,
but would expose numerous gaps in our present knowl-
edge, gaps that leave several important questions in the
balance. We shall assume frankly therefore that the
chromosomes are the bearers of the determiners of prac-
tically all of the hereditary characters that have been in-
vestigated by pedigree culture methods, acknowledging
freely our ignorance on many points, but maintaining that
while no facts have been discovered which offer insur-
mountable arguments against the viewpoint taken, the
following logical sequence of truths discovered at various
times and by different methods of research make a pretty
sound case upon which to base our practical conclusions.

RELATIVE Importance OF NUCLEUS AND CyTOPLASM

There are several reasons for believing that of the two
parts of the cell, the nucleus and the cytoplasm, the former
plays the greater rôle in heredity.

In general it is believed that the two parents contribute
equally in the production of offspring—that the male and
female contribution of potential characters is practically
the same. If there were a difference it would be shown
by divergent results in reciprocal crosses, but the investi-
gations following Mendel’s method make it probable that
with the exception of sex and sex-linked characters, the
results of reciprocal crosses are generally alike. This
being true, it would appear that the principal basis of
inheritance must be sought elsewhere than in the cyto-
plasm, for in most observed cases the sperm is very much
smaller than the egg, and this difference is largely a dif-
ference in the amount of cytoplasm each carries. Is one
not to look for some significance in this disparity in size?
Strasburger, as well as other botanists, has even gone so
far as to declare the male generative cell in certain angio-
sperms to be simply a naked nucleus that slips out of its
cytoplasmic coat into the embryo sac, leaving the dis-
carded coat behind, and that stimuli proceeding from the
nucleus control the assimilation of food in the cell and
determine even the character of the cytoplasm itself.

No. 584] HEREDITY AND ITS MEANING 459

This belief may be too radical. The machine must have
, all of its parts to do proper work; and it may be, as Conk-
lin suggests, that such characters as polarity, symmetry
and localization of organ bases in the egg have their chief
seat in the cytoplasm. This is only a possibility and not
a fact, however, for one must admit that cytological inves-
tigation has not disclosed the presence of a material basis
of heredity in the cytoplasm, though he may not be con-
vinced that it is unimportant. Does the same statement
hold for the nucleus?

The nuclear cavity contains four substances as they are
ordinarily described in connection with morphological in-
vestigations. These are nuclear sap, linin, nucleolar ma-
terial and chromatin.

Nuclear sap probably belongs as much to the cytoplasm
as to the nucleus, and we know nothing as to its possible
significance and importance within the nucleus.

Linin by some investigators is regarded as very similar
to chromatin. Others (Strasburger) consider it to be
the framework of the chromosomes, and the only real sub-
stance within the nuclear cavity that is continuous from
generation to generation. It is a thread-like material
staining lighter than chromatin upon which the chromo-
somes appear to be strung in the early prophases of nu-
clear division.

Nucleolar substance, though it stains in a different
manner from chromatin, is considered by many to be
chromatin-like in its nature. It is the substance of which
the nucleoli are composed; but as these bodies become
vacuolated and finally disappear during nuclear division,
one is led to believe with Strasburger that they are tem-
porary storehouses of some necessary food material.

Chromatin, however, as the material of which the
chromosomes are composed, plays such a peculiar part in
the activities of the cell, that hypotheses as to the méan-
ing of its behavior are certainly more than shrewd guesses,
as will be seen.

The chromosomes may be described as morphological

460 THE AMERICAN NATURALIST [ Vou. XLIX

elements, of various shapes and sizes that are found
within the nucleus; they are especially demonstrable as
deeply staining bodies, definite in number for each cell at
the period of division. In many cases in both plants and
animals they have been found to be made up of small
particles, the chromomeres, and various investigators
have expressed the belief that these, too, are definite in
number and play an important part in the larger collective
entity, the chromosome.

Almost from their discovery, the chromosomes have .
had an especially important part assigned to them in the
drama of heredity because of the previous philosophical
deductions of Weismann. Weismann reasoned that if
there were no reduction of heritable substance in the life
cycle of an organism, it would pile up indefinitely because
of the nuclear fusion at fertilization. He, therefore, pre-
dicted the discovery of some mechanism by which the
character conserving substance would be divided. A few
years later his prediction was verified in its important
details by actual observation of the chromosome reduc-
tion in the formation of germ cells in Ascaris. From this
discovery and from the facts that a specific number was
found for the cells of each species, that all the cells of an
individual appeared to possess the same number (except
when they were halved at gametogenesis), that they were
apparently permanent organs, that they were longitudi-
nally halved in division so as to give each daughter cell
the same number as well as an exact half of each chromo-
some possessed by the mother cell, investigators were
early tempted to place upon chromosomes the whole
burden of inheritance.

Our observations regarding chromosomes and the re-
duction divisions in plants now rest on a basis of cyto-
logical investigation of over 250 species, representing
over 150 genera and divided among the four great groups
of this kingdom. Montgomery’s 1906 list of chromosome
numbers in animals represents investigations on 185 spe-
cies, comprised in about 170 genera, distributed among

No. 584] HEREDITY AND ITS MEANING 461

nearly all the phyla of the animal kingdom. Sex chro-
mosome studies have undoubtedly increased these figures
for the animal kingdom to date, by hundreds of species.

Variation in chromosome number among the cells of an
individual plant or animal is a recognized fact among
cytologists, but this variation is not regarded as of par-
ticular significance, as commonly it is held to exist only
among old cells, cells highly specialized, or, at any rate,
cells which will never have anything in common with re-
production. To quote from Strasburger,
the number of chromosomes in the nuclei of the somatice cells of both
the sexual and the asexual generations have been found to vary. But so
far as my experience goes, these observations are always to be observed
in the nuclei of cells which are no longer embryonic, like those in an
embryo or growing point, but which, on the contrary, are to some ex-
tent histologically specialized and are not destined eventually to give
rise to reproductive cells. The determinate number is still more fre-
quently departed from in nuclei which are definitely excluded from the
sphere of reproduction.

In the reproductive cells, chromosome division is, on
the other hand, very exact, and the numbers found, almost
invariable, with one exception. This exception is the so-
called accessory chromosome or chromosomes, that ap-
pear to be coupled with sex differentiation. And the
very fact that such accessory chromosomes do exist and
by their presence or absence parallel sex distribution,
forms one of the most unanswerable arguments in favor
of the chromosomes being the chief bearers of character
determinants.

MORPHOLOGICAL InpIvIDUALITY OF THE CHROMOSOMES

The next topic to consider is whether there is sufficient
evidence to support the idea that these bodies—the chro-
mosomes—are morphological entities persisting from one
cell generation to another.

Prochromosomes are deeply staining bodies found in
the resting cell nuclei of plants, which probably corre-
spond in number, but not in size, to the chromosomes
which are found in the dividing nuclei. These bodies are

462 THE AMERICAN NATURALIST [ Vor. XLIX

thought to represent the resting nuclear condition of the
chromosomes. Prochromosomes have been found in at
least sixty species of plants, and various structures com-
parable to them in many others. These investigations
favor the thought that the chromosomes are persistent
morphological entities; nevertheless they are not suffi-
cient to establish the matter if there were no other data
at hand.

There is a series of facts, however, which is more con-
vincing. We are told that in addition to each species of
animal or plant having in the larger part of its cells a spe-
cific number of chromosomes, there is a constant reap-
pearance of the different shapes and sizes of these chro-
mosomes in the same positions relative to one another
during cell division after cell division.

Strasburger says: ‘‘The observation of such a series
of stages of nuclear division as can be obtained by the
laying open of embryo sacs in which development of
endosperm tissue is commencing, makes it difficult to re-
sist the impression that it is always the same chromo-
somes which make their appearance over and over again
in the repeated divisions. In the prophase, the chromo-
somes are seen to appear in precisely the same position
that they occupied in the preceding anaphase, and if the
picture of the anaphase were proportionally enlarged, it
would exactly correspond to that of the succeeding pro-
phase.’’

The facts from which these general conclusions have
been drawn can not be denied. Baltzer found odd-shaped
chromosomes of similar shape in many maturing eggs of
sea urchins. Boveri, Montgomery and later Schaffner
pointed out a constant difference in the form and the size
relations of the two chromosomes of Ascaris megalo-
cephala univalens. Sutton thought he could recognize
each individual chromosome in eleven consecutive cell
generations of the maturing germ cells of the lubber
grasshopper Brachystola magna. The so-called sex chro-
mosome which has been found in so many insects and

No. 584] HEREDITY AND ITS MEANING 463

other animals, is a clear case of constancy in appearance.
In plants the same phenomenon has been observed. Ro-
senberg investigated the pollen mother cells of Crepis
virens and in certain stages in division invariably found
two long, two intermediate and two very short chromo-
somes. Division figures in the somatic cells showed the
same differentiation, and in an examination of the nuclei
of the pollen grain he found only one chromosome of each
kind present. Such other species of this genus as have
been investigated also show some variation in chromo-
some form, although it is not so striking as in C. virens.
Hieracium venosum, exceptionally good material also in-
vestigated by Rosenberg, has shown the same thing.
Edith Hyde remarks on the fact of the constant reappear-
ance of certain chromosome forms among hundreds of
division figures which she observed in Hyacinthus orien-
talis. Sauer mentions a very long chromosome constantly
present in pollen mother cell preparations of the lily-of-
the-valley, and Strasburger and Lutz found a large
chromosome among many small ones in Lychnis dioica.
In certain species of Yucca this chromosome differentia-
tion takes on a dimorphic aspect, ten of the chromosomes
being very large and about forty-five very small.

Taking into consideration all of these facts, of which
hardly more than a random sample has been given, one
is clearly justified in concluding that these cell characters
are reproduced generation after generation. Why this
constancy if they are not important?

PHYSIOLOGICAL INDIVIDUALITY OF THE CHROMOSOMES

There is also considerable reason for believing that the
various chromosomes of a cell may have different func-
tions.

Boveri was the first to endeavor to test this hypothesis
by allowing sea-urchin’s eggs to be fertilized by two sper-
matozoa. Three nuclei, each with eighteen chromosomes,
were thus present in the same egg, two male and one
female. Although cytoplasmic division seemed to pro-

464 THE AMERICAN NATURALIST [Vor. XLIX

ceed normally, the chromosomes were usually distributed
irregularly by a three-poled or a four-poled spindle. As
a result three or four cells were produced at the first divi-
sion of the doubly fertilized egg, instead of the two cells
that arise after normal fertilization. Various abnormal
larvæ were produced later. In such embryos, Boveri found
the organism to be divided into definite regions, thirds or
fourths, each part traceable to one of the three or four
original cells, and the cells of each part differing from
the cells of the other parts in their combination of chro-
mosomes and usually in their chromosome number. In
rare cases normal embryos were produced, but these were
more commonly developed from a doubly fertilized egg
which in its first division was three-celled, than from one
in which it was four-celled. The thought occurs at once
that three cells have a better chance than four cells in
securing a full set of chromosomes, both as to number and
kind. If the division were normal, each nucleus would
receive a full set in the case of the chromosome distribu-
tion to three cells, but the division is usually irregular,
and because of this irregularity each cell does not usually
secure its normal set of chromosomes. Nevertheless it
is clear that the embryo parts developed from the three-
celled cleavage stand a much greater chance of being
normal than those from the four-celled type, although
through irregularities in division an eighteen-chromo-
some-celled region might be formed even where the first
division was four-celled.

In some cases, the embryo was completely normal as
regards skeleton and pigmentation in one or even two of
its thirds, while the remainder was entirely lacking in
these characters. Nearly normal embryos occurred which
were perfect as to parts and specific characters, but indi-
vidual variations which normally should have appeared
in separate larvæ were present among the thirds. Asym-
metrical larvæ also were formed.

More important still are the results Boveri obtained by
isolating the three cells of the three-fold type and the

No. 584] HEREDITY AND ITS MEANING 465

four cells of the four-fold type and allowing them to de-
velop into larve. When the four cells of a four-celled
stage of a normal embryo are separated, each cell pro-
duces a normal dwarf embryo alike in every respect, but
the three- or four-celled embryos from double fertilized
eggs, when treated in the same manner, never produce
normal dwarfs even when the chromosome distribution
has been numerically equal. Large numbers of larve
brought into existence through this experiment showed all
possible combinations of characters, just as all possible
chromosome combinations were found in their nuclei,
and from these and other data the conclusion is drawn
that ‘‘not a certain number, but a certain combination of
chromosomes is necessary to normal development, and
this clearly points out that chromosomes have different
qualities.’’ In other words, the sea urchin has a set of
eighteen chromosomes, each chromosome performing at .
least some different functions from its neighbors, making
it necessary for the whole set to be present in order to
Insure normal development.

In further investigations, Boveri placed sea-urchin eggs
which had been normally fertilized and were about to di-
vide under pressure. As a result, division of the nucleus
took place, but often no division of the cytoplasm. Such
eggs on again dividing often formed more than two poles,
resulting in inequalities in chromosome distribution and
abnormal larval development. Boveri puts upon these
cases an interpretation similar to that of the preceding
experiments, as the irregular chromosome distribution
seems to be all they have in common.

Morgan comments on Boveri’s experiments as follows:

The evidence makes probable the view that the different chromosomes
may have somewhat different functions and that normal development
depends on the normal interactions of the materials produced by the
entire constellation of chromosomes.

Artificial parthenogenesis and experiments with enu-
cleated eggs have proved that only one set of chromosomes
is necessary to normal development of embryos, but it is

466 THE AMERICAN NATURALIST [ Vou. XLIX

important, in considering these experiments, to note that
two sets of similar chromosomes are present in a normal
sexually produced organism.

Pairs of chromosomes of each shape and size (if they
differ in shape and size) are nearly always found in the
somatic cells—the exception being when the so-called
accessory chromosomes are present. And since but one
of each kind is found in the two gametes that fuse to form
the new organism, it is only natural to suppose that one
set was contributed by the maternal parent and the other
by the paternal parent.

_ The numerous cases in which this phenomenon has been
demonstrated are to many the most convincing evidence
of some sort of a morphological individuality of the chro-
mosomes. To them the fact implies pairs of freight boats
loaded with the essential materials of life, to others—the
minority—it is no more wonderful than the constant re-
currence of other plant organs. At any rate, it has been
shown that these sets of chromosomes continue an appar-
ently independent existence for some time. Moenkhaus
produced hybrids between the two species of fish, Fundu-
lus heteroclitus with long straight chromosomes and
Menidia notata with short curved chromosomes, and the
early divisions of the fertilized egg showed clearly com-
plete sets of chromosomes from each parent. Rosenberg
obtained similar results in crosses between the two sun-
dews, Drosera longifolia, which has forty small chromo-
somes, and Drosera rotundifolia, which has twenty large
chromosomes. In some eases similar to the latter, where
one parent contributes a greater number of chromosomes,
it should be noted that the organism seems to have regula-
tory powers. The chromosomes unnecessary for a double
set are either thrown out or take no part in the activities
of cell division. For example, in the supposedly hybrid
sundew, Drosera obovata, Rosenberg found that its thirty
chromosomes behaved in the following peculiar manner.
Ten of them paired with another ten, but the other ten
remained unpaired and acted in a very abnormal fashion

No. 584] HEREDITY AND ITS MEANING 467

in the reduction divisions. The ten pairs separated nor-
mally, one of each pair going to each pole; but the ten
unpaired were irregularly distributed, sometimes nearly
all of them going to one pole, sometimes most of them be-
coming lost in the cytoplasm and forming small nuclei.
Embryos were produced in a very few cases and these
only through back-crossing with pollen of D. longifolia.
Unfortunately these embryos only developed through a
few cell divisions.

These chromosome pairs have been distinguished by
the name homologous chromosomes. For a long time it
was thought that the paternal and the maternal set of
chromosomes separated from each other bodily at the re-
duction division. Now it is believed to be only a matter
of chance which chromosome of a pair passes to a particu-
lar daughter cell. There is some cytological evidence for
this view, but the main argument in its favor is that this
behavior is all that is necessary to fit nearly all the known
facts of heredity, with the chromosomes playing the
part of the active heredity machinery as will be seen
shortly. This statement is true in a broad sense, but the
word nearly is used because there is an exception to it.
Chance apportionment of either member of a homologous
pair of chromosomes to a daughter cell accounts for all
facts of alternative (Mendelian) inheritance except where
there are breaks in the correlation between characters
usually inherited together. Since such breaks in corre-
lation are common, it is clear that there must be a period
when chromosome pairs have such an intimate relation
that material can be exchanged. Many biologists believe
that such a period is found during the maturation of the
sex cells. The particular point at which such a conjuga-
tion or approximation of chromosome pairs takes place is
called synapsis; it occurs as a part of the prophase or
first stage of the reduction division. Some investigators
have been unable to demonstrate any real chromosome
fusion at this time, but all agree that there is an approxi-
mation between the two sets, and a chance for some kind
of an exchange or interaction to take place.

468 THE AMERICAN NATURALIST [ Von. XLIX

Evidence of the physiological individuality of the chro-
mosomes may be concluded by referring briefly to the so-
ealled accessory chromosome. This fraction of a chro-
mosome, whole chromosome, or in some cases, group of
chromosomes, pessesses no true synaptic mate, and there-
fore at reduction division two types of daughter cells are
found. The presence or absence of the ‘‘accessory’’ is so
closely associated with sex determination that most biolo-
gists now regard it as the morphological expression of a
germinal sex determinant. The essential result of re-
searches on this body may be summed up in the following
words of Wilson.

They have established the existence of a visible difference between the
sexes in respect to these chromosomes, and have shown that it is trace-
able to a corresponding difference in the nuclei of the gametes of one
sex or the other.

The simplest type of accessory chromosome, where the
male possesses an unpaired chromosome which passes
to one pole undivided in one of the spermatocyte divisions
and hence enters but half the spermatozoa, was discovered
by Henking (1891) in Pyrrhocoris. This work was con-
firmed in certain species of Orthoptera in 1902 by Mc-
Clung, who advanced the hypothesis that the odd chromo-
some was a sex-determiner. Shortly afterward this was
made more probable by Wilson and by Stevens who
proved for several species of Hemiptera that the body
cells of the males contain one less chromosome than the
females. Two accessory or X chromosomes are present
in the female, while but one is present in the male.

About the same time, both Wilson and Stevens inde-
pendently discovered another kind of dimorphism in male
germ cells of certain Hemiptera. Here the X chromo-
some of the male has a smaller synaptic mate Y. The
body cells of the female, however, show two of the large
X chromosomes. The sexes, therefore, both contain the
same number of chromosomes, but have the same type of
chromatin difference as was first discovered. The female
is XX and the male XY.

No. 584] HEREDITY AND ITS MEANING 469

Baltzer claimed in 1909 that in the sea urchins Sphere-
chinus and Echinus the sex with the dimorphic germ cells
is the female instead of the male, but the work of Tennent
has shown him to be in error and he has retracted the
statement. There is, therefore, no undisputed cytological
evidence demonstrating this type of dimorphic eggs; but
since breeding results on certain species of birds and of
lepidopters can be interpreted only on such an assump-
tion, it is safe to assume that sooner or later they will be
found.? Whether or not there are animals of this type,
however, is of no particular importance in the present
discussion. What we desire to emphasize is that a large
number of animals, including man, have been shown to have
a chromatic difference between the sexes, and that this
difference is readily explained by the fact that the eggs
are of a single type and the spermatozoa of two types.

In dicecious plants no such morphological differentia-
tion has been found. But this fact does not negate the
idea that the visible differences found in animals are really
sex-determining differences. We have only to suppose
that the dimorphism is primarily qualitative and second-
arily quantitative. Indeed Wilson has found that the Y
chromosome—the synaptic mate of the X—may vary in
different species from a size equal to that of X until it
disappears entirely, leaving X without a mate.

There is only one criticism in this whole matter. One
may admit these cytological differences between the sexes,
but hold that they are early appearances of secondary sex-
ual characters. Morgan, von Baehr and Stevens have
answered this impeachment. In the phylloxerans and
aphids all the fertilized eggs produce females; males arise
only by parthenogenesis, though females may arise in this
manner. The cytological facts are as follows: Under
favorable external conditions eggs develop without reduc-
tion and females are formed. Under unfavorable condi-
tions one or two chromosomes (the sex determiners) are
thrown out. If these eggs develop without fertilization

2Dimorphie eggs in Lepidoptera have recently been demonstrated by
both Doncaster and Seiler.

470 THE AMERICAN NATURALIST [ Vou. XLIX

males arise. The somatic condition of the females may
therefore be termed XX and that of the males XY. If
both reduced normally at any time, ordinary fertilization
might be expected to give both males and females. But
the spermatocytes without X degenerate, leaving only one
type of functional spermatozoa, which produces females.
Thus actual causal connection between the X chromosome
and sex determination appears to have been demonstrated.

These are the main cytological arguments in favor of
the chromosome view of heredity that seem to me to be
insuperable. There are minor arguments both pro and
con, which, as I said in the beginning, we have not space
to consider. Instead it seems more profitable to- show
how Mendelian results interlock with those from cytology
like the parts of a jig-saw puzzle.

CHROMOSOMES AND MENDELIAN INHERITANCE

The principal phenomena of Mendelian inheritance are:
(1) characters that breed true; (2) uniformity of the
population of the first hybrid generation in particular
traits in which homozygous parents differed; (3) inde-
pendent segregation of certain character determiners;
(4) recombination of certain characters; (5) perfect
coupling between certain characters; and (6) partial
coupling between certain characters. Let us see how
plausibly one čan picture the mechanism through which
such phenomena may result without imputing to the
chromosomes any behavior that is not known to occur.
To do this simply let the imagination portray a plant spe-
cies having four chromosomes, each chromosome having
three character determinants that can be followed through
the breeding results that are obtained.

Our figures represent the immature germ cells of the
plant just previous to the reduction division. Fig.
shows the germ mother cell with a duplicate set of heredi-
tary determinants. The mature germ cells are exactly
alike, therefore the plant breeds true to the characters.
concerned.

ni

No. 584]

HEREDITY AND ITS MEANING

Q W p

Meo

Fig. 1

Fic. 2

Que

Meo

aos

Meo

471

Suppose, however, that a change in the germ plasm has

occurred (Fig. 2) at some time or other.

In one member

of the first pair of chromosomes, determinant ‘‘A’’ has
The mature germ cells differ from each
other by one factor.

become ‘‘a.’?

Q Wp

=" ug

Fie. 3

QW

1
mo]

For this reason the plant does not
breed true, but gives a mono-hybrid Mendelian result.

472 THE AMERICAN NATURALIST [Vou XLIX

Again, if such a change occurs that A becomes A’ (Fig.
3), a series of triple allelomorphs giving monohybrid re-
sults with each other, is formed. ‘‘A’’ is allelomorphic
to 66 Ar? or &tg 2?

Qov
H
O p
QW

Aad

D
E
F

Eng e.
giok»

FIG. 4

But there are other character determinants in the first
pair of chromosomes. What happens if both ‘‘A’’ and
“B” become changed? There are two possibilities, as
shown in the two parts of Fig. 4. If one of the members
of the pair of homologous chromosomes becomes abC
while the other remains ABO, there is a positive corre-
lation between the inheritance of ‘‘A’’ and ‘‘B.’? On the
other hand, if the change is such that the two chromo-
somes are aBC and AbC, there is a negative correlation
between A and B. In other words, the determinants re-
main correlated in the same way they entered the com-
bination. There may be breaks in these correlations,
however, as Morgan has shown in Drosophila; and these
breaks in correlation occur in a constant ratio. Diagram-
matically, it may be said that A and B are always the same
distance apart in the chromosome structure and that the
determinants ‘‘cross over’’ from one member of a pair
to the other every so often. All of the gametes in the
first case are not ABC and abC, for example. Some of
them will be AbC and aBC. And the same percentages
of these cross overs are found in the second case where
“A” and ‘‘B”’ are correlated negatively. Furthermore,

No. 584] HEREDITY AND ITS MEANING 473

if C should become ec, and the chromosome pair take the
form ABC and abe, there are definite relations between the
three determinants. Breaks in correlation occur, and this
ratio is constant, so that if given the percentage of breaks
of correlation between “A” and ‘‘C’’ and “B” and ‘‘C,”’
the percentage of breaks between ‘‘A’’ and ‘‘B’’ can be
predicted. If there is a break in the correlation between
“A” and ‘‘C’’ 30 times in 100, and a break between ‘‘B’’
and ‘‘C’’ 10 times in 100, then there will be breaks in the
correlation between ‘‘A’’ and ‘‘B’’ 20 times in 100.

QW b&
QWs

Yeo
ha

Fie. 5

Likewise, the determinants in the second pair of chro-
mosomes are coupled together in their inheritance. D, E
and F have each their peculiar linkage to the other, a link-
age that remains comparatively constant. Yet the de-
terminants in the second pair of chromosomes are entirely
independent from those in the first pair in their inheri-
tance. For example, if, as shown in Fig. 5, ‘‘A’’ should
become ‘‘a’’ in either member of pair number one, and
“D” should become ‘‘d’’ in either member of pair number
two, Mendelian dihybridism would result. Furthermore,
if “A” and “D” should each have the function of affect-
ing the same general character complex in somewhat the
Same manner, there would be an apparent 15:1 ratio if
dominance were complete or a series of types ranging
from the type of one grandparent to that of the other, if
dominance is lacking.

These are the main features that have been established

474 THE AMERICAN NATURALIST [ Vou. XLIX

by recent work on hybrids. We have pictured them as
actual chromosome functions, because every part of the
description has been actual fact as far as the breeding
experiments go. Our picture, it is true, is fictitious, for
we do not know absolutely that the heredity mechanism is
of this nature. But the facts do fit perfectly all that is
known of chromosome behavior. It seems impossible,
therefore, that there should be so many coincidences.

There are also two other pieces of evidence that fit in
and round out the case. Bridges has shown that females
occasionally occur in Drosophila bearing the sex-linked
characters borne by the mother but showing no influence
of those borne in the father. Such exceptional females
were found to inherit directly from their mother the power
of producing like exceptions, and it was proven cytolog-
ically after the prediction had been made from the breed-
ing facts that these females resulted from the non-disjunc-
tion of the X chromosomes at the maturation of the eggs
from which they came, and that one half of their daughters
did in fact contain a Y chromosome in addition to two X
chromosomes. This appears to be definite proof that sex-
linked genes are borne by the X chromosomes.

The other important basis for regarding the chromo-
somes as the material basis for heredity also comes from
Morgan’s work on Drosophila ampelophila, this being the
only species upon which sufficient work has been done to
give a reasonable basis for the conclusion. All of the hun-
dred and thirty or so mutations in this species wpon which
Morgan and his students have worked are so linked to-
gether in heredity that they form four groups correspond-
ing to the four pairs of chromosomes found in the species.
If one single character should be found that did not fit into
one of these four groups, the whole theory would break
down. But no such character has appeared.

This completes the case for the chromosomes as regards
the main facts, and it seems only proper that a fair-
minded jury of scientists should render verdict for the
plaintiff. No case is so bad, however, that a lawyer can

No. 584] HEREDITY AND ITS MEANING 475

find nothing to say for the defense and scientists in this
respect resemble the men of the bar. Certainly there are
some outlying facts, but they are comparatively unimpor-
tant. If a series of important facts should at any time
be found which do not fit, the chromosome mechanism
should be looked into. It is likely that the explanation
will be found in an abnormal chromosome behavior as was
the case in the aphis.

_ PRACTICAL CONCLUSIONS AND Discussions

If now it be accepted as a reasonable premise that the
chromosomes are the chief if not the sole bearers of he-
reditary determinants of body characters, and that their
behavior is a rough indication of the mechanism of he-
redity; what cytological facts, if any, can be made useful
at present or in the future to plant and animal breeders?
If such data exist, they should be put to service; if it is
likely that such facts can be found, investigations should
be undertaken. The broad question may be divided into
three parts which will be discussed in regular sequence :

1. What are the relations of chromosomes to somatic
characters?

2. What are the relations of normal chromosome beha-
vior to the transmission of characters?

3. What are the relations of peculiar or unusual chro-
mosome behavior to the transmission of characters?

RELATIONS OF CHROMOSOMES TO INTERNAL CHARACTERS

Some very interesting observations have been made on
. the relations of internal and external characters to chro-
mosome number.

Farmer and Digby in a comparative study of the cells
of a fern of the genus Athyrium with similar cells of three
of its varieties, found that the measurements were suc-
cessively larger in the three varieties than in the species,
and that there was a corresponding increase in the number
of chromosomes, the gametic numbers for the species and
its varieties being estimated at 76-80, 84, 90 and 100,

476 THE AMERICAN NATURALIST [Vou. XLIX

respectively. Investigations on another fern, Lastrea, did
not corroborate these results, however, in one variety the
chromosomes being more numerous and the cells smaller -
than in the parent type.

Gates by comparing nuclei and cells of different tissues
of Enothera Lamarckiana and similar structures in its
‘mutant’ O. gigas with double the number of chromo-
somes, found that the O. gigas cells and nuclei were always
larger, varying from a comparative ratio of 1:1.5 to 1:3.
At the same time, it would hardly be wise to maintain that
this is always the case, for ~~ a few individuals were
investigated.

Primula sinensis has two foxstia in cultivation, similar
except as to size. The giant form has flowers about one
and one half times the size of those produced by the ordi-
nary form. Gregory investigated these two forms cyto-
logically to determine the cause of this increase. The
nuclei and the chromosomes of the giant form were a little
larger, though the difference was hardly a measurable one.
The chromosome number was the same in both the forms.
In a later investigation he has found that some exceed-
ingly large plants with nuclei distinctly larger than those
of the normal form had double the number of chromo-
somes normal to the species.

Boveri investigated this same relation of cells and nu-
clei to chromosome number in N, 2N and 4N larve of the
sea urchin. From these studies, he concludes that chro-
matin is non-regulatory, and in the case of decrease, un-
regenerable, the cytoplasm in contrast showing the fullest
regulatory activity. Further, the size of the larval cells
is governed by the chromosome mass and the cell volume
is directly proportional to the chromosome number. On
the other hand, Conklin’s investigations on annelids, mol-
lusks and ascidians lead him to take a position opposed
to that of Boveri. He says:

The size of the nucleus, centrosomes and chromosomes is dependent
upon the volume of the cytoplasm is clearly shown in Crepidula, where
in large and small blastomeres, these structures are invariably propor-
tional in size to the volume of cytoplasm.

No. 584] HEREDITY AND ITS MEANING 477

_ Neither chromosomes nor nucleus control, the size of the
cell in annelids, mollusks or ascidians.

RELATIONS BETWEEN CHROMOSOMES AND EXTERNAL
CHARACTERS

Thus there seems to be no constant relationship even
between nuclear or cell size and number of chromosomes,
and bonds of union between external taxonomic charac-
ters and chromosome number seem to be still more tenu-
ous. It is true tliat certain giant Primulas and Ginotheras
had more chromosomes than were characteristic of the
normal forms, but it is just as clear that all giant Primulas
(and the same is probably true of Ginotheras, from the
work of Heribert-Nilsson and of Geerts) do not have ab-
normal chromosome numbers.

Results on several species of both animals and plants
are interesting in this connection.

The thread worm, Ascaris megalocephala, has two va-
rieties, bivalens and univalens, the former having as a 2N
number four chromosomes, the latter two chromosomes.
Nothing is known as to the origin of these two forms.
They are found parasitic in the same host individual and
neither form is rare. According to Herla, they hybridize
freely and produce embryos whose cells have three chro-
mosomes, but no mature hybrids have ever been found.
Meyer could distinguish no anatomical differences þe-
tween the two varieties.

Rosenberg investigated the reproductive structures of
two species of sundew and found one to have double the
chromosome number of the other. A subsequent com-
parison of anatomical and taxonomic characters failed to
show any sharply marked differences between them ex-
cept in size. The form having the smaller chromosome
number was smaller and less robust. They inhabit the
same territory and produce natural hybrids which are
sterile.

Rosa canina has two varieties which have the same taxo-
nomic characters, but one form has thirty-four while the

478 THE AMERICAN NATURALIST [VoL XLIX

other has only sixteen chromosomes. The form with
thirty-four chromosomes is apogamous and reproduces
without fertilization, but that one must not conclude that
apogamy is necessarily associated with a double or an in-
creased chromosome number, is clear from the case of
Rumex. Rumex was investigated by Roth; one species,
R. cordifolius, having forty chromosomes as its 2N num-
ber, required fertilization to produce offspring; another
species, with only sixteen chromosomes, was apogamous.
A short list of nearly related species or species with two
varieties varying in their chromosome numbers with their
character differences, if present, is given below.

Name Date | N | 2N Characte rs Investigator
Alchemilla SSAB 1904 |32| 64 |Apogamous Strasburger, E.
aphanes....... 1904 | 16| 32 > =
Ascaris ioe hap 1883 | 2| 4 |Alike externally |Van Beneden
s4 e e rg 1895.{| 2| 4 Meyer, O.
rh Ne Ge eee 1 99 T “ and others
Ascaris lumbricoides....... 1887 24 Boveri, T.
n De eee 1887 48 S 4
Dahlia variabilis S E 1911 |16| 32 Ishikawa, M.
vee ee 1911 | 32} 64 3 S
Drosera Baramin P T 1909 |10| 20 Rosenberg, O
ola... 20 | 40 More robust, ete. i s
Echinus microtuberculatus 1888 | 9| 18 Boveri, T.
1902 |18| 36 oh A
Heliz pomatia...........: 1903 |24 | 48 |Alike externally |Ancel, P.
: ie eee iat oe 1896 |12| 2 v. Rath, O.
N ephrodium molle onas 1908 |64 |128 None mentioned | Yamanouchi, 8.
ree ane 1908 66 132 :
Cnothera seul shame ay 191) rn Gates, R. R.
eh 909 | 14 |. 28 Large and 2 ERS
coarser j
FUA sinensis wa 1909 |12| 24 Gregory, R. P.
t form ing ROK 1909 12| 24 More robust n ee
a ee 1914 |24| 48 = Ly g ee
Rosa canina Pa aah ian ap a 1909 34 |Apogamous Rosenberg, O
Cie eis OS c Les 190: 8} 16 Strasburger, E
Thalictrum WRUNG os 1909 | 12 | 24 Overton, J
rpuras 1909 |24 | 48 |Apogamous a $6:
Zea was meee Flint’ a 1911 !10 Kuwada, Y. “
r e Bo e ooo, 1911 |12 | he

What conclusions can be drawn from these facts? Cer-
tain botanists have attempted to connect chromosome
doubling with apogamy, as usually the chromosome num-
ber in apogamous species is higher than in the normal
species of the same genus; but there is no evidence of

No. 584] HEREDITY AND ITS MEANING 479

apogamy in @nothera gigas, and in Rumex the form with
the low number of chromosomes is apogamous while the
form with the high chromosome number requires fertili-
zation. On account of these exceptions, therefore, it |
seems probable that the cause of apogamy is deeper than
a mere doubling of the chromosomes, even though doub-
ling may usually accompany such a change in reproduc-
tive habits.

Variation in chromosome number in the same species
has been proposed as a cause of general variation in so-
matic characters, but the evidence is not clearly in favor
of such a theory. In the fern Nephrodiwm molle Yama-
nouchi found spermatid cells to be of two sorts, those with
sixty-six and those with sixty-four chromosomes. This
would mean that Nephrodium has two gametophyte forms
and two sporophyte forms, externally identical, so far as
our present knowledge goes, but differing in their chro-
mosome numbers.

Further, sporophytes developing from the prothallia of
ferns without the intervention of a sexual process have
the N instead of the 2N chromosome number, yet apoga-
mously developed fern sporophytes, except as to chromo-
some number, are indistinguishable from normal sexually
produced individuals of the same species.

Many writers have been tempted to postulate a causal
relation between the numerical variation of chromosomes
among the species of a genus and the genera of a family
and their specific and generic characters. The thirty or
more species of Composite investigated have shown a
remarkable variation in their chromosome numbers, the
2N numbers ranging between six and sixty, and, as is well
known, the Composite possess an infinite variety of
sharply contrasting characters. But the lily family also
has an enormous number of characters in its species and
genera, and the genus Liliwm, with its great variety of
characters distributed among forty-five species, is typical
of the other genera of the family, as far as present inves-
tigations go, in having the same chromosome number for

480 THE AMERICAN NATURALIST [Vor. XLIX

all of its species. Others suggest that the more chromo-
somes a plant species possesses the greater is its varia-
bility. Thus Spillman? speaks of the low variability of
rye, suggesting its small chromosome number (six or
eight) as a possible reason; for maize, having probably
from twenty to twenty-four chromosomes, is infinitely
more variable than rye. However, Britton’s ‘‘Manual’’
selects Crepis virens for special mention as an extremely
variable species from among the four or five other species
listed under that genus, and it is known that C. virens
has only six chromosomes, while three other species of
Crepis investigated all have higher numbers. Again, ac-
cording to Wiegand, the Canna has only six chromosomes,
yet every gardener is well acquainted with the infinite
variety in Cannas.

THE CHROMOSOMES AND VARIABILITY

After a consideration of the above facts, one may well
hesitate to state that there is even a high degree of corre-
lation either between variability in chromosome number
and general variability, or between high numbers of chro-
mosomes and a high degree of variability in specific char-
acters. On the other hand, it is not certain that the data
upon which our discussion is based are relevant to the case
in hand. We have discussed a possible relationship be-
tween chromosome numbers and species complexity and
variability as found in the wild. This is not at all the
same thing as discussing the relationship between chro-
mosome number and true variability. It is true that com-
plexity and specialization of plants and animals seem to
have no connection with chromosome number, and that
within a family a genus or a species profusion of taxo-
nomic characters do not go hand-in-hand with high chro-
mosome numbers. But in these cases our data come from
persistent forms. What the actual inherent variability
of the protoplasm is in most cases we do not know. Dro-
sophila ampelophila, a species with only four chromo-

3 Six according to Westgate’s unpublished data; eight according to Nakao.

No. 584] HEREDITY AND ITS MEANING 481

some pairs, is considered to be very constant in its char-
acters from the taxonomist’s standpoint, yet by careful
continued observation Morgan has succeeded in detecting
over 130 mutations.

From a strictly mathematical standpoint, it would seem
that if other things are equal, variability would take place
in proportion to the number of chromosome units. The
difficulty is that in no case do we know anything whatever
about the relative complexity of any particular chromo-
some unit. One must infer, however, that the 47—48 chro-
mosomes in man are individually much more complex than
the 128-132 chromosomes in the fern Nephrodium molle.
If this inference be correct there are reasons why altera-
tion in determinants may occur in direct proportion to
the number of chromosomes or rather to the mass of chro-
matin without there being visible somatic variability in
the same ratio. Let us construct an imaginary plan for
preventing visible variation without preventing change
in chromosome determinants. Unquestionably the sim-
plest means is to double the chromosome number. Se-
lecting, for example, a species with four chromosomes, let
us suppose that fertilization occurs without a reduction
division at some time or other. Then instead of a dual
organism with two sets of chromosomes of similar func-
tion, one from the male and one from the female parent,
there would be a quadruple organism with two sets of
similar chromosomes from each parent. Any germinal
change which would produce a new dominant character
would be apparent immediately, but for a recessive change
to appear—and these are many times as numerous as the
others—it would be necessary to have identical changes
occur in two chromosomes. Following out this line of
reasoning, it is not hard to see what a great possibility
for uniformity there is in further chromosome duplication,
provided the actual fact of duplication makes no great
change in the organism. That chromosome doubling has
no decided visible effect we have seen from the cases
already described; and since so many nearly related spe-

482 THE AMERICAN NATURALIST [ Vou. XLIX

cies and varieties have their chromosome numbers in
series 1:2:3:4, ete., it seems by no means improbable that
what we have imagined above has actually occurred many
times. And if one may believe that the eyent has the
result supposed, all the worry about relationships between
chromosome number and height of species in the scale of
evolution may be eliminated.

Norman CHROMOSOME BEHAVIOR AND HEREDITY

The second query, concerning the relation of normal
chromosome behavior to the transmission of characters,
is much more important than the one just examined, but
it can be discussed more briefly. By normal ‘‘chromo-
some behavior’’ is meant a reduction division where ma-
ternal and paternal chromosomes approach each other in
definite pairs (if homologous pairs are present), chance
only governing the passage of either to a particular
daughter cell. This is probably the usual behavior in the
higher plants and animals, and upon this behavior depends
Mendelian heredity in the narrow sense. The thesis to be
submitted and scrutinized is the following: The maximum
possible difficulty in the improvement of animals and
plants by hybridization usually depends directly upon the
chromosome number.

When a mutation in a single determinant takes place in
the germ cells of a plant, such as may cause the loss of
red color in the corolla, crosses between such a form and
the normal give a monohybrid Mendelian result. Two _
mutations in non-homologous chromosomes gives in a
similar way a dihybrid result. Such simple conditions,
however, are not met with very frequently. For example,
White found that a fasciated tobacco when crossed with
the type from which it sprang and from which it probably
differed only by this single determinant, gave a mono- —
hybrid Mendelian ratio in the F, generation; but when the
fasciated type was crossed with other types the result was
a complex F, population. This population was suscepti-
ble of analysis, nevertheless, and showed that the various

No. 584] HEREDITY AND ITS MEANING 483

varieties with which the fasciated type was crossed dif-
fered from it by several determinants, each of which was
transmitted independently though they every one aff ected
the development of fasciation. This illustration is not
one of a rare phenomenon. It is what geneticists find
constantly in their experiments. Presence or absence of
a particular character may depend upon the presence or
absence of a particular essential determinant, but, given
this determinant, sooner or later the investigator finds
several other determinants which modify the expression
of the character. The existence of these modifiers has
been the cause of a great deal of confusion in the analysis
of breeding results, but in reality the inheritance is sim-
ple. The experience that all investigators who have
worked intensively have had with them shows that prac-
tically all somatic characters are due to multiple determi-
nants in the germ cells. It merely depends on the rela-
tive difference between the germ plasms brought together
in crosses, how complex the resulting F, populations ap-
pear. Since even apparently simple characters are thus
due to complex germinal interactions, that results of
crosses made for the purpose of improving such intangi-
ble things as yield, size, quality, etc., should be complex,
is not astonishing. In the comparatively extensive expe-
rience that the writer has had in breeding tobacco, maize,
peas and beans the wide variability of the F, population
in crosses between distinct varieties leads him to think
that it is extremely common for such varieties to differ
qualitatively in every chromosome. Furthermore, the
relative complexity of the segregating populations is
much greater in tobacco than in corn and greater in corn
than in peas or beans. What can this mean but that when
varieties are found that differ qualitatively in all of their
chromosomes, the complexity of the result varies directly
with the number of chromosomes present.

In Mendelian inheritance the number of actual types
(both homozygous and heterozygous) present in the F,
population when all are represented is 3", and the number

484 THE AMERICAN NATURALIST [ Vou. XLIX

of individuals that must be present to give an equal chance
for the presence or absence of an individual of every type
is 4", where n represents the number of allelomorphic
pairs. This being true, if differences in all of the chro-
mosomes are predicated in tobacco and in pea crosses, the
maximum number of individuals necessary in the F, gen-
eration to allow for one reproduction of each of the grand-
parental forms is 42 in the first case and 47 in the second
ease. It is clear that there is an absolutely overwhelming
difference in the difficulty of recovering the grandpar-
ental forms in the two examples.

Now this is about what one wishes to do in many plant-
breeding problems. It is desired to combine one or two
characters from one parent with all of the other qualities
of the second parent. And such has been my experience
that I believe that this maximum possible difficulty in the
operation as predicated by qualitative differences in all of
the chromosomes often occurs. There can be no question
on these grounds of the importance of determining the
number of chromosomes in a species before beginning
a complex plant-breeding problem, and thus being able to
comprehend the maximum possible difficulties that may
be encountered. How greatly these difficulties vary may
be seen in the very incomplete list of chromosome counts
in common plants that is given below.

Common Name | Scientific Name ee 2N Date Investigator
Banana..... Musa sapientum, "dole 8 16" 1910 |Tischler, G.
ree gi ald Musa sapientum, |
"Radjah Siam’”’...... i 26 a 1910 z md
T eas aor anime Ss 24 48 1910 oe Mi
Bean; oY. Phaseolus vulgaris......| 8 16 1904 |Wager, H.
Calla lily.. pps Africana.....| 8 16 1909 (Overton, J. B.
chy eRe Can indica Cat eee es ge 6 1900 |Wiegand, K. M.
Cee ee en eee eet | 8&8 more than
| 10 1904 |Koérnicke, M.
Coti. ss cee Zea Mays, * ‘yellow
starchy” ‘‘amber

pop,” “black ma nage
“golden broach field,”
Cahita Mat coo, 10 “op? 1911 |Kuwada, Y.

cece, 9-10 | “18-20” | 1911 j T

RD Eo pruebas 9-12 1911 585 2

No. 584] HEREDITY AND ITS MEANING 485
Common Name. Scientific Name N 2N Date | Investigator
Pes A ye ei Zea M ays, re sugar.. 12 1911
Cottone cc] Gossy ypium hybrid’. ..] 28 “567 1903 bai W. A.
ob he diane! ti i “Egyptian” 20 1910 Balls, W. L.
EE ATA ER E E bt “16” | 1906 |Tischler, G.
Dandelion. ` Tarazacum confertum.. 8 “16” 1909 Rosen i,

Pi ere ener) O .|120r13, about 26 | 1905 (Juel, H. O.
Elderberry.. 3 Sambucus 2 PO une eee 18 38 1909 Lagerberg, T.
Ev vening |

primrose .. \@nothera grandiflora... ra 14 1909 Davis, B. M.
Evening | |
primrose... O. lamarckiana........ “Jy 14 1907 Geerts, J. M.
| 7 1911 Gates, R. R.
Evening | |
primrose. . . 0. gigas... i. 5 arous 14 28 1909 Gates, R. R.
eae pee N ia i II RNT E a 64 128
or or 1908 Yamanouchi, 8.
66 132
okies as squalens...........| 12 2 | 1900 Strasburger, E.
tea WRT 7 ae lanceolata var. | |
platyphylum......... 5 10 2014 Tahara, M., and M.
| Ishikaw:
we (Grupis WORE aan 3 6 | 1909 Rosenberg, O
p \Crepis tectorum........ 4 8 | 1905 el, H. O.
=e Crepis japonica........ 8 16 1910 Tahara, M.
NAGS cae & Lilium martagon.......| 12 24 | 1884 Guignard, L.
Lily-of-the- |
alley . Convallaria majalis 18 “36” | 1899 Wiegand, K. M.
Lily-of-the- :
Valley os oe ue majalis..... 6 3 09 Sauer, L. W.
berry... -Morus orus alba, ‘‘Shirowase’’ 17? 40-50 1910 Tahara, M
rS orus indica.......... 14 28 1910 Tahara,
Nightshade Solanum nigrum....... 36 72? 1909 inkler, Hans
D E ein Cone. E AT E y A "107 1898 Schaffner, J. H.
Pees see See Paeonia spectabilis..... i2 oe" 1893 |Overton, E.
eee isum sativum......... 7 14 903 (Cannon, W. A.
Persimmon. .|Diospyros virginiana 30 or
more 1911 (Hague, Stella M.
Bg) RR ah dea, Pinus cate SOA 12 24 1899 Cha ST n, ©. d.
ren PA E Orana oliot. inari 12 244 1910 Kuwada, Y.
irer Rosa va 3 species..... 8 16 1904 retni eane E.
Tobacco..... Nicotiana ap: .. 6.45 ss: 24 48 1913 |White, O. E.
hea bees pasan eA.. 12 24 1909 Winkler, Hans
Eeoae ulipa Gesneriana......| 12 24 1901 t, A.
Wake-robin Trillium grandiflorum 6 3 1899 Atkinson, G F
Seok Triticum vulgar Sees « 8 16 Koérnicke, M
a See re 8 STE" 1893 (Overton, E.
e PE a Pe piu ee ee 8 ek | 908 (Dudley, A. H.

Among these figures are found four of our most impor-
tant crops—wheat, tobacco, corn and cotton. They con-

trast strikingly in their chromosome numbers.

eat

and tobacco, species in which the flowers are naturally
self-pollinated, have 8 and 24 chromosomes, respectively,

4‘*But we often find a larger number.’’ Quotation marks refer to in-
ferred numbers rather than actual countings.

486 THE AMERICAN NATURALIST [ Vou. XLIX

in their gametes. Corn and cotton, species usually cross-
pollinated, have 10-12 and 20-28 chromosomes, respect-
ively, in their germ cells. These species all have been
under cultivation since before there has been recorded his-
tory. Many varieties of each exist. It is not at all im-
probable that with thousands of years of cultivation and
selection under diverse conditions, mutations in most of
their chromosomes have persisted. If, then, improvement
means working on character complexes that involve al-
most all of the plant functions, it does not seem improb-
able that the actual difference in the difficulty of improving
wheat and tobacco is as 48:44, or about 1 to 4,295,000,000.
In like manner corn and cotton compare in the ratio
41°: 428. or 1 to 68,720,000,000. And is it not true that
modern improvement in most of these crops does involve
nearly all the plant functions? Yield in wheat involves
number and size of grain, and number of culms, with all
that these things include in plant economy; yield of to-
bacco involves number, size and thickness of the leaves.
Quality, a mystical word, is perhaps still more complex.
In wheat, it takes in habit of growth of both root and stem
and such other characters as go to make up strength and
hardiness, thickness of pericarp, size of aleurone cells, and
the physical and the chemical character of both endo-
sperm and embryo, as well as their size ratios in regard
to each other. In tobacco, it includes thickness and
strength of leaf, color, texture and all chemical and physi-
eal characters that make for flavor and ‘‘burn.”’

One may say that this is all very well as a theory, but
that it is all theory, and ask what support is given to it by
practise. I have had personal experience with but two
of these four crops. I have worked extensively and in-
tensively with corn and tobacco for some ten years. But
I have followed carefully the published experiments in
breeding wheat and cotton and have seen several of the
more important experiments. And I may say that it was
my observation of the extreme difficulty in the experi-
ments with cotton and tobacco as compared with corn and
wheat that led to this theory of the cause.

No. 584] HEREDITY AND ITS MEANING 487

In proposing this thesis, the chromosomes have been
considered as pairs of freight boats loaded with character
determiners, shifted bodily to the daughter cells by in-
ternal forces of which we are ignorant. Yet this is not
the whole truth. The determiners in particular chromo-
somes seem to be tied together more or less tightly, but
they are not always transferred as one package. They
are coupled in their transmission to the next generation,
but this coupling is not perfect. Breaks in the coupling
occur and there is order and regularity in these breaks.
Our knowledge on these matters rests upon the researches
of Morgan on Drosophila, Bateson on the sweet pea, and
Tanaka on the silkworm, so it is not certain whether these
are common grounds for this regularity or whether each
species has regular laws which control the breaks in cor-
relation. But in either case, these breaks do not inter-
fere with our proposition. They only complicate matters.
In most of the cases in Drosophila, where they are best
known, linkage is comparatively tight, 7. e., breaks are
somewhat rare; but they may become so frequent as to
simulate inheritance from separate chromosomes. In
those cases our theory is of no value, but if Drosophila
is any criterion by which to judge, such conditions are
very unusual.

ABNORMAL CHROMOSOME BEHAVIOR AND HEREDITY

The third query concerning the relations of peculiar or
unusual chromosome behavior to the transmission of
characters may be passed over with a word. In certain
insects, for example, bees, wasps, aphids, phylloxerans,
etc., odd sex ratios and attendant complexities have long
been known. These have been cleared up more or less
-completely by cytological studies. They depended upon
chromosome behaviors that are not usual in animals or
plants. Similar peculiar chromosome mechanisms may
be present in many other species. Attention is merely
called to the fact that if experiments on any plant species
appear to show that its characters do not obey the laws
that have been demonstrated for so many types, their

488 THE AMERICAN NATURALIST [ Vou. XLIX

cytological eccentricities should be looked into. In them
will probably be found the key to the situation. The
(Enotheras may be mentioned as a ease in point. Their
heredity in many cases is not what would be expected by
analogy with other plants. We know that in some ways
the behavior of their chromosomes is irregular. Further
study will probably show that this is the sole cause of their
anomalous heredity.

LITERATURE CITED
Allen, C. E
1905. Nuclear nsike in the Pollen Mother-cells of Lilium canadense.
An , 19: 189-258.
Atkinson, G. F.
899. Studies on Reduction in Plants. Bot. Gaz., 28: 1-26.
W.L

1910. The Mechanism of Nuclear Division. Ann. Bot. 24: 653-665.

1910. Ueber die ag! zwischen dem Chromatin und der Ent-
een und Vererbungsrichtung bei Echinodermenbastar-
den. v. f. Zeliforsch., 5: 497.
gerse W.
1909. iDa ’s Principles of Heredity. Cambridge, R Uni-
sity Press. See bibliography reported ther
Bonnevie, K
1909. Chromosomenstudien I. Archiv. f. Zellforsch.,1: 450-514, 1908.
II. Ibid., 2: 201-278.
Boveri, Th.
1892, Die Entstehung des Gegensatzes zwischen den Geschlechtszellen
und den somatischen Zellen bei Ascaris megalocephala. Sitzber.
Gesell. Morph. u. Physiol. München., Band 8
1902. "ap Mitosen als Mittel zur pares ‘ina Zellkerns. Verh.
u. Med. Gesell. zu Würzburg., N. F., 35.
Bridges, C. B.
1914. Direct Proof Through Non-disjunction that the Six-linked Genes
of Drosophila are Borne by the X-chromosome. Science, N. 8.,
40: 107-10

Cannon, W. A.
1902. A igge Basis for the Mendelian laws. Bull. Torr. Bot.
: 657—661.

, 29:

1903. Studies i in Shit Hybrids: The Spermatogenesis of Hybrid Cot-
Bull, Torr. Bot. Club, 30: 133-172.
1903. Saroiy in Plant Hybrids: The Spermatogenesis of Hybrid Peas.
Bull. Torr. Bot. Club, 30: 519-543,

Cardiff, I. D.

1906. A Study of Synapsis and Reduction. Bull. Torr. Bot. Club, 33:
~ 271-306.

No. 584] HEREDITY AND ITS MEANING 489

Chamberlain, C. J.
899. Odgenesis in Pinus laricio. Bot. Gaz., 27: 268-280.
Conklin, E. G.
1901. The Individuality of the Germ-nuclei During the Cleavage of
the Egg of Crepidula. Biol. Bull., 2: 257-265
1905. The Bapes Theory iag the Standpoint of Cytology. Sci-
» N. By- 21: 525-
1908. The panis of de Science, N. S., 27: 89-99.
Coulter, J. M., and Chamberlain, C. J.
1903, Merphology of the Anigiospérina, N. Y., Appleton.
Davis, B.
1909. Cytological Studies on @Œnothera: I. Pollen Development in
Gnothera grandiflora. Ann. Bot., 23: 551-571.
Delage, Yves,
1903. L’hérédité et les grands probléms de la biologie générale. 2™°*
ed., Paris, Reinwald.
DeVries, H,
1910, Intracellular Pangenesis: Including a Paper on tae
and Hybridization. C. S. Gager, trans, Chi. Open Court.
Dixon, H. N
1894. Fertilization of Pinus silvestris. Ann. Bot., 8: 21-34.
Driesch, H.
1899, Die oe morphogenetischen Vorgänge. Ein Beweis
vitalistischen Geschehens. Leipzig, Engelmann.
Farmer, J. B., and mete. J. E B.
1905. On the Maiotic Phase Segrar a in Animals and
ants. Quart. Jour. Micr. Sci 489-
Farmer, J. B., and Digby, L.
1907. Studies in Apospory and Apogamy in Ferns. Ann. Bot., 21:
161-199.

1905. Betrachtungen über die Chromosomen, ihre Individualität, Re-
duktion und Vererbung. Arch. Anat. u. Physiol., Anat. Abt.,
179-228.

1908. Zur Conjugation der Chromosomen. Arch. f. Zelforsch., 1:
604-611.

Gates, R. R.
1907, Pollen Development in Hybrids of @nothera lata X O. Lamark-
and lation to Mutation. Bot. Gaz., 43: 81-115.
1909. The Stature and Chromosomes of Cnothera gigas DeVries.
h. f. Zeliforsch., 3: 525-552.
1911, Sa ion of Chromosome Reduction. Bot. Gaz. 51: 321-344.
Geerts, J. M
1908. Beiträge zur Kenntnis der cytologischen Entwicklung von
othera Lamarckiana. Ber. Deutsch. bot. Ges., 26a: 608-

-r
1907. praa die Zahl der Chromosomen von (nothera Lamarckiana.
er. deut. bot. Gesell., 25: 191-195.

490 THE AMERICAN NATURALIST [Voi XLIX

— V., and Wygaerts, A.
ta construction du noyau et la formation des chromosomes dans
es cinéses somatiques. I. Racines de Trillium grandiflorum
et cet homeotypique dans le Trillium cernuum. La Cel-
lule, 21: 7-76.
Grégoire, V.
1905, Les résultats acquis sur les cinéses de maturation dans les deux
règnes. Revue critique de la Littérature. (I Mém.) La
Cellule, 22: 221-376.
1907. Les fondements eena = théories courantes sur l’hérédité
Mendelienne. Ann. Soc Zool. et Malacol Belgique, 42:
267-320
1908. Les shini im 1’étape oT represéntent-ils une caryo-
cinése? La Cellule, 25: 87-9
1910. Les cinéses de ane ation haa les deux règnes. L’Unité e
sentielle du processus méiotique. (II Mém.) La Cellule, on: :
223-422,

9. Note on the Giant and Ordinary Forms of Primula sinensis.
Proc. Cambridge Phil. Soc., 15: 239-246.
1909. The Mode of Pairing of Chromosomes in Meiosis. New Phy-
tologist, 8: 146-153
1914. On the BCE of Tetraploid Plants in Prunula sinensis. Proc.
Roy. Soc., B. 87.
Guyer, M. F.
1911. Nucleus and Cytoplasm in Heredity. Am. Nat., 45: 284-305.
Haecker, V.
1907. Die Chromosomen als angenommene Vererbungstriger. Ergebn.
u. Fortschritte d. Zoologie, 1: 1-136.
1910. Ergebnisse und Ausblicke in der a o ea Zeitschr.
Abstammungs. u. Vererbungslehre, 3: 181-

s

Hague, 8.
1911. A Morphological Study of Diospyros virginiana. Bot. Gaz., 52:
3442,

Hegner, R. W
1911. Geri cell Determinants and their Significance. Am. NAT., 45:
385-397.

Henking, H.
1891. Ueber Spermatogenese und deren Beziehung zur Eientwicklung
bei Pyrrhocoris. Zeitschr. wiss. Zool., 50: 1891.

Heider, K ae
1906. Vererbung und Chromosomen. Versamml. deut. Natur u. Artze.
Jena.
Herbst,

C.
1909. Vererbungsstudien VI. Arch. f. Entw., 27: Heft 2.
Hertwig, O.
1898. Die Zelle und die Gewebe. Jena, Fischer.
Hyde, E.
1909. The Reduction Division in the hanes of Hyacinthus orientalis.
Ohio Nat., 9: 539-544.

No. 584] HEREDITY AND ITS MEANING 491

Ishikawa, M.
1911. Cytologische Studien von Dahlien. Bot. Mag., Tokyo, 25: 1-8.
Janssens, F. A,
1905. Spermatogenése dans les Batraciens: Evolution des auxcytes
males du Batrahoseps attenuatus. La Cellule, 22: 380-425.
Juel, H. O.
1897. Die es in den Pollen-Mutterzellen von Hemerocallis
a und die bei egy auftretenden Unregelmissigkeiten.
Jahrb. wiss. Bot., 30: 205-226.
Koernicke, M. Der heutige aa ‘des See Zellforschung.
. Ber. deut. bot. Gesell., 21:
Kölliker, A.
1885. Die Bedeutung der Zellkerne fiir die Vorgänge der Vererbung.
Zeitschr. wiss. Zool., 52: 1-46,
Kuwada, Y.
1910. A Poses ae Study of Oryza sativa. Bot. Mag., Tokyo, 24:
—281,

“1911. TBa in the Pollen-mother Cells of Zea Mays, L. Bot. Mag.,
Tokyo, 25: 163-181.

Lagerberg, T
909. Studien über die Entwickelungsgeschichte und eae ae
Stellung von Adoxa moschatellina. Kungl. Sv. Vet. Aka
pon 44: 1-86
Lawson, A. A.

1911. Nuclear Osmosis as a Factor in Mitosis. Trans. Roy. Soc.
Edinburgh, 48: 137-161,
1911. The Phase of the Nucleus Known as Synapsis. Trans. Roy.
Soc. Edinburgh, 47: Pt. III: 591-604,
Lotsy, J. P.
1906. Ponta di über Descendenz-theorien. Two vols. Jena, Fischer.
Lundegård, H
1909. Tones eo in den Pollenmutterzellen einiger diko-
yl zen. Svensk. Bot. Tidskr., 3: 78—124.
1910. hee Kernteilung in den Wurzel-spitzen von Allium cepa and
a faba. Svensk. Bot. Tidskr., 4: 174-196.
Lutz, Anne es
1909. Notes on the First Generation of Hybrids of O. lata X O. gigas.
Science, N. S., 29: 263-267.
Moenkhaus, W. J.
1904. The Development of Hybrids between Fundulus heteroclitus
and Menidia notata with Especial Reference to Behavior of
nya and Paternal Chromosomes. Am. Journ. Ss ot

Montgomery, pe +.
9 T A in the Spermatogenesis pA the Hemiptera heterop-
tera. Trans. Am. Phil. Soc., N. S., 21: 97-173.
1908. On the Morphological Difference of the piee of Ascaris
megalocephala. Arch. f. Zellforsch., 2: 66-75.
Morgan, T. H.
1910. Chromosomes and Hevedike. Am, NAT., 44: 449-496.

492 THE AMERICAN NATURALIST [Von XLIX

1913. Heredity and Sex. N. Y., Columbia University Press,
Mottier, D. M.
1907. The ac aa of sigan ok Chromosomes in Pollen Mother-
. Ann. Bot., 21: 309-347.
Nakao, M.
1911. a. studies on the nuclear division of the pollen mother
of e cereals and other hybrids. Jour. Col. Agr. Sap-
ro joaki 4: 173-190.
Nawaschin, toa
1 Nitkin über die Bildung der Spermakerne bei Lilium Marta-
gon. aces d. Buitenzorg, 31ème Supplément, 2e Partie:
871-9
Némee, B.
1909. Zur Mikrochemie der Chromosomen. Ber. deutsch. Bot. Gesell.,
43-47.

27:
1910. Das Problem der Befruchtungsvorgänge und andere piaga
Fragen. Berlin, Borntraeger.
pare J.-B:
909. On the Organization of Nuclei in'the Pollen- aother Cells of
Certain Plants with Special Reference to the Permanence of
the Chromosomes. Ann. Bot., 23: 19-61,
Rabl, C.
1885. Ueber Zelltheilung. Morphol. Jahrb., 10: 214-330.
Remak,
1858. Ueber die Theilung der Blutzellen beim Embryo. Müller’s
Arch., S. 178-188.
TR >
190 Bor Kenntnis der Reduktionsteilung in Pflanzen. N. Bot. Not.,

S. 1-24.
1906. Cytological Investigations in Plant Hybrids. Rpt. 3d Intern.
Conf. on Genetics, pp. —291,
1907. ERT — Cytological Studies on the Apogamy in Hiera-
Svensk. Bot. Tidskr., 2: 143-170.
1909. Ueber. _ des Ruhekerns, Svensk. Bot. Tidskr., 3: 163-
173, Taf. 5.
1909. yras die Se bei Tarazacum und Rosa.
nsk. Bot. Tidskr., 3: 150-1
1909. Oytoiogisete und mo Se Studien uber Drosera longi-
a X D. rotundifolia. Kungl. Svensk. Vetensk. Handl., 43:

a
1909. Zur Kenntnis von den PESA der Compositen.
Svensk. Bot. Tidskr., 3:

Sargent, E.
1896. The Formation of Sexual a in Lilium Martagon. I.
ogenesis. Ann. Bot., 10:

1897. The Formation of Pc ‘na in gga Martagon. II.
permatogenesis. Ann, Bot., 11: 187-

Sauer, L. W.

1909. Nuclear Divisions in the Pollen Mother-cells of Convallaria
majalis. Ohio Nat., 9: 497-505

No. 584] HEREDITY AND ITS MEANING 493

Schaffner, John H.
1909. Chromosome Difference in Ascaris megalocephala. Ohio Nat.,
06-508.
Stevens, N. M.
1905. cone in Abo est with Especial Reference to the
‘Ace y Chromosome.’’ Carnegie Inst., Wash., Pub. 36, I.
1906, Studies 3 in agit stays ma II. A Comparative Study of the
Heterochromosomes in Certain snr of Coleoptera, peA
tera and Lepidoptera, hak Especial ee a to Sex Det
mination. Carnegie Inst., Wash, o Pub. 36,
Stomps, T. J.
1911. Kerndeeling en Synapsis by Spinacea oleracea. S. 1-162.
1910. Review Biol. Centralbl., 31: 257-320
T E.
894. pro — of Number of Chromosomes in ny Life His-
ing Organisms, Ann. Bot., 8: 281-
1904, Vener sera ha Sitzber. K. Preuss, prai, Wiss., 18:

1905. _ Tyrie wid allotypische miga Pe ibe eg Beiträge
ererbungsfrage. Jahrb.
1907. Ueber die Individualität m Pa gaa mt ai Pfropfhy-
riden F . Jahrb. f. u Bot., 44: 482
1907. Die Ontogenie der Zell seit an eee: ret ‘ial a
1-138

1908. Chromosomenzahlen, ie ee Vererbungstrager, und
duktionsteilung. Jahrb. f. w -, 45; 479-570.
1909. Meine Stellungnahme zur Frage der Pfropfbastarde. Ber.
D sell. $

utsch.
1909. Zeitpunkt der Bestimmung oe ee. Apogamie, Parthe-
nogenesis und er aan ee Histologische Beiträge, Heft
-124. Jen
1909. The Minute Biden of Cells in Relation to Heredity. Dar
dM vege Seopa pp: 102-11. Cambridge, Cambridge Uni.

rsity Pre
1910. diseno oi. . Flora, 100: 398—446.
Sutton, W. S.
1902. On the Morphology of ees pe Group in Brachystola
magna. Biol. Bul., 4:
1903. The Chromosomes in koo "Biot Bul., 4: 231-251.
Tahara, M.
1910. Ueber die Kernteilung bei Morus. Bot. Mag., Tokyo, 24:
281-289

1910. Ueber die Zahl der Chromosomen von Crepis japonica, Benth.
Bot. Mag., Tokyo, 24: 23-28
opty M., and Ishikawa, M.
1911. The number of chromosomes of Crepis lanceolata var. platy-
phyllum. Bot. Mag., Tokyo, 25: 119-121
Tanaka, Y.
1913. A Study of Mendelian Factors in the Silkworm, Bombyz mori.
Journ. Col. Agr., Sapporo, 5: 91-113.

494 THE AMERICAN NATURALIST [ Vou. XLIX
Tennent, D, H.
1908. The Chromosomes in Cross-fertilized Echinoid Eggs. Biol. Bul.,
15: 127-134.
Tischler
Tii die Entwicklung m paee und y Sh ereriaay bei
Ribes-Hybriden. Jahrb. f. w ot.,
1908. carw an sterilen SAE Ee Nabi piy Z Zellforsch.,
: 3-151.

1910. Untersuehungen über die Entwicklung des Banana-Pollens.
h. f. Zellforsch., 5: 622-670.
Wager, H.
1904. The TEOR and goons Divi ision in the Root-apex of Phaseo-
Ann. Bot 29- (Number of chromosomes deter-
mined b White aes a study of Wager’s figures.)
Weismann, A
1892. Das Keimplasma. Jena.
White, O.
1913. The Bearing of Teratological epera in Nicotiana on
Theories of Heredity. Am. Nart., 52:
Wilson, E. B.
1904. ee! ag in Development and Inheritance. New York, Mac-

1901. Puao Studies in Cytology. A Cytological Study of
ra ae! Parthenogenesis in Sea Urchin Eggs. Arch. f.
. 12: 529-596,
1909. ane on Chromosomes, V. The Chromosomes of Metapodius.
A Contribution to the Hypothesis of bord Genetic Continuity of
Chromosomes, Journ. Exp. Zool., 7-205.
1910. The See ay omes in Relation to the Determination of Sex.
Science Progress, 5: 570-592.
1911. The ya Chromosome. Arch. mikr. Anatomie, 77: 249-271.
(Bibliography on accessory chromosome. )
Me Hans
1908 tue Parthenogenesis und Apogamie im Pflanzenreiche. Pro-
sus rei Botanice, 2: 293-454.
1910. Ueber die Nachkommenschaft der Solanum-Propfbastarde ane
die Chromosomenzahlen ihrer Keimzellen. Zeitschr. Bot.,
1-38.
Yamanouchi, S.
1908. Sporogenesis in Nephrodium. Bot. Gaz., 45: 1-30.
Ziegler, H. E.
1906. Die Chromosomen-Theorie der Vererbung in ihrer Anwendung
auf den Menschen. Arch. Rassen-Gesellsch. Biologie, 3: 797-
812.

Zoja, R.
1895. Sulla independenza della cromatina paterna et materna nel
nucleo delle cellule embrionali. Anat. Anz., 11: 289-293.

REGENERATION POSTERIORLY IN ENCHY-
TRÆUS ALBIDUS!

H. R. HUNT

THE primary object of the following experiments was
to determine whether Enchytreus albidus can regenerate
posteriorly, when cut at regions of the body varying from
near the posterior end to near the anterior end. Sec-
ondly, an attempt was made to compare the rates of re-
generation per day posteriorly at the different levels at
which the worms were cut in two.

No experiments have been published in which the ca-
pacity of this species to regenerate posteriorly has been
tested. Nusbaum (’02; ’04) studied the histological
processes in the regeneration of the Enchytreide ante-
_ riorly and posteriorly. He found that regeneration ante-
riorly does not take place as readily as regeneration pos-
teriorly, and that never more than two or three segments
regenerate anteriorly.

The animals used in the present experiments were col-
lected in abundance from the coarse gravel of the tidal
zone on the seashore at Cold Spring Harbor, Long Island,
New York. Six sets of experiments were conducted.
Each of the worms was cut into two pieces, the anterior
and the posterior pieces being preserved. The average
number of segments in this species is not far from sixty.
The regions selected for cutting were such as to give
fairly comprehensive data as to the regenerative capacity
posteriorly at different levels. In the first set of experi-
ments the worms were so cut as to leave only about eight
anterior segments; in the second set about sixteen anterior
segments; and in the third set about twenty anterior seg-
ments. In the fourth set the cut was made near the
middle of the worm; in the fifth about sixteen posterior

1 Contributions from the Zoological Laboratory of the Museum of Com-
parative Zoology at Harvard College, No. 260.

495

496 THE AMERICAN NATURALIST [ Vou, XLIX

segments were removed; and in the sixth eight posterior
segments. The worms were anesthetized with chloretone,
and the operation was performed under a dissecting mi-
eroscope. The pieces were placed in small sterilized glass
bottles, each containing a strip of filter paper and enough
sterilized sea water to keep the animals well moistened. -
Ten pieces of approximately the same length were kept
in a single bottle. Throughout the experiment the bot-
tles, each one stoppered with a cork, were kept in an ice
chest to restrict the growth of bacteria. The work was
begun early in July, 1913, and was continued until the
first of October. At the middle of August it became nec-
essary to carry away from the seashore the material then
living. After this, fresh water was used for moistening
the worms and cleaning out the bottles. The worms,
however, seemed to regenerate as well in the fresh-water
as in the salt-water environment. The analysis of the
results of the experiments was done in the zoological labo-
ratory of Harvard University.

It was found that the length of the regenerated seg-
ments, as compared with that of the segments in the adja-
cent unregenerated part of the worm, was a fairly accu-
rate criterion for determining the number of regenerated
segments. To test the accuracy of this criterion, parts
of eight worms consisting of the twenty most anterior
segments were allowed to regenerate for about eight
weeks. Having taken the precaution to determine accu-
rately the number of segments in each of the pieces at the
time of the operation, it was easy to determine how many
segments had regenerated, for of the total number of seg-
ments at the end of the experiment all except the original
twenty were, of course, regenerated segments. The re-
sult thus obtained was compared in each worm with that
obtained by counting in the same worm the number of
segments posterior to the point where there was an abrupt
change in the length of the segments, that point indicat-
ing the region of the cut. Table I gives the data for this
comparison. The results show that the method which

No. 584] REGENERATION IN ENCHYTRAEUS 497

was used to determine the number of regenerated seg-
ments is accurate to within one or two segments, for it
will be noted that the results by the two methods never
differ by more than two segments, usually by only one.
The worm’s body is so short that it was found impracti-
cable to secure exactly eight, sixteen, etc., segments in
every piece used in the whole series of experiments.

TABLE I
Number of Segments Regenerated :
As Determined by e

Number of the Worm Total Number Minus 20 Segment Length
1 6 rf
2 18 18
3 20 22
4 23 24
5 21 22
6 15 14
7 18 17
8 10 12

The results obtained in each of the six sets of experi-
ments have been condensed, for convenience, and are
shown in Table II. In the first vertical column of this
table the Roman numerals designate the number of the
set of experiments. The horizontal lines corresponding
to each of these sets give in succession, (1) the number of
segments in the pieces used in the experiments, (2) the
number of worms operated on, (3) the number that sur-
vived long enough to be observed, (4) the per cent of
worms that survived and were observed, (5) the period
during which the regeneration took place, (6) the number
of segments (0 to 24) regenerated by the surviving worms,
(7) the average number of segments regenerated in each
set of experiments, and (8) the mean rate of regeneration
per day of the worms in each set expressed in segments.
This mean rate of regeneration was obtained by first com-
puting the rate of regeneration per day (in segments)
for each worm in the set, and then averaging all the re-
sults. In some worms the number of segments regen-
erated was observed twice, several weeks elapsing be-

[ Vou. XLIX

THE AMERICAN NATURALIST

498

‘Axis moqe st Apod oy} ur syuəugəs JO LaqUINU OFBIBAG OY, T

oor | 889 °° THIOL
‘ ig 7 FTIR e o aO
#80 g'z ; ‘lltipig tp „ TE- ST SI oor 8 IA
SIT i : FETITELE ET CETE 3 hee
6c. v9 PAP : TP Sl Ee eee Pera tae » 8POF TE TS FOL 9T A
oe i A 3 NoT ELINT S €j ejs » TE-93
TSI clelt z Tell fl rb bel bbl a E8 eL
8Es" T6 BS ee a a Ba TILE (EEIE T Iei l a » Ori OF PP OTTI 0E AI
8'8 ews oe : PIPL TLE LSS AICi PIE een a oe! » Verse
297 SPT I A eas eee I TELE ET E SALICE TAEA » 6S-FP £g OL 0E OF III
gor izi = mamrna naa er o
LOZ" v6 ze er : LECEEPELCIS Syl rere ne a n: 36-76 VS TE 661 FP II
6z Ss Sk (ake lelelelefellelrlelslgl-lelpig) . ezz
T6 -a : Thea ite lee eg
OFT o'g reike i ar edeler elie] ssp ge 9 9 g6 ze I
om) s}aeu
sjuemeg | syuemBag | Fé | S| 6c SI | oI de ue: 6) 8i/L/9191F ZITO uorwaaasqO Ta ory cane oe Basin io
at etl don JO popoq (9384099 suo M | STUTO M ey : Bret
youg 8} PZ J3 Sox S[VNPTATpUy Jo səquny “d | JOON | J °ON | 50 -oN | Jo “ON
Il Wiavib
a

No. 584] REGENERATION IN ENCHYTRAEUS 499

tween the two observations, so that the total number of
observations recorded in any one set of experiments may
be larger than the number of worms observed in the
same set.

The results of these experiments are summarized in the
graph shown in Fig. 1, where the rates of regeneration
per day (expressed in hundredths of the length of one
segment) are measured on the axis of ordinates (Y),
and the length (in number of segments) of the pieces that
produced the regenerated parts, are measured on the axis
of abscisse (X). Since sixty is about the average num-
ber of segments in this species, that is the value which
has been used in plotting the curve. A mathematical
analysis of the rates of regeneration at the different levels
shows that the difference in the mean rates of regenera-
tion at any two successive levels is significant. But the
temperature of the worms was not carefully controlled,
and the periods during which the wounds were healing
and the worms preparing to form new segments were in-
cluded in the computation of the mean rates of regenera-
tion. Therefore, the ratio between the rates of regenera-
tion, as here computed, at any two of the six levels only
approximates the ratio which would have been obtained
between the rates at these two levels by subjecting all the
worms to the same temperature conditions and by using
in the computation of the mean rates of regeneration only
the periods during which the segments were being formed.
The curve suggests, however, that the rate of regenera-
tion for the posterior half of the body is proportional, or
nearly so, to the number of segments removed. Anterior
to the twentieth segment the rate of regeneration de-
creases. May we not have here a curve depending on two
opposing sets of factors; one which tends to increase the
rate of regeneration as more segments are removed, the
other to decrease the rate? In the latter set of factors
the amount of available building material may be the most
important element.

500 THE AMERICAN NATURALIST [ Vor. XLIX

The worms seemed to regenerate equally well in a frei
water or in a salt-water environment. Thirty-one of the
one hundred and sixty surviving worms lived for about
forty days in a fresh-water environment and regenerated.
Twenty-six worms from which the sixteen posterior seg-

8 i6 2 30 44 52

Fic, 1. Curve showing the daily seme of pon Bae by pieces of six dif-
ferent lengths. The unit selected to measure the mean rate of regeneration at
each of the six levels was 1/100 of a catia ent, nent that oh to measure nse
lengths of the pieces which produced the regenerated segments was one segment

ent (on axis Y) was the same as that chosen = EE one segment (axis
of X) of the pieces mea nth yp reniawnee par

ments had been removed, and twenty-six others from
which the posterior half had been removed, regenerated
almost contemporaneously for about thirty days in the
same ice chest, and in a salt-water environment. Later
in the season in a different ice chest eighteen worms from
which the sixteen posterior segments had been removed,
and thirteen from which the posterior halves had been re-
moved, regenerated contemporaneously for about forty
days in a fresh-water environment. When the sixteen
posterior segments were removed the rate of regeneration
in the salt-water environment was 0.02 segments per day
less than in the fresh-water environment, while when the
posterior halves were removed the rate of regeneration
in the salt-water was 0.07 segments per day greater than in

No. 584] REGENERATION IN ENCHYTRAUS 501

the fresh-water surroundings. These facts show that
the worms regenerate in both fresh and salt water. This
is not surprising, since individuals of this species are
normally found both on the seashore, where they live in a
salt-water environment, and also in earth moistened with
fresh water. Furthermore, with the exception of the sa-

Fic. 2, Camera lucida iawii of the posterior end of a normal worm. Magni-
fied 17 diameters.

Fig, 3. Camera lucida drawing pags the pe regenerated posterior a ho a worm,
The region posterior to X is regenerated gnified 17 diam
Fie. 4. Sketch of a Aa double tail, pero about 17 pagia EA
linity of the water used to moisten the worms, the worms
which regenerated in the fresh-water surroundings were
probably subjected to about the same conditions as those
which regenerated in the salt-water. Therefore, the sa-
linity of the water in the environment does not seem to
affect the rate of regeneration. The data used in plotting

502 THE AMERICAN NATURALIST [Vou. XLIX

the curve shown in Fig. 1 were secured from worms which
regenerated in the fresh-water, as well as from those
which regenerated in the salt-water, environment. The
above observations make it seem probable, therefore, that.
the form of the curve does not differ fundamentally from
the form which it would have had if all the worms had
regenerated in salt-water surroundings.

In. Fig. 2 is shown the normal appearance of the ven-
tral aspect of the posterior end of a worm in which there
has been no regeneration. It will be noticed that the
length of the segments gradually decreases toward the
posterior end; but in Fig. 3, which is a camera lucida.
drawing of the posterior portion of one of the regen-
erated worms, the length of the segments decreases ab-
ruptly at the point X, showing that to be the point at
which the tail was removed.

Three worms from which eight posterior segments were
removed regenerated double tails. Morgan (’97) and
Michel (’98) observed the same phenomena in Allolobo-
phora fetida. One of these worms is shown in Fig. 4.

Some attempts were made to determine the rate of
regeneration anteriorly at different levels on the worm’s
body. At present all that can be said is that regeneration
posteriorly takes place much more frequently and rapidly
than anteriorly.

The conclusions that follow from these experiments are:

1. Enchytreus albidus regenerates posteriorly when
cut off at any level between eight segments from the pos-
terior end of the body and eight segments from the an-
terior end. It will be noticed that although the mortality
in pieces containing only the eight most anterior segments.
was about 94 per cent., yet those that did survive regen-
erated from three to eleven (on the average seven) seg-
ments. In other words, a piece from the extreme anterior
end, containing only one eighth the number of segments.
in the whole worm, can regenerate nearly as many seg-
ments, on the average, as it had at the beginning of the
experiment. Morgan (’97) found that in Allolobophora:

No. 584] REGENERATION IN ENCHYTRZUS 503

fetida anterior pieces of less than thirteen segments
rarely, if ever, regenerate posteriorly. In Enchytreus
the anterior limit of the capacity to regenerate posteriorly
was not found.

2. The rate of regeneration seems to increase from the
posterior end of the worm up to its middle almost in di-
rect proportion to the number of segments removed. An-
terior to about the twentieth segment the rate decreases.

3. Regeneration can take place either in a fresh-water
or in a salt-water environment. Also, the salinity of the
water seems to have little or no effect upon the rate of
regeneration.

4. Double tails can be regenerated when the eight
most posterior segments are removed.

5. Regeneration posteriorly takes place more readily
than it does anteriorly.

I am indebted to Professor C. B. Davenport for pro-
posing the problem and for making many helpful sugges-
tions. I also wish to express my gratitude to Professor
E. L. Mark and to Professor H. W. Rand for corrections
and suggestions in the preparation of the manuscript.

REFERENCES CITED
Michel, A.
1898. Recherches sur la Régénération chez les Annélides. Bull. sci.
France et Belgique, Tome 31, pp. 245-420, pl. 13-19.
Morgan, T. H.
1897. Regeneration in Allolobophora fetida, means = Entwick-
lungsmechanik, Bd. 5, Heft 3, pp. 570-586,
Nusbaum,
1902. Veber ks morphologischen Vorgänge bei der Regeneration des
lich abgetragenen hinteren Körperabschnittes bei Enchy-
seo Arch. polonaises Sci, biolog. et med., Tome 1, pp,
292-347, (Cited from Zoologischer Jahresbericht, 1902.) Pre-
preii account in Biologisches Centralblatt, 1902, Bd. 22, pp.
—298.

1904. Veber ni Regeneration des Vordertheiles si D
körpers nach einem künstlichen Operation. lonaises
8. Diolog. e t Med., Tome 2, pp, 233-258. ovis frum Zoolo-

gischer Jahresbericht, 1904.)

THE ORIGIN OF BILATERALITY IN
VERTEBRATES:

Proressor A. C. EYCLESHYMER

DEPARTMENT OF ANATOMY, UNIVERSITY OF ILLINOIS

Many attempts have been made to determine how early
in development the vertebrate egg becomes bilaterally
symmetrical. The conclusions have been as varied as the
attempts.

Before the subject can be discussed it is necessary to
consider two fundamental propositions. The first is that
there exists an active pole in the egg, and the second is
that the anterior end of the embryo develops in this
region, or at least in the active hemisphere.

The active pole is indicated at an early period by cer-
tain phenomena, such as secretory activity, accelerated
yolk metabolism, formation of pigment, position of
nucleus, expulsion of polar bodies, ete. Hatschek says
that ‘‘it is probable that a polar differentiation is present
in the unfertilized ova of all the metazoa, through which
the most active and least active poles can be determined.’’
Whether or not Hatschek’s statement be true, it is certain
that if the area in which cleavage grooves first appear
be traced backward a differentiation in this area can
be found in a very early stage. We are thus enabled to
speak of an active pole and an opposite inactive pole. A
line passing through the two is designated as the primary
ovic axis.

That the active pole or hemisphere gives rise to the
embryo was first pointed out by Jan. Swammerdam
in his ‘‘Bibel der Natur.’’ This view was later supported
by Prevost and Dumas, von Baer, Reichert, Cramer, New-
port and others. Pfliiger, however, believed that the
greater portion of the embryo was formed from the in-
active hemisphere and his view was supported by Roux,
O. Hertwig and others. Most of the later investigators

1 With observations by C. O. Whitman on Bufo.

504

No. 584] BILATERALITY IN VERTEBRATES 505

including Morgan and Tsuda, Assheton, H. V. Wilson,
King, Smith and others have generally agreed that the
head end of the embryo forms from the active hemi-
sphere and the caudal portion from the inactive. My
own experiments on a considerable number of Amphibia
have led to the conclusion that the head of the embryo
forms from material which lies at, or near, the active pole
of theegg. It thus seems fair to assume that the cephalic
portion of the embryo is formed from the active hemi-
sphere.

As stated there have been many attempts to deter-
mine how early in development the egg shows bilateral
symmetry. Some claim bilateralism for the primitive
ovum. Others hold that this condition is not present
from the first, but originates at some later period. This
period may precede or follow the deposition of the egg.
Those who regard the egg as bilaterally symmetrical
before deposition claim that this is manifested either
through an excentric position of the egg nucleus, or an
excentric pigmentation. Those who regard it as fixed
after deposition are not in accord. By some the path of
the spermatozoon is considered as the determining factor,
by others the first or second cleavage groove, and by still
others areas of accelerated segmentation.

The assumption that the egg is bilaterally symmetrical
from the beginning is based upon nothing more than
plausible hypothesis and naturally falls beyond the range
of experimental proo

Some (Schultze) hold that the excentric position of the
egg nucleus together with the primary ovic axis deter-
mine bilaterality. The work by Roux, Jordan and others,
shows that this is highly improbable.

Others (Roux, Morgan and Tsuda) maintain that the
excentric arrangement of pigment enables one to deter-
mine bilaterality. Professor Whitman’s observations
which are recorded in a later paragraph, together with
his drawings, indicate that the arrangement of the pig-
ment is of significance in Bufo. The observations of
Moskowski on Rana, Morgan’s later observations on

506 THE AMERICAN NATURALIST [ Vou, XLIX

Bufo, together with my own on Amblystoma, have thrown
doubt upon this conclusion.

Still others (Newport, Roux) believe that the path of
the entering spermatozoon and the primary ovic axis
determine bilaterality. Jordan has shown that this view
is untenable for Diemyctylus. Professor Whitman’s ob-
servations, recorded in a later paragraph, show that this
is not true in Bufo.

Thus each of these assumptions has been met by
serious objections.

The idea that the first plane of cleavage determines
the axis of the embryo was expressed as early as 1853 by
Newport in the following words:

I have long been aware that the axis of the embryo was in the line

of the first cleft of the yolk.
From a series of experiments on the frog’s egg Roux
came to the conclusion that the first cleavage plane coin-
cides with the median sagittal plane of the embryo. In
the same year Pfliiger reached the same conclusion. Sup-
ported by these eminent investigators the theory was very
generally accepted. In working over the same field
Rauber found that in the axolotl and frog the median
plane of the embryo coincided with the second cleavage
groove instead of the first. Shortly after the publica-
tion of Rauber’s work, O. Hertwig working on the egg
of Triton confirmed the observations of Rauber. In 1892
Roux modified his earlier view and stated that the second
groove as well as the first often coincided with the median
plane of the embryo. ;

In the following April the writer found from a series
of puncture experiments on the egg of Amblystoma that
exovates on opposite sides of the first cleavage groove
were later found on one side of the embryo. The conclu-
sion was that in these cases the first cleavage groove did
not separate the right and left halves of the embryo.

In 1893 Jordan and the writer reviewed the experi-
ments up to this date. We found that even in the de-
scriptions and figures given by Newport, Roux, Rauber,
there was evidence sufficient to show that the median

No. 584] BILATERALITY IN VERTEBRATES 507

plane of the embryo often deviated widely from the first
or second cleavage planes. We accordingly undertook
ansextended series of observations on the living segment-
ing eggs of Amblystoma, Diemyctylus, Rana and Bufo.
Our conclusions were as follows:

The first and second cleavage planes undergo, even in the earlier
stages, extensive torsion. Everything indicates that the extent of this
shifting increases greatly in later stages. This led us to conclude that
the earlier cleavage planes and the embryonie axes have no vital con-
nection and that the coincidence where it exists is of no fundamental
significance.

The later observations by Grönroos, v. Ebner, Morgan
and Tsuda, Kopsch and others have likewise emphasized
the significance of these variations.

It is scarcely necessary to state that if these cleavage
planes mark embryonic areas, the amount of material
set apart in different eggs for similar parts of their re-
spective embryos, must be exceedingly variable, and these
excesses and deficiencies must be corrected by a corre-
sponding retarded or accelerated growth until the norm
is reached, but there is not the slightest evidence that
such corrections occur.

These wide variations have been repeatedly observed
not only in various amphibia but also in practically all
classes of vertebrates: in Amphioxus by Wilson; in
Petromyson by McClure, Kupffer, Eycleshymer; in Dip-
noans by Semon; in Ganoids by Salensky, Dean, Whit-
man and Eycleshymer; in Teleosts by Coste, Hoffmann,
His, Agassiz and Whitman, Kingsley and Conn, Clapp,
Sobotta and others; in Reptiles by Agassiz and Clark,
Oppel, Sarasin; in Aves by Coste, Koelliker, Kionka ;
in Mammals by Duval, v. Beneden, Assheton, Sobotta
and many others.

The inevitable conclusion from such a mass of evi-
dence can not be other than that neither the position or
direction of cleavage grooves has the slightest signifi-
cance as far as the setting apart of definite embryonic
areas is concerned.

If then it may be considered an established fact that

508 THE AMERICAN NATURALIST [ Vou. XLIX

neither the position nor the direction of the cleavage
grooves enables one to predict the long axis of the em-
bryo, we are naturally led to look for other phenemena
which may be of significance. As stated in an earlier
paragraph my experiments showed that the head end of
the embryo is formed at, or very near, the active pole,
and since this area is the one in which cell division is
most rapid, it was concluded that the anterior end of the
embryo, which is the first to differentiate, was indicated
by this increased cellular activity. I accordingly stated
that an area of increased cellular activity indicates the
position of the head end of the embryo. As is well
known, this area can be located with the advent of the
first cleavage groove.

While the head end of the embryo may thus be readily
located, the median plane of the body may lie in any one
of an indefinite number of meridians. The question
which now arises is which one of these meridians will
represent the median plane of the future embryo.

The writer’s studies on Rana, Bufo, Acris, Ambly-
stoma, Necturus have shown that in another portion of
the egg there is an area of smaller cells, and that this
area of smaller cells always marked the region of the
forthcoming blastopore. The blastopore in turn defi-
nitely fixes the posterior portion of the embryo.

With the recognition of these areas of accelerated cel-
lular activity, the one at the active pole, indicating the
position of the future head of the embryo, the other at
the side of the egg, indicating the position of the forth-
coming blastopore, it necessarily follows that the median
plane of the embryo must coincide with a line passing
through the centers of the two.

When these observations were first published in 1898,
many questioned the existence of such a secondary area
of cellular activity. Yet a search through the literature
showed that such an area had been observed in many
groups of vertebrates. Lwoff found such an area at the
posterior end of the embryo of Amphioxus. The figures
of the segmenting blastodises of Elasmobranchs, given

No. 584] BILATERALITY IN VERTEBRATES 509

by Balfour, Riickert, Gerbe and Sobotta all show that in
these forms such an area is present. In the Reptilia,
Vay’s studies on Tropidonotus show that an area of small
cells represents the posterior end of the embryo. v.
Koelliker first called attention to such an area in the
blastodise of the chick and suggested that it determines
the position of the posterior end of the embryo. The
later investigations of Duval and Kionka leave no doubt
as to the frequent and probably constant appearance of
this area in the locality which later becomes the posterior
end of the embryo.

In 1904 the writer made a study of the egg of Necturus,
which from its size is especially favorable for surface
study. This work was undertaken with a view of ascer-
taining how early this secondary area could be located.
It was found that as early as the fourth or fifth cleavage,
the cells on one side began to divide more rapidly than
any others, excepting those of the primary area. It was
possible to predict in this form the median plane of the
forthcoming embryo at an extremely early stage of
cleavage.

The following year de Bussy from his studies on the
Japanese Cryptobranchus emphasized the fact that he
could find no secondary area of accelerated cell division
such as had been described by the present writer. Yet
Smith working on the American Cryptobranchus says
that he finds ‘‘an accelerated cell division about a radius
ef the blastodise which gives a condition of bilateral sym-
metry.

The writer felt that it was scarcely necessary to follow
the subject further and should not have rehearsed the
findings had it not been that certain material came into
his hands last year which bears directly upon this sub-
ject. This material consists of unpublished descriptions
and drawings made by the late Professor C. O. Whitman
in June, 1894. These were turned over to me by the de-
partment of zoology of the University of Chicago. Pro-
fessor Whitman ’s notes run as follows:

510 THE AMERICAN NATURALIST [Vou. XLIX

Hitherto we have obtained eggs the first week in June. This year
we could find none until July 1. We had several night rains, enough to
flood the low ground behind Breakwater Hotel. On the evening of
June 30, the day after the rain fell copiously, the toads swarmed in this
place, and had a carnival of noise; the whole place rang with so many
voices as to be almost deafening. On the morning of July 1, we found
a great many eggs. The following night the singing followed but much
reduced, and only two pairs of toads were captured. The next night
the water had gone except in one of the ditches and no toads were to
be heard, and of course no eggs. It would seem that rains stimulate
them to lay; and the lateness of the season may have been the reason
that the egg-laying was confined almost entirely to a single night.

The unfertilized eggs are by some said to be unoriented, that is they
are said to be unable to take the normal position assumed by the fer-
tilized egg. The sperm is supposed by some observers to mark the first
plane of division and to give the egg the power to right itself. I find
that it is not true that the eggs will lie just as they happen to fall,
although they do so more nearly before fertilization than after it. I
an egg be separated from the rest and turned about for some moments
with needles, so as to loosen its adhesion to the membrane, and then
rolled to one side so that the equator is vertical, one observes that it
slowly turns and in the course of a minute, or sooner, it takes the
normal position with the blacker pole uppermost and the whiter show-
ing a little on one side, when viewed from above. This was repeated
several times and on several eggs with the same result. The motion is so
slow that one does not notice it until after the lapse of some seconds.

I cannot affirm that all unfertilized eggs will right themselves; ordi-
narily they do not if left to themselves. They assume an irregular
wrinkled appearance and have so little power of righting that they stick
to the membrane enough to prevent it. When fertilized they contract
and round up and get freedom of space to move in, The entrance of
the sperm evidently increases the disproportions between the weights of
the upper and lower pole. The upper pole becomes lighter and the egg
rights itself more readily and quickly. The orientation of the egg is
complete before fertilization.

e eggs which are in the stage of first cleavage there is a small
depression which I have found by examination of earlier stages is the
“fovea germinativa” of Max Schultze, or the “fosette germinative ”
of Bambeke. I find further after fertilization, a second point or de-
pression, which probably is the place of penetration of the spermato-
zoon. The fovea marks the upper pole, but is not placed at the middle
of the upper hemisphere; it is excentric.

I followed two eggs which showed both the fovea and the spermatic
dent. In neither did the first cleavage plane pass through this dent.
In one case it passed far from it while the second cleavage passed near
to it. In another ease the dent is in the middle of one of the first four

No. 584] BILATERALITY IN VERTEBRATES 511

cells, and on the darker side of the upper hemisphere. If this be the
sperm track it does not determine the median plane of the embryo.

CLEAVAGE

The eggs were obtained in the two-cell and four-cell stages. At this
time the pigment is excentric, falling a little short of the equator on
the one side and a little beyond it on the opposite. [The notes nowhere
state that the antero-posterior direction of the embryo is indicated by
the distribution of pigment, yet I think an examination of the figures
ean not fail to convince all that their interpretation can not be other-
wise.—A. C. E.] When the first cleavage groove runs in the plane «f
symmetry the second cleavage grooves are at right angles and appear
at about the same time in both halves as shown in Figs. 3 and 4. When |
the first cleavage groove is transverse to the plane of symmetry the
second cleavage grooves do not appear at the same time, but the one

1
upper ais on the blastoporie (posterior) side, but leaves considerable
below the upper cells on the opposite (anterior) side. e secon
equatorial usually cuts off all the pigmented cells on the anterior side
of the egg and non-pigmented cells on the posterior (blastoporic) side.

The blastomeres on the posterior (blastoporic) side are smaller than
on the anterior side, from the very first. It is the blastoporie side that
takes the lead in division and the cells are smaller here all the way up
to the time when the blastopore appears.

It is thus obvious that the findings by Professor Whit-
man not only lend confirmation to my observations on
bilaterality, but that they in reality anticipate them.

It may be said with added confidence that bilaterality
in the vertebrate egg is revealed through the early cleav-
age grooves. The cephalic portion of the embryo is in-
dicated by the area in which cleavage grooves first appear
and in which cellular division is most rapid. The caudal
portion is indicated by a secondary area of cellular activ-
ity in the blastoporic region. These two areas pass into
each other constituting an embryonic tract.

In addition to the above observations, Professor Whit-
man’s manuscript and drawings give the results of a
series of puncture experiments in the blastoporie lip.
Since these observations have an important bearing on
the question of epiboly, emboly and concrescence, they
are appended,

512 THE AMERICAN NATURALIST [ Vou, XLIX

ha

&9
9
dg

No. 584] BILATERALITY IN VERTEBRATES 513

EXPERIMENTS

On June 5, 1894, sixteen eggs in the thirty-two cell stage were punc-
tured at the equator, in the middle of the white cells, as shown in Fig.
11. In twelve the blastopore appeared near the puncture as shown
in the accompanying cut. The extraovates were found in the positions
shown in A, 1-12, at 10:00 a.m. the next morning. The variations in
positions are doubtless due to my punctures falling at different points,
sometimes hitting as in Fig. 12, at other times in the very edge of the
pigment. In the four remaining eggs two showed no extraovate and
two showed no blastopore.

On June 4, 1894, pricked egg B at middle of lower pole, soon after
the blastopore was sharply marked on the side of the embryo. Ven-
trally this outline was not clearly marked. At 4:30 this blastopore was
outlined all around and nearly circular or about 1% diameter observed

at 3:00. At 6:30 the blastopore was far advanced and nearly circular.
At 8:30 it was nearly closed. It will be noted that the extraovate re-
mained central throughout.

Another egg C was punctured in the ventral edge of the blastoporie
rim, and the extraovate was carried along by the closing blastopore.
I ought to have made two punctures, one in the middle as well, so that
this approach could have been seen. However, my notes show that the
blastopore advanced evenly. In this case the extraovate is carried

point, approached by the blastopore from the opposite side.

June I pricked a number of eggs in the early cleavage stages
(8-64 cells) at lower pole. In most of these eggs the extraovates were
found after two to three hours to lie at or near the equator of the egg.
This was long before the appearance of the blastopore. The extra-
ovate has evidently moved and if one should leave the egg until the
blastopore appeared and then look at it, it might be found at the middle
of the body; and thus it might appear as if the embryo had lengthened
across the lower pole (Roux). Sometimes extraovates have moved and
the punctures healed.

514 THE AMERICAN NATURALIST [ Von. XLIX

No. 584] BILATERALITY IN VERTEBRATES 515

EXPLANATION OF PLATES
Since no explanation of the figures could be found
other than those included in the preceding pages, I have
endeavored to give an explanation in accord with the
text. It should be remembered however that the figures
may be open to other interpretations than those pre-
sented. The figures show the distribution of pigment
and the relation of the embryo and the cleavage planes
to the pigment. It will be noted that the eggs when
viewed from above show a lighter area or crescent on
one side. This excentric position of the pigment is like:
wise well shown in profile. The arrows in all cases show

the direction of the forthcoming embryo.

Shows the upper hemisphere of an egg in which e embryonic
axis is indicated by a line passing through the centers of the noe erescent and
the e deeply pigmented area. In this case the first cleavage plane (I)
passed at right angles to the embryonic axis. It is of interest to pbk that the
second cleavage (II) has appeared in that portion of the egg nearest the og y
crescent and prera Agere it coincides with the median plane of sha mbry'

Fig. 2. Sho profile view of an egg in which the first cleavage s
with tie recog piane rae the embryo while the second is at right angles to the
sam loss to understand the extent of the arrow in this and the
MEER a pineg views. It may be ea bata Ww hitman intended thus to
indicate the limits hee the embryonic

Fic. 3. Shows the upper ae ct an egg in the four cell stage. In
this case ae median plane of the forthcoming embryo coincides with the first
cleava i

s the upper hemisphere of an egg consisting of eight cells.

It is fi be noted that the formation of the first equatorial sharply ego cer the
prs shies our vege stag of the egg on the one side but not on the opposite
In t case the median plane of the embryo cotachiak with either the

groove.
1G. 5. Represents the profile view of either the same egg or another egg
in the same stage. In this case the differences in the distribution of the pigment
are again shown.
Fic. 6. Shows the upper hemisphere of an egg in koria TE the sa
nor the second cleavage grooves coincide with the median
Fie. 7. Shows the upper hemisphere of an egg at a Se when ot scare
cleavage grooves are present. It is impossible to say whether the median T
of the embryo coincides with either the first or the second cleavages.
pearance of the cleavage grooves leads me to infer that the direction res 68
arrow is parallel with either the first or pt second,
Fic. 8. Represents a profile view Aui api the same egg or another egg in the

Fig. 9. agga the a hemisphere of another egg in which the fourth
cleavage grooves are present. In egg the median plane of the embryo
coincides with bee eg or second Psa groove.

516 THE AMERICAN NATURALIST [Vou. XLIX

No. 584] BILATERALITY IN VERTEBRATES 517

Fic. 10. Shows the upper hemisphere of an egg in a later stage of cleav-
age. It should be noted that the lines representing the primary grooves are
entirely pra by a “ee of the blastomeres.

Fie epresents a profile view of the same (?) egg, viewed from the

e on wh i the blastopore is forthcoming. The small crossed lines represent

the localities in which Professor Whitman punct the of this stage.

nts a profile of the opposite side of the same (?

the u hemisphere of an eg a later stage of cleav-
age. It should again be glee that it would be impossible to

two grooves it must be setae irregular.

Fic. 14. Represents a profile view of the same (?) egg viewed from the
side in which the blastopore will por appear. On this side cell division is
decidedly in advance of the opposite
Fig. Represents a profile view Ne "the opposite side of the same (?) egg.

10. 1

Represents a profile view = an Pea in roy segmentation. The
side ph Niak the Trea will appear indicated not only = the dis-
Oe ae x pigment but also by a ESE paaien in cell divisi

b Neurite a gana view of an egg at the time when ane pE SE
appears. on figure shows that it appears on the side of the egg which is least
pigmen

SHORTER ARTICLES AND DISCUSSIONS
THE TORTOISESHELL CAT

In The Journal of Genetics (June, 1913), Doncaster has sum-
marized genetic data dealing with the tortoiseshell cat. The
records are collected from fancy breeders and from the work of
Dr. C. C. Little

Aside from certain disputed points the inheritance is in ac-
cordance with simple sex-linkage and is analogous to the human
defeects—color-blindness, night-blindness, nystagmus, and hemo-
philia, and to the thirty or more sex-linked factors of Drosophila.

If the factor for yellow be represented by Y and its allelo-
morph, the factor for black, by B, the lack of either by b, the
sex factor by X, and the allelomorph of X by x, the normal
zygotic possibilities are as follows: YX—bx=—yellow male.
BX —bx=black male. YX—YX=yellow female. BX—
BX = black female. YX—BX-=tortoiseshell female.

It is obvious then that there can be but two classes of males,
while there are three classes of females. Difficulties arise when
it is attempted to explain the occurrence of black females pro-
duced either by the mating of a black female to a yellow male
which should give only tortoiseshell females and black males.
or by the mating of a tortoiseshell female to a yellow male,
which should give only tortoiseshell and yellow females and
black and yellow males. The occurrence of the rare tortoiseshell
male is also the cause of considerable difficulty. In one mating
out of seventeen of yellow females to yellow males there were
produced three tortoiseshell females. There are recorded in
addition from the seventeen matings forty yellow females and
forty-eight yellow males which are in agreement with expectation.

In order to explain these discrepancies it is suggested that
possibly the linkage of Y with X is not absolute. Yellow males
may then produce gametes bX and Yx in addition to the normal
or more frequent gametes YX and bx. Gamete bX is female
determining, while gamete Yx is male determining and yellow
bearing. The latter gamete should produce a tortoiseshell male
when it meets an egg BX.

On this hypothesis we should expect the tortoiseshell males to
be as frequent as the anomalous black females from yellow
fathers. From the matings recorded there are eighteen anoma-
lous black females and only three tortoiseshell males, and one
of these tortoiseshell males had a black father. There is a fur-

518

No. 584] SHORTER ARTICLES AND DISCUSSIONS 519

ther objection to this hypothesis inasmuch as it is not explained
how gamete bX differs from BX. Doncaster admits these diff-
culties, stating that further work is necessary before a definite
conclusion can be reached.

In a more recent paper! Doncaster has suggested non-disjunc-
tion of the sex-chromosomes in oogenesis as a possible explana-
tion. This explains the matroclinous black females, but fails to
account for the lack of an equal number of patroclinous yellow
males. It also fails to account for the tortoiseshell male and
the occurrence of tortoiseshell females among the offspring of
yellow by yellow.

In a series of experiments begun upon cats at the University
of Pennsylvania during the last year, the tortoiseshell problem
has been especially investigated. A yellow Persian male was
crossed with common cats—black, maltese and tabby. The re-
sults, although not at present extensive, are sufficient to explain,
at least in part, the anomalies observed, and to suggest a simple
explanation for the occurrence of unexpected classes.

When the yellow male was crossed with a maltese female, a
maltese male and two blue and cream females were produced.
The blue and cream is the maltese or dilute tortoiseshell. When
mated to a black female the yellow male produced both dark and
dilute kittens. This shows that the black female was hetero-
zygous for dilution. Two of the males were black and two mal-
tese. The two females were dark tortoiseshell. When the yel-
low male was crossed with a dark tabby, there were produced
dark and light tabbies and maltese. Blacks are also to be ex-
pected from this mating. The mother is evidently hybrid be-
tween tabby and black and between black and maltese. The
female offspring showed yellow: the male offspring were without
yellow except for tabby striping.

The female offspring obtained from these matings may be ar-
ranged in a series, ranging from one that is predominantly
yellow to one that is maltese except for a few cream-colored hairs,
The maltese with a few cream hairs occurred in the litter of
three above mentioned, which included also a maltese male and a
maltese female with a small cream patch.

It may be readily understood how a maltese cat with a few cream
hairs or its intense form, a black cat with a few yellow hairs, would
be recorded as maltese or black, and it is reasonable to suppose
that further segregation of distribution factors in the direction
of black would have produced a fully black female. This may

1 Quarterly Journal of Microscopical Science, February, 1914.

520 THE AMERICAN NATURALIST [ Vou. XLIX

be compared with conditions in the guinea-pig in which yellow
spotting is continuous with total black. The essential differ-
ences are that in the cat we have a factor for yellow allelomor-
phic to a factor for black, that these allelomorphs are sex-linked,
and that either alone is sufficient to produce its expected color,
but that when one is balanced against the other, as in the tortoise-
shell female, other factors governing the relative amounts of the
two colors can act and produce continuous variation from yellow
to black.

The three tortoiseshell females from the mating of yellow by
yellow may be explained by supposing that the mother was
gametically a tortoiseshell plus a sum of yellow extension factors
and minus a sum of black extension factors.

The occurrence of the tabby factor brings in a restriction of
the black pigmentation producing yellow stripes. It is there-
fore much more difficult to distinguish a tabby from a tabby-
tortoiseshell than a black from a tortoiseshell. We have had a
few tabby-tortoiseshells that would have been recorded as tab-
bies if close examination had not been made.

Another source of error in records involving the tortoiseshell
pattern may be introduced by the occurrence of white spots.
Doneaster makes no mention of these in his paper, so that it is
possible that they did not occur in the animals recorded. In
what is genetically a tortoiseshell and white cat the incidence
of the white spotting may happen to be at just those points
which would otherwise be yellow. Thus the occurrence of black
and white daughters from yellow males may be explained. It
is possible also that the yellow mother of the three tortoiseshell
kittens recorded from the mating of yellow by yellow may have
been white at points which, if pigmented, would have been black.
She would then have been genetically a tortoiseshell and white
and some tortoiseshell kittens would have been expected.

I would suggest as a plausible hypothesis that the rare tor-
toiseshell male is genetically a yellow with an extreme of black
extension factors or a black with an extreme of yellow extension
factors. This hypothesis is rendered more probable by some
slight evidence showing that male tortoiseshells breed like
yellows.

There is then no need for assuming in the cat either breaks
in sex-linkage or non-disjunction of the sex chromosomes in
oogenesis. PHINEAS W. WHITING

UNIVERSITY OF PENNSYLVANIA

The American

Naturalist

intended for EREN and books, etc., intended for review should be

MSS.
sent go agen of THE AMERICAN oo RAL

articles containing summ

IST, Garrison-on-Hud dson, New York.
aries of research work bearing on the

problems of oe evolution are especially welcome, and will be given preference

in —

red reprints of nia a are supplied to authors free of charge.

oe
Further p arai will be supplied at c
Subscription

s and adv sarina eero be sent to the publishers.

subscription price is four dollars a
anadian postage twenty-five

a yea
cents oiditional.

The
oreign postage e is fifty cents and
The c a for single copies is

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

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, cae 2, 1908, at the Post pa at Lancaster, Pa., under the Aet ot
ongress of March 3

FOR SALE
ARCTIC, ICELAND and GREENLAND
BIRDS’ SKIN
Well dees Low Prices
articulars of
G. DINESEN, Bird Collecto
Husavik, North Iceland, Via Leidle, England

JAPAN NATURAL HISTORY plore
erfeot Condition and Lowest

Specialty: Bird Skins, Oology, Entom fesse Marine

Animals and others. Catalogue free, Correspond-

ence solici

T. FUKAI, soo

dapsone Japan

ATES

WANTED
BIRDS OF AMERICA by J. J. Audubon, 7 volumes,
please report cash price, stating condition, binding and dates
of volumes.
F. C. HARRIS,
Box 2244 Boston, Massachusetts

The University of Chicago

ers instruction e Su
met E a þasis aal
cade yan a ers of the

ce Summer Quarter, 1915
sc gone a ar mele
Detailed announcements will gi

sent upon application.

Marine Biological Laboratory
Woods Hole, Mass.

— for research in
gaiki E

INSTRUCTION ©
July—August

SUPPLY
DEPARTMENT
Open the Entire Year

GEO. M. GRAY, Curator, Woods Hole, Mass
The annual announcement will be snt on npplioation 1
The Director, Marine Biological Woods
Mazs-

THE
AMERICAN NATURALIST

VoL. X LIX. September, 1915 No. 585

A STUDY OF ASYMMETRY, AS DEVELOPED IN
THE GENERA AND FAMILIES OF
RECENT CRINOIDS

AUSTIN H. CLARK

PHOLSCG yok os coh tenes Me Cae oe es a he ee fe eH a os ws 521
The Different Types of as AMOS 6 9 wae tarse INESS sc : 523
The Asymmetrical Crinoids ..........-+. 22 esseenes WE D Se ee 524
The Phylogenetic iisa of Moy tishatty elie pepe mec ea Vee oe 526
The Geographical Distribution of Asymmetry ....................-- 527
Bathymetrical Distribution of the Asymmetrical Crinoids ............ 530
Thermal Distribution of the Asymmetrical Crinoids ................. 535
The Asymmetrical Features in Detail ..... 250.202. secs ces cces sae 538
Jamar a eee BN eRe EA a Eh Sp i ew NA 546
PREFACE

In the animal kingdom there are few, if any, forms
which can be properly described as perfectly symmetrical,
either from a bilateral or a radial standard. We have,
however, become accustomed to refer to many types as
“asymmetrical.” In the sense in which we employ this
word we do not intend to convey the meaning that these
types alone of their respective classes depart from true
bilateral or radial symmetry, but rather to indicate that
they exhibit more asymmetry than the maximum contem-
plated in our generalized concept of, or arbitrary stand-
ard for, those classes.

Thus we readily recognize and confess the asymmetry
in the skull of the narwhal (Monodon) with its single
greatly elongated and twisted incisor, and the asymmetry
in the bones in the skull of the whales, while at the same
time we commonly consider man to be symmetrical,
though careful measurement shows the right arm and

521

522 THE AMERICAN NATURALIST [Vot. XLIX

hand to be larger than the left, and the left leg and foot to
be larger than the right.

It is clear, therefore, that in dealing with asymmetry in
any group we must work inward from the most asymmet-
rical types toward the least asymmetrical, arbitrarily
erecting a barrier between what we call asymmetry and
what we are pleased to consider as ‘‘symmetry’’ at any
point we choose.

Asymmetry—that is to say the maximum departure
from perfect bilateral or radial symmetry—appears to
follow certain definite lines wherever it appears, quite
regardless of the type of animal, or the form, in which
it is manifested.

In the following pages we shall consider the wider
variations from the typical pentamerous symmetry among
the recent crinoids, which is phylogenetically most exten-
sively developed at the consummation of the phylogenetic
lines, and physico-economically most extensively devel-
oped in the situations most unsuited to crinoidal exist-
ence, particularly in the very warm water of the Hast
Indian and north Australian littoral, and the very cold
water of the Antarctic regions and the deep abysses of the
oceans, and is least evident among phylogenetically con-
servative types, and in the situations which appear to be
best suited for crinoid life.

As an indication of the possible fundamental impor-
. tance of the light thrown on the study of asymmetry by an
examination of the data offered by the recent crinoids, it
may be noticed and borne in mind that among the mam-
mals the phylogenetically aberrant asymmetrical narwhal
(Monodon) is exclusively arctic; the phylogenetically
aberrant asymmetrical whales occupy a habitat very aber-
rant for the class; and the anthropoid apes, which are
pronouncedly right or left handed, live in very warm
regions; that among the birds the curious crook-billed
plover (Anarhynchus), with the beak twisted to the right
and one side of the body lighter in color than the other,
occurs only in New Zealand, the home of many phyloge-

No. 585] A STUDY OF ASYMMETRY 523

netic oddities; the hornbill Rhinoplax, with an asymmet-
rical tail, further peculiar in having a solid casque, an
elongate central rectrix, and a naked patch on the back
extending to the sides of the head, is found in the warm
Malayan region; the crossbills (Loxia), with the tips of
the mandibles crossed and a corresponding distortion in
the bones of the head are all subarctic or cold temperate
forms; and the owls with one ear greatly larger than the
other, so far as has been determined are, like the cross-
bills, birds of the colder regions; and that among the
fishes and similar types the very asymmetrical Anableps
lives in the warm tropical littoral, while the flatfishes
(Pleuronectidæ) are chiefly developed in the warm trop-
ical littoral, and in cold and shallow water, and the asym-
metrical forms of ‘‘ Amphi ’? (using the term in its
broadest sense) occur in warm and shallow water.

Further it is interesting to recall that animals under
domestication—that is, living under conditions which
typically lead to a more or less degenerate diversity in
form and color—commonly develop asymmetry of action
which, though usually occurring in the form of individual
variation, may become very marked as in the case of the
Japanese waltzing mice, as well as pronounced, though
irregular and sporadic, asymmetry in color pattern, denti-
tion, and other features.

More or less pronounced asymmetry undoubtedly exists
in many types in which up to now it has been overlooked,
and the conclusions reached in the present paper may be
modified somewhat when a better knowledge of the sub-
ject is attained; but on the other hand it is scarcely prob-
able that many instances of marked asymmetry have es-
caped the notice of naturalists.

THE DIFFERENT Types or RINOIDAL ASYMMETRY

In the great majority of the recent crinoids the body is
almost perfectly pentamerous, being composed of five
similar sectors. The presence of a small muscular cone
in the posterior interradius, at the summit of which is

524 THE AMERICAN NATURALIST [ Vou. XLIX

the posterior opening of the spiral digestive tube, gives
the only visible indication of a departure from true pen-
tamerous symmetry.

In certain types, however, a more or less marked devia-
tion from the characteristic symmetry occurs. This devi-
ation follows four different lines:

1. A rearrangement of the five primary groove trunks
upon the disk whereby (a) the left posterior increases in
size and gives off more branches than any of the others;
(b) as a result of the anterior migration of the mouth,
the two posterior become much longer and the anterior
much shorter than the others and a condition of bilateral
symmetry is attained; (c) correlated with the anterior
migration of the mouth, all of the primary groove trunks
become merged into a horse-shoe shaped ring which
skirts the lateral and anterior borders of the disk, giving
off branches to the arms, the mouth being in the right
center of the ring so that the ambulacra on the left are
more developed than those on the right, or the ambulacra
leading to the left posterior arm disappearing altogether
so that the ambulacra on the right are more developed
than those on the left;

2. A dwarfing, or an overdevelopment, of the left poste-
rior, more rarely of both, posterior radials with their
post-radial series;

3. The intercalation of additional radials and post-
radial series which alternate with the original five, and
the associated dropping out of one of the five radials; and

4. The suppression of two of the primarily five basals.

THE ASYMMETRICAL CRINOIDS

In the following list are given all the families and
genera of recent crinoids which include asymmetrical
species.

After the families the bathymetrical and thermal
ranges are given, and after the genera the bathymetrical
range.

Certain families are represented in the warm littoral

No. 585] A STUDY OF ASYMMETRY 525

water of the Malayan region and northern Australia, but
the highest actual temperature record is considerably less
than the temperature of this water; in these cases the
temperature 80.5° is given after the gadertapied maximum
as more nearly representing the true maximum.

Of the nine families the four in which asymmetry is
most markedly developed are marked with an asterisk
(*); and of the twenty-seven genera the sixteen which
include the most notably asymmetrical species are simi-
larly distinguished.

Depth (Fathoms) Temperature (F.)

Capillaserings: 00s 0s a co a ie 0-830 44.5-78.5 (80.5)
ETT he) Creo: BER MIRE AU EE ioe ERR E Marner 0-106
N pooma 2. SiS ks Ba 10-830
OOO WIECH oo ss 8 ee ee E 140-153
CODIR 64 a ae as 0-160
A I ic Ne OEE Oc PEER LS oS 0-194
CERRO Ste eas oa 100
LOONE OO Oe oe ee ee 42-163
Ean oa os oi ces es. a ee 0-288 62.0-71.9 (80.5)
Comot cock ts ES a A 0-160
Damien: ian aa e 10
ONAN E E E EEE ee ee 0-288
COMA Se cr a can vee 262
“CRONIN SS e a ie eee a 0-140 52.3-80.0 (80.5)
E AE A ic ahs cee eS 0-95
PCOMURTNOTUE © 6 orci oe es ere 0-83
COMONINRA GS, Se cae ec ees 2
POCORN oo a a 0-140
TROOP es ok en os 2-—1,600 28.7-60.5
ESA E RER I C E E S 10-222 28.7
FPentamotroiaide <... oee 103-1,800 33.5-60.6
WE RAmIOtOCHINEN Ooo. cas ce chen 861-1,800
VIN ons oo ak as oo eee 565-940 36.7-38.1
Nk TOMMOCTTNER oe ey ob en a ss 40
SCOTPOREOPOOTINUE o enana eed es 565
Bourgueticrinide aa Ca open bay ue ee 62-2,690 29.1—70.7
P RMNO ee ss CAs Bs he 77-1,300 32.2-48.7
RENOOTURNG oi os in Waka es Coens’ 687—2,419 37,4—40.0
*Rolopodido oae, «aed ee 5-120 71.0
ODER i E E oe Cie ees awake 5-120
"PF iedioerinide ... OS eG 266-2,575 31.1-43.9
"UGIOMIOONAUE -n.a oc Vs eee 392-782
"Pe Oerioe. ia Ca is es Ak 266-2,485
sf GNU cc Su Fis Soe a ee a 575
,103

526 THE AMERICAN NATURALIST [ Vou. XLIX

TEE PHYLOGENETIC DISTRIBUTION OF ASYMMETRY

The phylogenetic distribution of the asymmetry among
the recent crinoids is very interesting.

Asymmetry is almost universal in the comatulid family
Comasteridæ, which includes the most specialized of all
recent forms; in this family the first and second types
occur, though the latter is much less common.

Asymmetry is characteristic of the genus Promacho-
crinus, which is probably rightly considered as the most
specialized genus in the subfamily Heliometrinæ, the
largest and most universally distributed subfamily of the
at present dominant family Antedonidæ; in the genus
Promachocrinus the first and third types occur.

Asymmetry is equally characteristic of the genus Thau-
matocrinus, the most specialized genus of the family
Pentametrocrinidæ; in this genus the third type is found.

Asymmetry exists in all of the genera of the Plicato-
crinidæ, which includes the last highly specialized expo-
nents of the ancient order Inadunata, which flourished
from the Ordovician to the Carboniferous, with one
family extending into the Permian and Trias and another
(the present family) appearing in the Jura; in the Plicato-
crinidæ the first, second and fourth types occur in recent
genera, while the third is also found in fossil genera.

Asymmetry is characteristic of both of the recent gen-
era of Apiocrinidæ, which are the most specialized genera
in the family; in these the second type occurs.

Asymmetry of the second type is characteristic of the
only recent genus of the Holopodide.

Asymmetry characterizes both of the species of Rhizo-
crinus—which is at least as highly specialized as any of
the genera of the Bourgueticrinide—existing in the pres-
ent seas, and one of the species of Monachocrinus, a genus
of which the exact phylogenetic position is uncertain,
although it is probably on a par with Rhizocrinus; in
these the third type occurs.

In the following list the recent asymmetrical types are

No. 585] A STUDY OF ASYMMETRY 527

given in the order of the extent of their departure from
the normal pentamerous symmetry:

Plicatocrinide: Asymmetry of Types 1, 2, (3) and 4.
Comasteride : Asymmetry of Types 1 and 2.
Promachoerinus: Asymmetry of Types 1 and 3.
Apiocrinide Asymmetry of Type 2
Holopodide Asymmetry of Type 2

haumatocrinus Asymmetry of Type 3
Rhizocrinus Asymmetry of Type 3.
Monachecrinus : Asymmetry of Type 3.

The asymmetry of the Comasteride is considered more
fundamental than that of Promachocrinus for the reason
that it is characteristic of practically the entire family,
and also because it results in a much greater degree of
irregularity. It is interesting to note that asymmetry
of Type 3 is not uncommon among the Comasteride, in
the form of individual variation.

The asymmetry of the Apiocrinide and Holopodide is
considered more fundamental than that of the genus
Thaumatocrinus for the reason that it affects the entire
family, at the same time inducing a greater departure
from the normal form.

The asymmetry of Rhizocrinus is considered less funda-
mental than that of Thawmatocrinus because, though
affecting all of the species, exactly as in Thaumatocrinus,
it is less extensively developed.

The asymmetry of Monachocrinus affects only one of
the seven species of the genus.

Briefly stated, it appéars that, no matter in what form
it may manifest itself, metameric asymmetry in the recent
crinoids is an attribute of the most specialized types in
the groups in which it occurs.

From the conditions in the Plicatocrinide, the last
remnants of the once abundant Inadunata, it would appear
that asymmetry is an attribute of phylogenetically deca-
dent types—types in which type senescence has so far ad-
vanced as to inhibit the normal course of development.

THE GrocrapHicaL DISTRIBUTION oF ASYMMETRY
The geographical distribution of asymmetry is as inter-
esting as the phylogenetical distribution.

528 THE AMERICAN NATURALIST [ Vou. XLIX

Although occurring everywhere except in the Arctic
Ocean and in the Mediterranean, Bering, Okhotsk and
Japan seas, asymmetrical types are most frequent and
most highly developed (1) in warm shallow water from
southern Japan southward throughout the Malay Archi-
pelago to northern Australia and westward to Ceylon, and
(2) in the Antarctic and in the cold abysses.

Though present among species inhabiting the west
Atlantic from North Carolina to Brazil, and characteristic
of many forms living at intermediate depths in the west-
ern Pacific and in the Indian Oceans, in these it is never
more than slightly developed, even though they be very
closely related to types in which it is, in other situations,
carried to an extreme.

Depth (Fath Number of Number of ati gr va une
re Torrens Asymmetrical Genera | Symmetrical Genera | ‘or = eT mara
0-50 16 50 32%
100 15 53

100-150 13 51 25
150-200 10 44 22
200-250 5 39 13
250-300 5 34 14
300-350 3 30 10
350—4 4 32 12
400-45! 5 29 17
450-500 5 27 18
500-550 5 26 | 19
550-600 6 26 23
5 26 19
650-700 6 22 27
700-750 6 22 27
750-800 6 18 33
800-850 5 18 28
850- 4 18 22
5 19 26
950-1,000 5 1 31
1,000-1,100 5 16 31
1,100-1,200 5 12 41
1,200-1,300 5 9 55
1,300-1,400 4 9 44
1,400-1,500 4 7 57
1,500-1,600 4 7 57
1,600-1,700 5 3 166
1,700-1,800 5 3 166
1,800-1,900 4 3 133
1,900-2,000 4 3 133
2,000-2,500 4 3 133
2,500-3,000 1 3 33

In short, though almost universal, occurring every-

Genera with and without Asymmetrical Spe-

with very Asymmetrical and without As

metrical Species at Different Depths (===),

Genera with Depth,

ym-
of

and the Decrease in the Number
expressed in Percentages of the Total Number (- - - -).

),

Fic. 1. The Relation between the

cies at Different Depths (

5380 THE AMERICAN NATURALIST [ Von. XLIX

where except in inland seas, asymmetry is especially
developed in the warm waters of the eastern tropics, par-
ticularly in the Malayan region and in northern Australia,
and in the Antarctic and the cold abysses.

BATHYMETRICAL DISTRIBUTION OF THE ASYMMETRICAL
CRINOIDS
The number of genera of recent crinoids including
asymmetrical species, the number of genera including

| Number of | Per Cent. of the
Depth (Fathoms) Asymumetrical Genera | , Number of | Latter ———
Which Are Marked * | 5 bs | by the Former
0-5 7 50 | 14.0
50-100 7 53 | 13.2
100-1; 5 51 | 9.8
150-200 3 44 6.8
200-250 2 39 5.1
250-300 2 34 5.8
300-350 2 30 6.6
350-400 4 32 12.5
400-450 4 29 13.8
450-500 4 27 14.8
500-550 4 26 15.4
550-600 5 26 19.2
600-650 | 4 26 15.4
650-700 | 4 22 18.1
700-750 | 4 22 18.1
50-8 | 4 18 22.2
00- 3 18 16.6
850-900 3 18 16.6
900-950 4 19 21.0
950-1,000 4 16 25.0
1,000-1,100 4 16 25.0
1,100-1,200 4 12 33.3
1,200-1,300 4 : 9 44.4
1,300-1,400 3 9 33.3
1,400-1,500 3 ff 42.8
500-1, 3 7 42.8
,600-1, 4 3 133.3
1,700-1,800 4 3 133.3
1,800-1,900 3 3 100.0
1,900-2,000 4 3 100.0
2,000-2,500 3 3 100.0
2,500-3,000 1 3 | eee

only symmetrical species, and the percentage of the num-
ber of symmetrical genera represented by the number of
asymmetrical genera at different depths are given in the
table on page 528 and shown in Fig. 1.

Considering the percentages only, these may be re-
grouped as follows:

No. 585] A STUDY OF ASYMMETRY 531

ee Ri ek a a a a a S a 27
v00- G50 irs eo cle se Fs 5 ORs Cece ew ae Ce thew wa cee 16
GOO) 100 so ea cs eee ee a ee 28

py Nt, | Ee erg rr emma RE ee I angen 7 oes Senet aa 92

Considering only the genera marked with an asterisk
(*) we find the representation at different depths given
in the table on page 530 and in Fig. 1.

Considering the percentages only, these may be re-
grouped, as follows:

I ace aie ee CHa we eo owe a ey cakes 13.6
100e S00 se Ceres er ek ag a vag be ee ee hss eed Gate
BO 900 ae ee ee OU as as eh ek A ee ees 16.6
P003,000 -aoea ko ee ee sl heer 61.8

The number of families of recent crinoids including
asymmetrical species, the number of families including
only symmetrical species, and the percentage of the num-
ber of families including only symmetrical species repre-
sented by the number of families including asymmetrical
species at different depths, are shown in the table on page
933 and in Fig. 2.

The proportion of the genera including asymmetrical
species to those composed entirely of symmetrical spe- —
cies, about one third between the shore line and 50
fathoms, decreases to a minimum of one tenth at from 300
to 350 fathoms, and then increases, with greater and
greater rapidity, to 1,600 fathoms and below.

It is everywhere less than one quarter between 100 and
650 fathoms. Thus it is evident that the genera including
asymmetrical species are chiefly developed in shallow
water, and in deep water, and are least developed in water
of intermediate depth.

Taking the ocean as a whole, the temperature at 100
fathoms is 60.7°, and at 650 fathoms 38.6°; the optimum
temperature for the recent crinoids appears to be between
50° and 65°; when we remember that most of the asym-
metrical species, and all of the most asymmetrical ones, in
the genera which give us our numbers for 0-50 and for
50-100 fathoms, are confined to a littoral belt of scarcely
more than 50 fathoms, it becomes at once evident that
asymmetry among the crinoids is developed chiefly in

Ne — ee ee

6008-0062

Proportion of Families with and without Asymmetrical Species

Fia. 2. The
at Different Depths (

), and the Percentage of the Total Number of ©

at Different Depths (- - --).

Crinoid Families represented

No. 585] A STUDY OF ASYMMETRY 533

| t. of Sym-
| Number of The Families Number of metroa r siiis
(rations) | Abymmstrical |yarkod wih ane | SZameirionl |Hepresentd by
|
0-50 5 2 i 33
50-10 6 - as 4
100-150 7 2 u H
150-200 | 5 0 : “
200-250 | 5 0 0 oa
250-300 | 6 0 : ae
300-350 | 5 0 8 62
350-400 5 0 : n
400-450 5 0 : li
450-500 | 5 0 i n
500-550 | 5 0 5 >
550-600 | 6 1 : ie
600-650 6 1 : s
650-700 | 6 1 :
700-750 | 6 1 7 a
750-800 | 6 1 : a
800-850 | 6 1 ~ n
850-900 | 5 1 ` =
900-950 5 1 :
50-1, 4 0 é -
1,000-1,100 4 0 : =
1,100-1,200 4 0 : bes
1,200-1,300 4 0 $ 10
,300-1, 4 0 í sh
1,400-1,500 4 0 : i
1,500-1,600 4 | 0 : po
1,600-1,700 3 | 0 : xo
1,700-1,800 3 | 0 } | ps
1,800-1,900 2 | 0 : -
1,900-2,000 2 | 0 : re
1,900-2,000 2 | 0 l -=
2,000-2,500 2 | 0 : | ~
2,500-3,000 2 0

water above and below the optimum, and least at and just
below the optimum temperature. :

Onia only the genera marked with an asterisk
(*), that is, the genera with the most highly developed
asymmetry, we find the same general facts emphasized as
in the case of all the genera including asymmetrical
forms; but here the minimum is between 50 and 400 fath-
oms instead of between 100 and 650 fathoms. The tem-
perature at 400 fathoms is 41.8°. This approximation of
the minimum to the zone of optimum temperature when
only the most asymmetrical. types are considered
strengthens the hypothesis that the zone of optimum
temperature really represents the zone of least-developed
asymmetry.

534 THE AMERICAN NATURALIST [ Vou. XLIX

Comparing the proportionate abundance of asymmet-
rical genera at different depths with the frequency of all
the genera expressed as percentages of the total, we find
that the former decreases while the latter increases to
50-100 fathoms; from this point the two run roughly
parallel to 300-350 fathoms, after which the former in-
creases with progressively greater rapidity. while the
latter decreases steadily and gradually to 3,000 fathoms;
the two cross each other between 600 and 700 fathoms.

The proportion of the families including asymmetrical
species to those composed entirely of symmetrical species
increases from one third at 0-50 fathoms to three times
as many at 1,600 fathoms and twice as many at 1,900
fathoms and over. The increase, though irregular—
largely as a result of the small numbers involved at the
greater depths—is constant.

The number of families at different depths, expressed
as percentages of the total number, increases from 0-50
to 50-100 fathoms, and then decreases to 1,800 fathoms
and beyond. Except for a minimum between 350 and
500 fathoms the decrease is fairly regular.

The two lines cross between 200 and 300 fathoms.

The reversal of the direction of the line representing
the frequency of the families including asymmetrical spe-
cies as a percentage of the number of the families not in-
cluding asymmetrical species at different depths, as com-
pared with the line representing the frequency of the
families at different depths expressed as percentages of
the total number, indicates that the less favorable the
environment for crinoids as a whole the greater becomes
the proportion of asymmetrical forms.

In the proportion of genera including asymmetrical
species to those composed entirely of symmetrical species
we find a minimum between 100 and 650 fathoms or, con-
sidering only the most markedly asymmetrical types, be-
tween 50 and 400 fathoms, the numbers above 100 (or 50)
fathoms and below 650 (or 400) fathoms being greater.

Considering families in the same light we appear te

No. 585] A STUDY OF ASYMMETRY 535

have an increase between 350 and 500 fathoms—that is,
more or less coinciding with this minimum.

In the frequency of families at different depths ex-
pressed as percentages of the total number we notice a
minimum between 350 and 500 fathoms which reaches a
point not again touched until 750-800 fathoms and beyond.

This indicates the occurrence here of a proportion-
ately large number of families including asymmetrical
species, but at the same time a proportionately small
number of genera including asymmetrical species within
those families.

THERMAL DISTRIBUTION oF THE ASYMMETRICAL CRINOIDS

In examining the thermal distribution of asymmetry
among the recent crinoids we find it advisable to employ
family instead of generic units, for the reason that our
records are insufficient to furnish us with even approxi-
mate thermal ranges for many of the individual genera,
though in most cases these may be estimated with rea-
sonable exactness. The records for the crinoids of the
deeper water are far more satisfactory than the records
for the crinoids of the littoral, and this is very fortunate,
for it justifies us in assigning a temperature of 65° and
‘Over to a number of species and genera which are of
great importance in the present study.

In considering asymmetry in relation to temperature
by family units it must constantly be borne in mind that,
whereas certain families (Capillasterine, Comactiniine
and Comasterine) extend from the warm littoral into
moderately deep water with a relatively low temperature,
the asymmetry among their component genera and species
is strongly marked only in very shallow water of high tem-
perature, and is only slightly marked—indeed not infre-
quently entirely absent, as in Comatilia—in genera and
species inhabiting deep and cold water.

Thus through a study of family units the amount of
asymmetry shown at intermediate temperatures is really
exaggerated, and appears in its relation to the higher

536 THE AMERICAN NATURALIST [Vor XLIX

and to the lower temperatures considerably greater than
it really is.

In the subfamily Heliometrinæ, the largest and most
widely distributed subfamily of the Antedonidæ, which
itself is the dominant crinoid family of the present seas,
the range of temperature is very great; but as only one
out of the ten genera of the Heliometrinæ is asymmetrical
it has seemed sufficient to consider and to tabulate the
temperature of this genus (Promachocrinus) alone.

he frequency of the families including asymmetrical
crinoids at different temperatures is as follows:

r eS Ae 4
Ce re E 4
e 2 oa i Peon th his ous Mae hae 4
T a ei a 3
ee a ee 4
OO a 3
owl A y E A 3
Da ee et a 3
SOE EE Ee S 4
OOS a ee re ee 5
aD ce OnE SUE up i TE Neen reese EO 4
Me a 1

Polow TE pa Soc bo Vins & Vdiein 6 oe Fie ee ORS VR ET SE 1
DUM E N A A E AT 4.3
BB -OD rs eta cick ce es weed eed bees aa were een 3.2
COBO. tinh ib hc ess RR EEE Oa Cie hee CER es Ve 3.7

or,

MUONS Dos og kw 6 oa Od ie oes aS T E 1
= oe APRESS GSES GRRE a eM Me mA at 4.3
ME I Sek Bei ea re NON Boa eRe oes ese tae 3.2
UPC CEU beeen cues os ae Ces mee cue Ly a eee 4.0

Considering the zone of optimum temperature (50°-
65°) in contrast to the temperatures above and below, and
omitting the exceptionally low temperatures below 30°,
we have:

No. 585] A STUDY OF ASYMMETRY 537

Bearing in mind always that the frequency between the
warm littoral and the cold abyssal temperatures is exag-

gerated because of the segrega-
tion in the warm littoral zone of
the most asymmetrical genera
and species in many of the fam-
ilies inhabiting intermediate tem-
peratures, it is clear that asym-
metry is least developed at the
optimum temperature for crinoid
life, and most developed in tem-
peratures which are phylogenet-
ically too warm or too cold.

This agrees perfectly with what
we found from an examination of
the bathymetrical distribution of
asymmetry.

A comparison between the fre-
quency of the families of crinoids
represented in the recent seas,
including only symmetrical spe-
cies, given in the actual numbers
and also as percentages of the
total numbers, and the frequency
of the families including asym-
metrical species, given in the same
way, follows (Fig. 3):

I
l
l
l
l \
l
l
I
l
l

'
i
nee eee
Te ome a
1

ow co 0
owe
: Li i]
o

Fie, 3. Frequency at Dif-
ferent Temperatures of the
Families eterna As epee”
rical Species (
those erapl Symmetriea
Species only (- - - -).

Families with — with
mperatur a Per Cent.
(Fate ubeit) Svecee Saiy ps ae a Total”

85°-80° 2 13 4 | 44
80 -75 2 13 4 | 44
75 -70 9 60 4 | 44
70 -65 9 60 3 | 33

—60 14 93 4 | 44
60 -55 12 80 3 | 33
55 —50 | 11 73 3 | 33
50 -45 | 7 47 3 | 33
45 —40 | 47 4 | 44
40 -35 | 7 47 | 5 | 55
35 -30 3 20 e 4 | 44
30 -25 1 7 | 1 | 11

538 THE AMERICAN NATURALIST [Vor XLIX

THE ASYMMETRICAL FEATURES IN DETAIL
In the following list are given the four types of asym-

metry occurring in the recent crinoids, with their geo-

graphical distribution and the genera in which they are
found.

1. Disk Not Radially Symmetrical

Geographical Distribution—Southern Japan south-
ward to Samoa, Fiji and southern Australia, thence west-

aa
-l
“i

|

|

|

Soa.

-_-—
—
-_-—
—_—
_—
p= e
pen
oe ee ee

(————-), the Genera with One or More Rays
the Genera with Six to Ten
( exp

d or gp rged (——
Rays (- -), and the Genera with Three Basal
ressed as Percentages of the “otal heuer in Each Clas

ward to east Africa, from the Red Sea to the Cape; north-
western Africa and southwestern Europe (in moderately
deep water), and from South Carolina to Brazil; antarctic
regions, littoral to abyssal, and northward along the

No. 585]

A STUDY OF ASYMMETRY

539

eastern shores of the Pacific (in deep water) to British

Columbia.

This character is most strongly
marked in the shallow water
from the Marshall Islands and
New Caledonia through the Malay
Archipelago and along the north-
ern coasts of Australia, and
thence westward to Ceylon; and
again in the antarctic regions and
the abysses of the east Pacific.

Systematic Distribution —
Capillasterinz
Comatella Capillaster
Neocomatella Ne
Paleccomatella
Comactiniine
omatula Comint
Comatulella Comantinta
Comaster Comanthina
Comantheria omanthus
inept
machocr
AETR
Ptilocrinus

2. One or More Rays Dwarfed,
or Enlarged
Geographical Distribution. —
Malayan region and north Aus-
tralia, and Caribbean Sea, bu

only in warm and shallow water;

oo oO
ow at]
[J ' J
led »
©

O O 0
rts.
ul ee S
nO n oO

oO eteo

o
a
t
o
bgd

Frequency at Dif-
Temi mperatures of Fam-
tae Species in
e Disk is not Radially
m-

including Species with
or More Rays Dwarfed or

ae

eeh ed ( Fam-
ves inc ppc: Species with
m Six to T =-=- -),

s (-
a Pantie including Species
a Three Bas
and the Total ot a these t
regularities.

Malay Archipelago to

southern Japan, and Galápagos Islands to Central Amer-

ica in deep cold water.
Systematic Distribution.—
Capillasterinz

Capillaster (part)
Comactinii

ine

Comasterine
Comaster (part)

Comanthina

540- THE AMERICAN NATURALIST [Vou. XLIX
Comanthus (part)

Apiocrinidæ
Carpenterocrinus

Holopodidæ

Comantheria (part)

Proisocrinus

Holopus
Plicatocrinidæ
amocrinus

3. Six to Ten (Sometimes Four) Rays

Geographical Distribution—Southern Japan and the
Hawaiian Islands to the Malay Archipelago, in rather
deep water; abysses of the Indian Ocean and the Ant-
arctic; Florida northward and northeastward to Iceland
and Norway in deep and cold water.

1

hie a
(
[i
i
-i

-
wwe meme
@----
steerer
aoe essameser==
wow eee

SEEN

~

~~
a“
Ste wm

|
\
|
\
\
|

1G. 6. Proportion at Different Depths of Genera only Symmetrical Species,
and Genera including Species Asymmetrical Disks ( ), Genera in-
cluding One or More of the Rays Dwa
Genera including Species with from Six to Ten Rays (—— — -), and Genera in-
cluding Species with Three Basals (--- - -- j.

No. 585] A STUDY OF ASYMMETRY 541

This feature as an individual variant occurs in the
warm water of the Malayan region, in the shallower por-
tions of the Caribbean Sea, and very commonly on the
tropical Brazilian coast:

Depth R Ite ap ee Dr cpr be Six to Ten Three Total (35
(Fathoms) | metrical (17) lorkelarged(10)| Rays (4) Basals (4) 04a: (25)
0-50 (15) 88 (7) 70 (1) 25 0 (23) 66
50-100 (13) 76 (6) 60 (2) 50 0 (21) 60
100-150 (11) 65 (4) 40 (2) 50 0 (17) 49
150-200 (9) 53 (2) 20 (2) 50 0 (13) 37
200-250 (4) 23 0 (2) 50° 0 6) 17
250-300 (4) 23 0 (1) 25 (1) 25 6) 17
300-350 (2) 12 0 (1) 25 (1) 25 4) 11
350-400 (2) 12 (1). 10 2) 50 (1) 25 6) 17
(2) 12 (H) 10 2) 50 (1) 25 6) 17
450-500 (2) 12 @) 10 2) 50 (1) 25 6) 17
500-550 (2) 12 (1) 10 2) 50 (1) 25 6) 17
550-600 (2) 12 (2) 20 2) 50 (1) 25 7) 20
65 (2) 12 (1) 10 (2) 50 (1) 25 6) 17
650-700 (2) 12 (1) 10 75 (1) 25 ) 20
700-750 (2) 12 (1) 10 75 (1) 25
750-800 (2) 12 (1) 10 75 (1) 25 ) 20
800-850 (2) 12 0 ) 75 (1) 25
850-900 (i). 6 0 75 (1) 25 ) 14
950 BD G (1) 10 ) 75 (1) 25 ee ys
950-1,000 (ly 6 0 ) 75 (2) 50
1,000-1,100 (1) 6 0 ) 75 (2) 50 17
1,100-1,200 Uy. 9 0 ) 765 (2) 50
1,200-1,300 (1) 6 0 ) 75 (2) ar
1,300-1,400 (1) 6 0 ) 50 (2) 50 ) 14
1,400-1,500 W 6 0 ) 50 (2) 14
: ,500-1,600 U6 0 (2) 50 (2) 50 (5) 14
1,600-1,700 1) 6 0 (2) 50 (3) 75 ) 17
: »700—1,800 (6 0 (3) 75 (6) 17
800-1,900 (i) 6 0 (1) 25 (3) 75 (5) 14
oa ,000 G) © 0 (1) 25 (3) 75 (5) 14
2,000-2,500 () 6 0 (1) 25 (3) 75 (5) 14
2,500-3,000 0 0 (1) 25 (1) 3
Disk Not | Oneor More |
Temperature Radially | Rays Dwarfed| Six to Ten Three Basals Total
(Fahrenheit) Aranaren | | or Enlarged Rays
80°-75° 2 | 3 0 0 5
-70 3 | 4 0 0 T
70 -65 3 | 3 0 0 6
65 -60 3 | 3 1 0 ri
60 -55 2 | 2 X 0 5
—50 2 | 2 1 0 5
50 —45 1 | I 2 0 4
45 -40 2 | 2 2. 1 7
40 -35 1 | 2 3 1 7
35 -30 1 | 1 2 1 5
30 -25 1 | 0 | 1 0 2

542 THE AMERICAN NATURALIST [Von XLIX

Systematic Distribution. —
Heliometrinæ
Promachocrinus
Pentametrocrinidæ
haumatocrinus
Bourgueticrinide
Monachocrinus (part) Rhizecrinus

The Numb f Genera
Number of Genera Number of Genera Dube accessed es os
Depth (Fathoms) with ee with p Fe geen Percentage of the
isks isks Number with Symmet-
ri isks

0-50 15 51 29
50-100 13 55 23
100-150 pS 53 21
150-200 9 45 20
200-250 4 40 10
250-300 4 35 11
00-3 2 31 6
350-400 2 34 6
00-4 2 32 6
450-500 2 30 7
500-550 2 29 7
550-600 2 30 7
600-650 2 29 7
650-700 2 26 8
700-7 2 26 8
750-800 2 22 9
800- 2 21 9
850-900 1 21 5
900-950 1 23 4
950-1,000 1 20 5
,000-1,1 1 20 5
1,100-1,200 1 16 5
1,200-1,300 1 13 A 8
1,300-1,400 1 12 8
1,400-1,500 1 10 10
1,500-1,600 1 10 10
1,600-1,700 1 7 14
1,700-1,800 1 7 14
1,800-1, 1 6 16
1,900-2,000 1 6 16
2,000-2,500 1 6 16

2,500-3,000 0 4 D

4. Three Basals

Geographical Distribution.—Antarctic regions, and
northward to northwestern Africa, the Caroline Islands,
and British Columbia, except in the antarctic always in
very deep water.

No. 585] A STUDY OF ASYMMETRY 543

Systematic Distribution.—

Plicatocrinide
Ptilocrinus Gephyrocrinus
Hyocrinus Thalassocrinus

The frequency of each of these four types of asymmetry
at different depths and temperatures is given in the
tables on page 541 and in Fig. 4.

The rg mber of —
: wi symmetric:
Dopik Puika | ete Aar aonn | GIER DPS ppi pape gior g
Rays Number with Sym-
metrical Rays
0-50 7 59 12
50-10 6 62 9
100-150 4 60 6
150-200 2 52 4
00-250 0 44 0
250-300 0 39 0
300-350 0 33 0
350-4 1 35 3
400-450 1 33 3
450-500 1 31 3
00-550 1 30 3
550-600 2 30 6
00-650 1 30 3
650-700 1 27 4
700-750 1 27 4
750-800 1 23 4
800-850 0 23 + 0
0-9 0 22 0
00-9 1 23 4
950-1,000 0 21 0
1,000-1,100 0 21 0
1,100-1,200 0 17 0
1,200-1,300 0 14 0
,300-1,4 | 0 13 0
1,400-1,500 | 0 11 0
,500-1,6 | 0 11 0
1,600-1,700 | 0 8 0
1,700-1,800 | 0 8 0
1,800-1,900 0 7 0
1,900-2,000 | 0 7 0
2,000-2,500 0 7 0
2,500-3,000 | 0 4 0

In the table showing the frequency at different depths
the numbers in parentheses represent the actual cases, the
other numbers being the percentage of the total number
of genera in which the feature under consideration is
found. This last is given in parentheses at the head of
each column.

544 THE AMERICAN NATURALIST [ Vou. XLIX

For a graphic representation of the data in the table
on the lower part of page 541 see Fig. 5.
These frequencies group themselves as follows:

Wei i ee See 6.2
Ga cp ee cet ee . nel eae Ree 4.7
aa, ee ee ri 6.3
A E Cor oN Ge er a oe ee Varese 2.0

or, segregating those occurring at the optimum tempera-
ture:

o o

BOBS E E SEN we NN RE EE E eK Ewe 6.0

iting ee TE Oe OR Se ae PEE eee Pe 5.6

yp OR EEE OTe ee ee E ee ee eee 5.7

NOE Dee Ee Ona fey ERTE E rip rere 2.0

he Number of Genera
Number of Genera Number of Genera rich More Than Five
Depth (Fathoms) with More Than with Always Rays Expressed as a
Five Rays Five Rays Poe of the Num-
h Five Rays

0- 1 65 1
50-100 2 66 3
100-150 2 62 3
150-200 2 52 4
200-250 2 42 5
250-3! 1 38 3
300-350 1 32 3
350-400 2 34 6
400-450 2 32 6
450-500 2 30 6
500-550 2 29 7
550-600 2 30 6
5 2 29 t
650-700 3 25 12
700-7 3 25 12
7 3 21 14
00- 3 20 15
850-900 3 19 16
900-950 3 21 14
950-1,000 3 18 16
1,000-1,100 3 18 16
1,100-1,200 3 14 21
1,200-1,300 3 11 27
1,300-1,: 2 il 18
1,400-1,500 2 9 22
1,500-1,600 2 9 22
1,600-1,700 2 6 33
1,700-1,800 2 6 33
1,800-1,900 i 6 16
1,900-2,000 1 6 16
2,000-2,500 1 6 16
2,500-3, 0 4 0

No. 585] A STUDY OF ASYMMETRY 545

The relation at different depths between the crinoids
in which the disk is not radially symmetrical and those in
which it is radially symmetrical is shown in the table on
page 542 and in Fig. 6.

The relation at different depths between the crinoids in
which one or more rays are dwarfed, or, more rarely, en-
larged, and those in which all of the rays are of the same
size is shown in the table on page 543 and in Fig. 6.

e boca a
5 x wi ree Basals
Depth (raa | A Ee | ee foe aape y frale i
ber with Five Basals

0-50 0 66 0
50-10 0 68 0
100-150 0 64 | 0
150-200 0 54 | 0
200-250 0 44 | 0
250-300 1 38 | 2
300-350 1 32 3
350-400 i Sa 3
400-450 1 33 3
50-500 1 31 3
500-550 1 30 3
550-600 1 31 3
00-650 1 30 3
650-700 1 rt § 4
700-750 1 27 4
50-8! 1 23 4
800-850 1 22 4
0-9 1 21 5
900-950 1 23 4
950-1,000 2 19 10
1,000-1,100 2 19 10
1,100-1,200 2 15 13
,200-1,300 2 12 16
1,300-1,400 9 11 18
,400—1,500 2 9 22
1,500-1,600 2 9 22
1,600-1,700 3 5 60
1,700-1,800 3 5 60
1,800-1,900 3 4 75
1,900-2,000 3 4 75
,000-2,500 3 4 75

2,500-3,000 1 3 3o

The relation at different depths between the crinoids
with more (less frequently less) than five rays, and those
with five rays, is shown in the table on page 544 and in
Fig. 6.

The relation at different depths between the crinoids
with three basals and those with five is given in the table
given above and in Fig. 6.

546 THE AMERICAN NATURALIST [Von XLIX

SUMMARY

Among the recent crinoids any wide departure from the
normal close approximation to true pentamerous sym-
metry indicates unfavorable conditions of one or other of
two main types, which are not mutually exclusive.

These two types are

1. INTERNAL UNFAVORABLE CONDITIONS, induced by incip-
ient phylogenetical degeneration through type-senescence,
as in the Plicatocrinide, which in the recent seas repre-
sent the almost exclusively paleozoic Inadunata; and

2. EXTERNAL UNFAVORABLE CONDITIONS, taking the
form of

(a) Phylogenetically excessive cold, which, to
cite one example, appears to be the determining
factor in the asymmetry of the genus Promacho-
crinus; or of

(b) Phylogenetically excessive warmth, which
appears to be the determining factor in the asym-
metry of the family Comasteride.

INHERITANCE OF HABIT IN THE COMMON BEAN

JOHN B. NORTON, M.S.

MASSACHUSETTS EXPERIMENT STATION

Hasır is the external form of a plant taken as a whole.
It is usually described by a few general adjectives, such
as erect, open, spreading, etc. However, to study the
inheritance of plant habit, a detailed analysis of the real
characters underlying habit must be made. It is usually
found that the general outer appearance of a plant, its
habit, is the result of a combination of independent char-
acters, units, the recombination of which by crossing
often results in plants much altered in appearance from
the parent varieties. Characters usually unimportant
may be found of primary importance in the formation of
plant habit.

An example of such inheritance of habit is found in one
of Webber’s pepper hybrids (6). A cross was made be-
tween Red Chili, a variety with many erect fine branches,
and Golden Dawn, with few, horizontal, coarse branches,
both being of medium size. In the second generation re-
combination and segregation of the three character pairs
occurred, although not in strict Mendelian proportions.
The important feature of the results, however, lies in the
apparent creation of a giant and a dwarf type, not by the
appearance of new units by mutation, but simply by the
transference of the characters fine and coarse branches.
Hybrids having erect, many and coarse branches were
giants, while those having few, horizontal and coarse
branches were dwarfs. Other combinations of these
characters gave intermediate forms.

The study here reported was made largely on third and
fourth generation plants and a few second generation
plants of hybrids made primarily for the study of pig-

547

548 THE AMERICAN NATURALIST [Vou. XLIX

mentation. The material worked with, owing chiefly to
lack of knowledge of earlier generations, offered many
limitations and is unsuited to a detailed analysis of the
characters in question. As the plants were usually not
more than six inches apart in the rows, the crowding in
the later stages of development hindered accurate judg-
ment of the habit type.

With reference to general habit bean plants are either
pole or bush. Pole beans are commonly long twining
vines, climbing when provided with poles or other sup-
port. The true bush type is usually short, erect and non-
twining. There are also certain races of beans really in-
termediate between the true bush and pole types, the run-
ner beans, which are non-climbing. Types classed as
bush beans also occur, which are spreading and possess
outstretched branches of a more or less runner-like char-
acter.

The following table contains a description of habit of
varieties of beans considered in this discussion. The de-
scriptions are from ‘‘American Varieties of Garden
Beans’’ (5). The varieties observed agree with these
descriptions except in the case of Mohawk, which is de-
seribed as without runners. The strain of Mohawk iso-
lated here produces runners.

TABLE I
DESCRIPTION OF BEAN VARIETIES
Pole Beans ALT1
Golden Carmine—Small, good climber
Creasback—Small, at first bush-like, poor climber when young.
Runner Beans ALt

White Marrow—Very large, very spreading, many runners.

Bush Beans AIT

Burpee Stringless—Large, medium, very erect when young, with a few shoots
high above the plant, = more or less spreading when mature; no runners.
Giant Stringless—Same as

1 For the meaning of these letters see page 550.

No. 585] INHERITANCE OF HABIT 549

Semi-runner Forms Alt

Refugee—Very large, very spreading, many semi-runners.
Refugee Wax—Large, medium, very spreading, many runner-like branches.

Spreading Forms aLT or alt
Longfellow—Large to medium, somewhat spreading, many outstretched
branches, no real runners.
Kenny Rustless—Large, very pee almost runner-like bran
Prolific Black Wax—Medium, more or less spreading, arc oa out-
stretched branches, no real runners.

Erect Forms alT or alt

Black Valentine—Large, medium, fairly erect, occasional drooping branches,
no real runners

Blue Pod—Medium, erect, no runners or spreading branches.

Ta rge, medium, fairly erect when y m but drooping when ma-

no runners or decided spreading branc

Bet Kidney—Large, no runners, but as drooping with fruit-laden
branches and spreading when mature

ee Maer Wax—Large, sometimes ‘with drooping branches, but no real

Challenge Black Wax—Very small, erect, no runners or spreading branches.
Cu rries—Medium , erect, no runners or — branches.
ax—La ine. medium, erect, no runners.
Early Refugee—Medium, very erect, no runners or spreading branch
German Black Wa Bg Fis ek erect when er usually borne pen with
fruit laden branches when mature, no runners.
Long Yellow Six Weeks—Medium, very cae. no runners or spreading
ranches

Low Champion—Very large, usually erect, no runners or — branches.

Mohawk—Large, very erect, no runners, sometimes drooping w

Red Valentine—Medium, erect, no runners or spreading branches,

Round Yellow Six Weeks--Ginall; medium, very erect, no runners or spread-
ing branches,

Wardwell—Large, medium, fairly erect, no runners.

Warren—Very large, usually erect, no runners or decided spreading
branches

Wa RART a erect, no runners or spreading branches.

R. A. Emerson in his experiments on heredity of plant
habit in beans found three main character pairs con-
cerned, namely, length of plant axis, developed in vari-
ous degrees; ; twining habit or circumnutation developed
in various degrees or not at all; and lastly, the position of
pods, axial or terminal. His data involve chiefly the
latter character pair, which is inherited in a 3:1 propor-

550 THE AMERICAN NATURALIST [Vor. XLIX

tion, the axial position of pods being dominant. The posi-
tion of pods or flowers influences plant habit in this man-
ner: when flowers are formed at the growing tip of a
main stem or branch, such a stem or branch must neces-
sarily cease to elongate; on the other hand, if no flowers
or fruits are formed at that point it may continue to grow
indefinitely.

The habit of all the varieties of beans can be accounted
for easily with only these three character pairs. In
Table I the varieties here concerned have been grouped
according to the probable presence or absence in them of
the characters mentioned.

I have designated the axial position of the pods as A,
the terminal position by a; long plant axis by L, short
by 1; a long axis was shown to be dominant over short in
some of Mendel’s crosses of beans (1). I have designated
circumnutation by T and its absence by t, as, judging
from Emerson’s statements, and according to my own
observations twining habit is dominant. The possible
combinations of these characters are as follows:

HABIT TYPES

Type a, ALT....Pole beans.
Type b, ALt.....Runner beans.
s.

Type a comprises the pole beans, as the vines are of
great length, both on account of long axis and not being
checked by any terminal inflorescence, and as they can
climb by virtue of cireumnutation.

Type b comprises the runner beans. They aredlike the
pole beans except that the climbing habit is not developed
to any great extent, if at all. Between these two types
it is difficult to draw sharp distinction, but the true
runner probably lacks the factor for twining.

No. 585] INHERITANCE OF HABIT 551

Type ¢ probably represents the varieties which early
send up a few shoots high in the air like Burpee String-
less. In such beans the growth of the main stems or
branches is not entirely prevented by the absence of
the character which produces a long axis, and as the
climbing habit is more or less developed, the characteris-
tic shoots are sent up.

Type d represents the semi-runners, caused by the
short axis.

Combinations of type e and e, are the spreading varie-
ties, with long outstretched branches. They are to be
distinguished from runners by terminal inflorescences.
Kenny Rustless is a representative of the e type of habit
and probably Prolific Black Wax also.

The last two combinations, f and f, are the typical
erect bush form, such as Blue Pod Butter and Challenge
Black Wax.

Table II gives the possible crosses of these types and
the F, proportions to be expected when the forms crossed
are the most nearly typical. In the cases of typical
forms, the F, types should be differentiated without
much difficulty. A circumstance that must be looked
upon as a possible cause of exceptions is the presence of
unknown factors that cause variations in the intensity of
the development of the twining habit and of the inter-
mediate lengths between long and short axis. If there
are various factors for length, as Emerson assumes to be
the case in all quantitative characters (3), and if the
twining habit is to be explained in much the same way,
results may be considerably at variance with the expecta-
tions indicated in Table II. It must be remembered that
the constitutions given for the varieties are only as-
sumed.

At present, owing to circumstances mentioned before,
TABLE II
Constitution Type F: Proportions

ALTX ALT axo a

ALT xX ALt axb 3a: 1b

ALT xX AIT aXe 3a: le

ALTX Alt axd

z

Ae wrrs

552 THE AMERICAN NATURALIST [ Vou. XLIX

ALT 7a
ALt 2 3b
AIT gametes 3c
Alt] 1d

5 ALT X aLT aXe 3a: le
6 ALT X aLt aXe
Ta
ALt 2 3b
aLT gametes 3e,
aLt le,
Qa: 3b: 4e

vå ALT X alT axf 7a

ALT 2 3b
AIT 3e
aLT gametes if
alT Qa: 3b: 3e: 1f
8 ALT X alt ax? 15a
ALT O- 19
ALt 4 Te
AIT So Eos Sad
Alt Te
aLT 5
aLt 3f
alT 1
alt 27a: 9b: 9c: 3d: 12e: 4f
9 ALt X ALt bXxXb b 5
10 ALt X AIT bXe 9a: 3b: 3c: 1d as in type No. 4.
11 ALt X Alt bxd 3b: 1d
12 ALt X aLt bXe 3b: le
13 ALt X aLT bXe Ta
AL 2 3b
ALt 3e
aLT gametes 1
aLt S S 4

14 ALt X alT bXf 27a: 9b: 9c: 3d: 12e: 4f as in type No. 8.
15 ALt X alt bxf
ALt

No. 585]
16 AIT X AIT exe
17 AIT X Alt cxd
18 ATT X aLT exe
19 AIT X al? cxf
20 AIT Sait oxe
Si: AIT X ale cxf
AIT
Alt
alT
alt
22 Alt X Alt axd
23 Alt X aLT aXe
24 Alt X aLt aXe
ALt
Alt
E gametes
alt
25 Alt X alT dxf
AIT
Alt
alT gametes
alt
26 Alt X alt aX f
27 aLT X aLT exe
28 aLT X aLt exe
29 aLT X alT xT
30 SLT X alt exf
aLT
aLt
alT gametes
sit}
31 aLt X alt ixe
32 alt Xx alT éxf
33 aLt X alt EXT
834 alT X alT PAT
35 alT X alt I XF
36 alt X alt [xf

INHERITANCE OF HABIT

3f

1

9c: 3d: 4f
d

27a: 9b: 90: Sd:

e
3e: 1f as in type 30.
3: If

F
f

f
: 9e: Id: 12e:

12e:

553

d !
: 3e: 1f as in type No. 7.

4f as in type No. 8.

4f as in type No. 8.

only general notes on the behavior of various types of
crosses can be given.

554 THE AMERICAN NATURALIST [Von XLIX

Tyre 2. ALT X ALt

In the third generation of a cross of Creasback, a typ-
ical pole bean with White Marrow, a runner bean with
probably a weak character for circumnutation, all lots
were of axillary inflorescence. The habit of climbing
was developed in various degrees so that classifications
of types was difficult.

Cross Tyre 6 or 7. ALT +aLT or alt

Notes on an early cross of Creasback by Prolific Black
Wax indicate that the generation F, were pole beans, the
generation F, segregating into 33 pole and 8 bush. The
latter is probably a 3:1 proportion as expected. Whether
all plants described as bush were of the spreading type
does not appear from our records.

Cross Tyre 8. ALT X alt or alT

In a cross of Creasback with Blue Pod, a typical bush
bean, there occurs one strain of homozygous pole plants,
and also in the F, generation heterozygous types. Pole
and runner forms and bush forms of various types occur
in the proportions of 9:7 in one lot and in another of 3:1,
as might be expected in an F, generation. In another
small lot occur plants with long outstretched branches,
in another two plants of c type of habit. Evidently
Blue Pod has the constitution alt.

The date from a cross of Creasback with Blue Pod do
not signify much, as the types isolated happen to be con-
stant, one a pole type and bush types, of which several
are described as somewhat spreading. In one there
occurs a runner bean.

Creasback and Warwick crosses in the F, generation
behave consistently with the cross type, as assumed. In
one lot, 1 2 have axial inflorescence and three terminal.
Lots with spreading plants occur and one plant was noted
which possessed a very long axis, along with a twining
habit, but also terminal inflorescence. According to the

No. 585] INHERITANCE OF HABIT 555

explanation of habit characters assumed, such a plant
would have the formula aLT. Without a support which
happened to have been placed near it, the peculiarity of
the plant would not have been so noticeable.

A cross of Mohawk and Golden Carmine, a pole bean,
gave in the F, generation 7 plants of the bush type and
28 plants more or less pole like. In the notes no separa-
tion of pole and runner beans were made, probably due
to a lack of clear distinction between the two as occurs
in many crosses.

Cross Type 10. ALt x AIT

White Marrow by Burpee Stringless is presumably a
cross of this type. In one case the F is described as a
pole and in another as a runner bean. The F, genera-
tion results in 38 bush to 108 described as runner beans.
This is consistent with expected results when the plant
is described as a whole. The expectations are 12 pole
and more or less pole like beans and four more or less
bush like forms.

Cross Type 12 or 13. ALt+aLT or aLt

A cross of White Marrow, a runner variety, with Pro-
lific Black Wax, which belongs to the type with spreading
outstretched branches, gave 20 bush plants and 58 plants
of the runner and pole types, no differentiation being
made between the two. This is consistent with the as-
sumed constitutions.

Cross TYPE 14 or 15. ALt X alt or alT

White Marrow with Currie behaves according to ex-
pectation, giving in the F, generation 41 bush plants e or
f in type, and 52 of the runner or semi-runner type.

In the cross of Blue Pod by White Marrow and its recip-
rocal, neither variety being pole in type, climbing plants
apparently occur as well representative of most if not all
of the other habit types. Some lots isolated were very
erect, others spreading in various degrees; one lot is de-

556 THE AMERICAN NATURALIST [ Vou. XLIX

scribed as having long tendril-like shoots above the plant,
another along side of this had shorter shoots, perhaps
AIT. Among the lots, all degrees of climbing were devel-
oped; one plant encountered was evidently aLT like the
one mentioned in a previously discussed cross; plants
with more or less outstretched branches were noted.
Type notes on F, and F, generations of an earlier cross
in type; F, segregates into 25 bush forms and 62 runners,
are significant. The F, generation is described as pole
probably including pole beans of the F, type. The ratio
is disturbed by the lack of a clear understanding of the
true basis for classification of plant type in beans. ‘The
F, of another cross involving the same varieties is noted
as having 41 bush and 5 runner beans.

White Marrow and Burpee Kidney yielded two lots of
bush beans and two heterozygote lots giving 6 plants with
terminal inflorescence and 15 with axillary.

Red Valentine and White Marrow crosses give similar
results. In an early cross, the F, generation plants have
been grouped according to the general plant type, no at-
tempt being made to separate intergrading types. The
notes give the results of segregations as 75 bush and 136
runner beans. Later generation heterozygotes approach
a proportion of 9 runner to 7 bush beans. The apparent
behavior probably depends on whether the intermediate
types are classed as runner or bush. In the cross in
which only the F, generation was observed, only constant
bush types seem to have been isolated.

Cross Tyre 19 or 21. AIT X alt on alT

Blue Pod crossed with Burpee is a representative cross
of this type. Only in a few cases was the Burpee type,
plants with shoots high in the air, observed, as most lots
isolated were homozygous and erect. In the F, genera-
tion of an early cross, plants described as runners ap-
peared. The proportion was 3 runners tolbush. Heter-
ozygote lots descended from these plants segregated in
the same manner, totaled 18 bush and 71 so-called run-

No. 585] INHERITANCE OF HABIT 557

ners. The runners are probably really c in type or c
and d.

In the cross of Giant Stringless and Blue Pod the
parent types were both isolated. No semi-runners were
noted, as would be the case if the cross were No. 21 in
type.

Cross Type 25 or 26. Alt X aLT or alt

Refugee Wax is a semi-runner bean. The F, isolated
lots of this variety crossed with Blue Pod were all more
or less erect. Some lots homozygous for axial branch-
ing were isolated, many individuals of which showed
signs of climbing. The semi-running and climbing
branches were short, confirming the assumption that
neither variety used possesses the factor for a long axis.
The climbing tendency exhibited shows that there must
be strains of Blue Pod that possess T. Previous data are
in harmony with this.

Cross Tyre 29, 30, 34 or 35. aLT or aLt X alt or alT

Many crosses of bush beans with those of spreading
type give a 3:1 proportion in the F, and later hetero-
zygous lots.

In Keeny Rustless, a variety of the spreading type,
with its almost runner-like branches, by Red Valentine
some lots with the spreading habit have been isolated,
also more or less runner-like forms and one with the erect
habit of Red Valentine. The axial and terminal inflores-
cence is inherited in a 3:1 proportion. Notes on type in
one heterozygous lot show five erect and 10 plants with
outstretched branches.

In the cross of Black Valentine and Prolific Black Wax
one lot with outstretched branches was isolated; all
others were of the erect type.

In the cross of Blue Pod Butter and Prolific Black Wax
no spreading types with outstretched branches were
noted, but this is not surprising, as in an F, generation
the parent plants selfed for planting may not have hap-

558 THE AMERICAN NATURALIST [ Vou. XLIX

pened to be of the spreading type, thus giving homo-
zygous erect offspring.

In the cross of Golden Eyed Wax with Prolific Black,
outstretched branches due only to axial inflorescence
were noted.

Spreading plants of this nature also occur in the cross
of Bountiful and Prolific Black Wax. In the latter two
crosses the twining habit was more or less developed in
the longer branches.

Cross Type 34, 35, or 36. alT Xx alT, alT X alT, on
alt X alt

In the crosses of this type only erect bush beans with-
out runners or spreading branches, should occur, al-
though contorted stems might possibly appear. Such is
the behavior of the following crosses of this type:

Low Champion X Blue Pod Butter

Blue Pod Butter X Golden Eyed Wax and reciprocal

Blue Pod Butter X Mohawk and reciprocal

Challenge Black Wax X Warwick

Currie X Mohawk and reciprocal

Currie X Red Valentine

Blue Pod Butter X Warren

Bountiful X German Black Wax

In the crosses, Challenge Black Wax by Davis Wax and
Blue Pod Butter by Davis Wax, lots have been isolated
with short shoots above the plants somewhat resembling
the habit of Burpee Stringless and Giant Stringless.
This behavior is unexpected if such a plant type is to be
described by the formula AIT. The Davis Wax type
used in the crosses may, however, have been of a different
strain from that described in the table. This variety is
the only one used in the crosses that was not under the
observation of the writer, as its growth was discontinued
the year in which these notes were taken.

While the factors discussed above primarily determine
the plant habit, there are several others of secondary con-
sideration. No special notes were taken with regard to

No. 585] INHERITANCE OF HABIT 559

these. Some of them are mentioned in the following
paragraph.

The character of the habit type is somewhat influenced
by the amount of branching the plants exhibit; open,
loose, bush beans are the result of few branches; the
close, dense habit of some forms is caused by profuse
branching. The size of a plant to some extent influences
the habit, although not as much in small ones like Chal-
lenge Black Wax. In Warren the size of the plant prob-
ably causes it to droop. In some varieties the number
and weight of the pods, as well as their position, cause
some plants to droop and assume a spreading habit when
old. Perhaps fineness and coarseness of branching
affect habit.

One further matter that comes up for consideration is
the question of the effect of environment upon plant
habit. Its greatest effect, as would be supposed, seems
to be upon such quantitative characters as length of the
plant axis and probably the twining character to some
extent. Instances of adverse conditions resulting in the
almost total suppression of a character were noted in
plants grown on poor soil. They exhibited the slender
tips, typical of vines with axial inflorescence, but were
otherwise bush-like and erect. The accelerating effects
of very fertile soil on the growth of runner was also
noted. However, the environmental explanation for the
sudden appearance of runners among bush beans or of
pole beans among typical runners is open to question.
The most probable cause of such phenomena lies pri-
marily in the regrouping of the unit characters of habit,
combined at times with checking and accelerating factors
external to the plant.

The investigations here reported offer a foundation
upon which more extensive study on the subject might
be based.

The following table suggests a few important cross
types and the varieties which might be used to ad-
vantage:

560 THE AMERICAN NATURALIST [ Vou. XLIX

CROSSES FOR FURTHER STUDY

Type Plant
Number Varieties Type
2 Golden Carmine X White Marrow and reciprocal aXb

3 Golden Carmine X Burpee Stringless. and reciprocal aXe
+ Golden Carmine X Refugee and reciprocal axd
5or6 Golden Carmine X Keeney and reciprocal axe
7ors8 Golden Carmine X Challenge Black and reciprocal aX f
10 White Marrow X Burpee Stringlessand reciprocal b X ¢

b

þei ped jed pi p eed
AkReneoocwmrannwnne 2
À

bad

A

©

ia

jd

oO

11 White Marrow X Refugee and reciprocal xd
120r13 White Marrow X Keeney and reciprocal bxXe
White Marrow X Challenge Black and reciprocal b Xf

17 Burpee Stringless X Refugee and reciprocal exXd
180r19 Burpee Stringless X Keeney and reciprocal cxe
20or21 Burpee Stringless X Challege Black and reciprocal cxf
23 0r24 Refugee Keeney and reciprocal axe
25 0r26 Refugee X Challenge Black and reciprocal d Xf
290r30 Keeney X Challenge Black and reciprocal exf

The Burpee crosses should be particularly watched to
determine if the assumed set of factors AIT is the cause
of the shoots and later spreading habit of the plant.

The axis should be studied by means of accurate meas-
urement as far as possible. The judgment concerning
circumnutation would probably be necessarily more or
less indefinite.

In crosses 4, 5, 8, 9, 11, ete. the type number should be
determined.

The conclusions that can be drawn from observations
reported in the preceding pages are:

1. That plant habit in beans is largely determined by
the presence or absence of three characters which have
been designated by the letters A, L, and T.

1. A, the presence of axial inflorescence permitting an
indefinite growth, of the main stem and main branches,
and a terminal inflorescence causing definite growth.

2. The length of the axis L, an important factor con-
trolling plant habit and probably governed by a series of
two or more factors for a length L,, L,, ete., which behave
after the fashion of Emerson’s hypothesis for the inherit-
ance of quantitative characters.

3. The climbing habit is due to a factor for circum-

No. 585] INHERITANCE OF HABIT 561

nutation. This factor may be called T. The cause of the
various degrees of the climbing habit has not been deter-
mined with any degree of certainty. The contorted stems
of erect bush forms are probably caused by T.

II. The factors A, L and T may be present in any
possible combination, giving rise to the various habit
types of beans.

III. When the types are crossed among themselves
they behave approximately after the manner sketched in
Table IT.

BIBLIOGRAPHY
1. Emerson, R. A. Pie in Bean Hybrids. Rpt. Agr. Exp. Sta. Neb. 17
(1904), p 34—43.
2. Emerson, R. A. Inheritance of Sizes and Shapes in Plants. AMER. NAT.,
44 (1 910), pp. 736-46 (1910).

oo

. Emerson, R. A., and East, E. M. Inheritance of Quantitative Characters
in Maize. Wuiversity of Nebraska Agr. Exp. Sta. Research Bulletin 2
913).

(1
4, Jarvis, C. D aang Varieties of Beans. Cornell Uniyersity Agr. Exp.
ta. Bulletin 2
5. Tracy, Jr rican Varieties of Garden Beans. U. S. D. A.
in No. 109.

6. Webber, H. J. Preliminary Report on Pepper Hybrids. A. B. A.
Reports, VII and VIII, p. 188.

ON THE MODIFICATION OF CHARACTERS BY
CROSSING!

R. RUGGLES GATES

UNIVERSITY OF LONDON

Iw the early years of Mendelian discovery there was
much discussion concerning gametic purity in hybrids,
and the question whether unit characters are modified on
crossing was keenly debated. Convinced by the numer-
ous instances in which Mendelian characters appear to
be unmodified by crossing, many writers came to the con-
clusion that characters universally segregate without
being modified or ‘‘contaminated’’ by association with
other characters in the hybrid. That such a conclusion is
far too sweeping is, however, indicated by many later
results, and there is now a disposition to admit that
changes in a character or the breaking up of a character
may be effected through crossing. But some writers con-
tinue to look upon a unit character as an entity, which is
unmodifiable and indestructible by hybridization.

Notwithstanding the admitted belief of Bateson and
others that characters may be modified by crossing, I
know of no extensive body of evidence that such modifica-
tions take place except the work of Castle and Phillips
(1914) whose conclusions have not been fully accepted
and are chiefly concerned with modification by selection.
It therefore seemed worth while to direct attention to
certain experimental results of a somewhat different kind
which appear to show beyond cavil that modifications of
characters sometimes result from crossing. The matter
is an important one because it affects the old question of
the swamping of new characters through crossing, as well
as various other aspects of evolutionary theory.

1 Read before the American Genetic Association, San Francisco meeting,
August 3, 1915.

562

No. 585] MODIFICATION OF CHARACTERS 563

Anticipating the conclusions which will be reached in
this paper, it may be pointed out that the swamping effect
is not so serious a check upon progressive evolution as

might be supposed, (1) because blending or modification
of a new character only takes place in certain crosses and
may be accompanied by segregation even in some of
those, and (2) because Mendelian characters usually come
out ‘‘pure’’ when crossed with the form from which they
were derived. Hence when Mendelian characters arise
through mutations in nature it may be expected that they
will be able to perpetuate themselves and spread, espe-
cially when dominant, unless they place the organism at a
disadvantage in the struggle for existence. The modifi-
cation of a Mendelian character will come, not from cross-
ing with its parent form but with a more distantly related
species.

Some writers appear to believe that it is practically
impossible to modify a unit character because it is repre-
sented in the germ plasm by a ‘‘gene’’ whose essential
characteristic is its unmodifiability. But if we consider
that each unit character is a difference which has arisen
through a change in one element of the germ plasm, prob-
ably in a chromosome, then it would seem possible that if
introduced into a foreign cytoplasm the chromosome may
become subject to permanent modification.

Castle and Phillips (1914) have produced evidence
from hooded rats tending to show that selection may
modify a unit character in certain cases, although the
nature of this result is not yet fully analyzed. They
moreover show that the hooded character is modified by
across. Davenport (1906) in his experiments with poul-
try, concluded that unit characters are frequently modi-
fied by crossing. He says (p. 80):

Very frequently, if not always, the character that has been once
crossed has been affected by its opposite with which it was mated and
whose place it has taken in the hybrid. It may be extracted therefrom
to use in a new combination, but it will be found to be altered. This we

564 THE AMERICAN NATURALIST (Vor XLIX

have seen to be true for almost every characteristic sufficiently studied—
for the comb form, the nostril form, cerebral hernia, crest, muff, tail
length, vulture hock, foot feathering, foot color, earlobe and both gen-
eral and special plumage color. Everywhere unit characters are changed
by hybridizing.

In crosses between Ginothera rubricalyx and Œ. grand-
iflora I have studied with care the modifications which
take place in the expression of the various character-
differences in F,, F, and later generations. Many of the
results have been recorded in detail elsewhere (Gates,
1914, 1915a, pp. 250-282). It need only be said that the
foliage characters in F, form an absolutely continuous
series so that it is impossible to apply to them usefully
the unit-character conception. In F, a large number of
races were obtained differing in many ways as regards
their foliage, many of them breeding true and others
varying within wide or narrow limits. Occasionally in
back-crosses an apparently complete reversion takes
place to one or other of the parents, but blending and
fractionation of the characters is the rule.

It is, however, difficult to obtain critical evidence from
the foliage because, while the original differences are
sharply marked, yet it is always possible to assume that
the continuous F, series and the numerous F, races result
from the presence of many independent units.? I will
therefore confine my attention to the sharp pigmentation
character (R) of rubricalyx, for in the inheritance of this
character crucial evidence may be obtained. The origin
of this dominant unit-character through a single muta-
tion, and the subsequent attainment of the duplicate con-
dition (RR’) for this character in some of the offspring
of later generations (1915b), have been pointed out else-
where. Here we will examine the modifications of R
which take place when rubricalyx is crossed with Œ.
grandiflora.

The main facts regarding the variability of R in these

2 The inheritance of pubescence-differences shows similar features and can
not be reasonably interpreted in terms of numerous units.

No. 585] MODIFICATION OF CHARACTERS. 565

crosses have already been published (Gates, 1914, p. 244
and 1915a, p. 257) and need only be summarized here, to
emphasize their significance. In the publications cited
I had not yet reeognized that the occurrence of 15:1 ratios
in later generations of rubricalyx is significant as indi-
cating that in such families the duplicate condition for R
had been reached, even although other ratios such as
5:1 occur as well.

The F, generation of the crosses between rubricalyx
and grandiflora contained 2,794 plants, in 20 of which the
red bud-character R showed decided modification so as to
be more or less intermediate between the two parents.
Since each plant in bloom produces scores of buds simul-
taneously, and hundreds during the season, there is ample
material for determining the exact degree of modification
or development of the character in every individual. As
will be seen from the original records, the 20 plants in
which the color pattern was more or less modified were
not all alike but formed a series, some being nearer the
normal R than others. In most other F, plants sharp
segregation took place, the buds being entirely either R
or r without the slightest doubt in classification. In addi-
tion to the 20 plants above mentioned, there were, how-
ever, a certain number in which the character R was more
or less underdeveloped, so that it was impossible to be
certain whether they represented mere fluctuations or real
modifications of the character.

The crucial test of modification is supplied by the F,
generation. Two of these last-mentioned intermediate
plants self-pollinated yielded offspring like themselves,
without any tendency to segregate into the R and r types.
These families numbered, respectively, 283 and 20 plants,
so that in the former case at least any tendency to segre-
gation could not fail to be observed. The buds of these
plants were intermediate, the pigmentation was pale and
was never fully developed on the hypanthium as is the
case in rubricalyx. The whole population was then inter-
mediate like the parent.

566 THE AMERICAN NATURALIST [Vou. XLIX

Another F, family (No. 149) was derived from an F,
plant (65. III. 12) having sepals weak red with the color
pattern as extensive as in rubrinervis 6 (i. e., nearly the
extreme condition), and in addition streaks of pale red
on the hypanthium. This plant was therefore nearer r
than R, and one may account for its occurrence through
‘‘contamination’’ before segregation took place in the
germ cells of the previous generation. In pure rubri-
nervis or grandiflora I have never found even a trace of
red on the hypanthium until the flower fades. The off-
spring of this plant numbered 186 individuals and their
pigmentation fluctuated about that of the parent plant as
amean. This condition closely approximated that in Œ.
rubrinervoides (1915c, p. 390), which may have orig-
inated in a similar way.

We must, therefore, conclude that plants which are
intermediate in pigmentation breed true, at least in all
cases tested, and that the degree of pigmentation in the
parent is adhered to in the offspring whether the parent
plant is an under-pigmented R or an over-pigmented r.
In this aspect, the inheritance in such cases is quantita-
tive and the offspring vary only within narrow limits.

The quantitative aspect is further emphasized when F,
and F, hybrids of Œ. grandiflora and Œ. rubricalyx are
crossed back with either parent. The pigmentation is
much intensified when crossed back with rubricalyx, and
greatly diluted when crossed with grandiflora, Thus in
(rubricalyx X grandiflora) < grandiflora if the female
parent is heterozygous for R, segregation into R and r
plants will occur in the offspring, but the R plants will
be much paler than in the selfed offspring of the female
parent.

Hence there are two somewhat antagonistic effects
which have to be considered, (1) the segregation of R and
r individuals, and (2) a permanent dilution of the pig-
mentation of the R individuals. The former effect can be
explained by the meiotic mechanism which segregates

No. 585] MODIFICATION OF CHARACTERS. 567

chromosome pairs. The latter effect may be due to a
modification of the chromosomes themselves, or perhaps
of the surrounding cytoplasm, or the inhibition in pig-
mentation may be explained by the presence of more
numerous grandiflora chromosomes. Everywhere, in an
accurate study of the inheritance of R, the quantitative
as well as the qualitative (presence or absence) aspect
has to be considered.

The dilution effect from crossing back with grandiflora
has been tested in six families numbering 673 individuals
and is always essentially the same. Although segregation
into the R and r types takes place when the parent is
heterozygous, yet R once diluted always remains so and
apparently never gives rise to the original deeply pig-
mented condition. In other words, a permanently blended
condition arises as regards the depth of pigmentation,
although this will still segregate from the unpigmented
- eondition in heterozygous plants.

It is not easy to furnish a complete explanation for this
diluting effect. The permanent dilution of R through
union with a grandiflora germ cell may perhaps be ac-
counted for by the fact that in the heterozygote the chro-
mosomes of grandiflora are closely associated in the same
nucleus with those of the other parent. The chromosomes
which are finally dissociated in the germ cells, after thou-
sands or millions of mitotic divisions in association,
might then be supposed to be somewhat modified. There
are, however, difficulties with this view, since the absence-
character, r, is usually not contaminated, but splits out
sharply and almost invariably without any trace of red-
production.

It is also difficult to account for the facts on the
assumption that the cytoplasm has been permanently
modified.

There is, however, one hypothesis which appears to
meet the case. If all the grandiflora chromosomes are
equally effective in inhibiting anthocyanin production in

568 THE AMERICAN NATURALIST [ Vor. XLIX

the hybrids with rubricalyx—a not improbable hypothesis
—then the dilution effect will be the same in F, or in cross-
ing back, whenever an R chromosome is present in the
next generation; and when such a chromosome is not
present there will of course be complete absence from the
buds of the rubricalyx pigment. On this hypothesis, in an
original cross between rubricalyx and grandiflora a cer-
tain (observed) reduction in pigmentation occurs. When
the F, hybrid is crossed back with grandiflora the addi-
tional grandiflora chromosomes thus introduced dilute or
inhibit the color still further, while the presence or ab-
sence of the diluted R will depend upon whether or not
the R chromosome from rubricalyx is present. It would
_ thus appear to be unnecessary to assume that this chromo-
some is itself modified by its different nuclear and cyto-
plasmic environment.

In other words, the grandiflora chromosomes may be
supposed to exert a mass effect in inhibiting the influence
of the R chromosome. It is, of course, possible that in
these circumstances the R chromosome itself may be
permanently modified, but it seems possible to explain all
the facts without making this assumption. In any case,
whatever the modus operandi, there can be no question
that the R character is permanently diluted by crossing
with grandiflora, and the degree of dilution is increased
every time the hybrid is again crossed back with that
species.

Another noteworthy fact is that as the pigmentation
becomes more dilute its morphological expression is more
irregular. The color pattern of the bud begins to break
up, and instead of continuous pigmentation of the whole
bud a patchy effect will be produced. This spotted condi-
tion of the buds is very marked in certain families, e. g.,
in the second generation of offspring from (rubricalyx X
grandiflora) X grandiflora (see Gates, 1915a, Fig. 113, p.
280). When it appears it is found to persist in later
generations. To account for this condition through the

No. 585] MODIFICATION OF CHARACTERS. 569

accession of a ‘‘spotting factor’’ is a gratuitous assump-
tion. Spotting appears rather to be the manner of ex-
pression of the character when the amount of pigment is
small. It must be said, however, that in some families
having no greater quantity of pigmentation there is a
strong tendency for it to remain uniformly distributed,
so that the whole bud is very pale red.

LITERATURE CITED

Castle, W. E., and Phillips, John C. 1914. Piebald Rats and Selection; an
p A Test of the Effectiveness of Selection and the
siy ory of eaor ne in Mendelian Crosses. Carnegie

ubl. No. 195, pp. 56, pls. 3.
Davenport, C. z 1906. sii te in Poultry. Carnegie Publ. No. 52, pp.
136, pls. 17.

Gates, R. R. 1914. Breeding Experiments which Show that ape
and Mutation are Independent Phenomena. Zeitschr. f. Abst
. Vererb., 11: 209-279, Figs. 2
1915a. The Matation Factor in Evolution. Macmillans, London, pp. 353,
19156. On Successive teil casey pie Biol. Bull., 29: In press.
1915c. Some Cénotheras from Cheshire and Lancashire. Annals Mo.
Bot. Gard., ii por Pie 20-22

SHORTER ARTICLES AND DISCUSSION

STUDIES ON INBREEDING. VI. SOME FURTHER CON-
SIDERATIONS REGARDING COUSIN AND
RELATED KINDS OF MATING?

IN the first of these studies? the writer dealt with the results,
in so far as concerned coefficients of inbreeding, which would
follow continued brother X sister, parent X offspring, and cousin
X cousin mating. Regarding matings of the latter type it is de-
sired now to record certain further facts.

PEDIGREE TABLE I (HYP L)

To ILLUSTRATE THE CONTINUED BREEDING OF First-CousIN X First-CousIN
— SINGLE COUSINS

°
ae poe Ney
. > 2

ee ae =
oe ee ee ee

A <
m
g
n
d
0
h
Ta P
i m
k
n
$
w
a
v
Generation number 1 2 3 4

1 Papers from the Biological Laboratory of the Maine Agricultural Experi-
ment Station No. 85.
2 AMER. Nat., Vol. XLVIII, 1913, pp. 577-614.
570

x

No. 585] SHORTER ARTICLES AND DISCUSSION 571

There are, of course, two possible sorts of first cousins, single
and double. In the first case one of the parents of any individual
is a brother (or sister) to the one of the parents of the other indi-
vidual in the mating. In the second case, both the parents
occupy this relation to the parents of the other individual in the
mating.

These two sorts of first cousinship are shown in Pedigree Tables
I and II.

PEDIGREE TABLE II (HYPOTHETICAL)
To ILLUSTRATE THE CONTINUED BREEDING OF First-CousIN X First-Cousin
— DOUBLE COUSINS

k oe

g |, ja

i oe

ee

a tr

k G

i 1, f

; P

i

A, < i jo
g |

et

m ee

+ ala

ta 4 r

i | tp

J U

Ae {p

n {4

Generation number 1 2 3 | 4 5

The values of the coefficients of inbreeding for continued single
and double cousin mating are shown in Table I.

It will be seen that Pedigree Table I and the third column of
Table I are different from the corresponding values given on
pages 591 and 592 of the earlier paper. The present values
should be substituted for the earlier ones, which were based upon

-

572

THE AMERICAN NATURALIST

[ Vou. XLIX

the erroneous assumption that half the double-cousin values

would give single-cousin values.

TABLE I

VALUES OF THE SUCCESSIVE COEFFICIENTS OF INBREEDING IN THE CASE OF
CONTINUED COUSIN MATING

Coefficient of Ancestral Generation Coefficient fo for Single | Coefficient for Double
Inbreeding Included Pins usins — _ Cou usins
Zo 3 0 o
Z, 2 0 0

Ze 3 25.00 50.00
Z; 4 50.00 5.00
Z, 5 68.75 87.50
Z; 6 81.25 93.75
Ze T 89.06 96.98
Z, 8 93.75 98.44
Zs 9 96.48 99.22
Zo 10 98.05 99.61
Zio 11 98.93 99.80
VE 12 99.41 99.90
Zy> | 13 99.68 99.95
Zis 14 99.83 99.98
Ziu 15 99.91 99.99
Zis 16 99.95 99.994
00 — ee
se as ee eatin
Dg oe
7
0 A Ka re y
Ere 7
,
ys Pa
9 Y ` Sa
K % t
Š oo | S
N / f ri
Ñ 3
T ig)
Sy 4
v i Í
/ r
p? /
20 + l É
/ i
‘WV
2
4 E 2 70 7 74
GENERATIONS
Curves of inbreeding, showing (a) the limiting case of continued

Fic. 1.
brother x sister breeding, fodder the successive coefficients of hoj have the
aximum values; (b) c ued p ing ma cng fe) continued first-
pei first-cousin aces pane the cousinship a double (C2 x C?), ang a con-
tinued first-cousin x first-cousin m e the cou ate is single (Ctx Ct’).
The continued mating of Se pam Sdan re same curve a

No.585] SHORTER ARTICLES AND DISCUSSION 573

The data of Table I are given graphically in Fig. 1, together
with the curve for brother X sister and parent X offspring.

From the table and figure it is seen that with continued in-
breeding according to any one of these four types the coefficient
approaches the value 100. The rate of approach is different,
however, in the different cases. The curves fall into two pairs.
The brother X sister and the double cousin curves are precisely
alike so far as concerns their curvature or shape at any given
point. Similarly, the parent X offspring and single cousin curves
are of the same shape. The essential point of difference is that
the cousin curves lag a generation behind the others.

Let us now consider the question of the degree of inbreeding
following continued matings of the avuncular type of relation-
ship. Pedigree Table III gives a pedigree in which each mating
is of uncle X niece.

PEDIGREE TABLE III (HYPOTHETICAL)

To ILLUSTRATE THE MATING OF UNCLE X NIECE

| | fu

| m is
poco
c A
0
h | n
i k
t 5 je
d í i?
j |
i
>. 4 4 1s
; í ts
h i
: s
: i +.
b 4 r
pa fe
te | {
f i (i
l d
an

|

~
or

Generation number 1 2 a4

574 THE AMERICAN NATURALIST [Vot XLIX

From this table it appears that the values of the coefficients of
inbreeding will be exactly the same for this type of mating as in
the case of single cousin mating. Or, in other words, Z’s form
the following series.

TABLE II

VALUES OF COEFFICIENTS OF INBREEDING FOR CONTINUED
X NIECE MATING

Coefficient Number of Ancestral Generations Value of Coefficient

Zo 1 0

Z, 2 0

Z, 3 25.00

Ze 4 50.00

7A 5 68.75

B 6 81.25

ete. ete. etc. as in Table I

From the data presented in this and former papers it is clear
that inbreeding continued for about ten generations, quite re-
gardless of the type of mating, provided only it be continuously
followed, leads to within one or two per cent. of complete ‘‘con-
centration of blood.’’ The bearing of this result upon the general
question of the degree of inbreeding which exists in the ancestry
of our domestic animals to-day is obvious. To consider but a
single case: In 1789° a law was passed prohibiting the importa-
tion of cattle into the Island of Jersey. Hence it follows that all
pure-bred Jersey cattle of the present time must be of the
descendants of the relatively few animals on the Island in 1790.
Taking three years as about the average generation interval in
eattle, this means about forty generations since the Island was
closed to importation. The concentration of lines of descent
which must have occurred in this time merely by the dropping of
lines and quite regardless of the type of mating is obvious. This
is not the place to go in detail into the discussion of inbreeding in
Jerseys, especially as I hope shortly to publish the results of an
extensive study of this matter, but it seems desirable to emphasize .
the bearing of such hypothetical pedigrees for particular types
of mating as are given in this and earlier papers, on the general
problem of inbreeding.

It is possible to extend now somewhat the table of general
equations given by Jennings‘ for coefficients of inbreeding after

8 Teste Rees’s odo Sagas and H. S. Redfield, Natl. Stockman and
Farmer, December 15, 1892.

4 Amer. NAT., Vol. XLIII, p. 695, 1914.

No. 585] SHORTER ARTICLES AND DISCUSSION 575

m generations of each particular type of mating. We have the
following values, where n denotes the number of ancestral gen-
erations concerned, or, as Jennings puts it, the number of suc-
cessive inbreedings which have taken place.

Type of Mating Coefficient of Inbreeding
2n — 1
E aT an a O E O a eee on
2n — 2
BrOK Or OC MSLOY eLan a i ees ek
Ən
Gow i z 2n — 2n
omn X COUMM, MINIS cii: ccc s wee e neces’ SE TEF
x i Qn — 22 (from n==2
Cousin X cousin, GOW oe ec es eae E ek ln a i nao
r on — Nn — I
Parent X offspring ........... Serre re on
Vacio x Pow o no an on es

RAYMOND PEARL .

AN ATTEMPT TO PRODUCE MUTATIONS THROUGH
HYBRIDIZATION

THERE is no more interesting problem to the experimental
evolutionist than the one relating to the cause or causes of the
origin of mutations. Until we are able to solve this problem we
can only accept what the gods give in our breeding experiments.
When a mutation arises it is usually a simple process to produce
a pure stock. By mutation is meant any deviation from the
normal type which reappears in some of the descendants. In
the following experiment most of the abnormalities that were
found never reappeared in the offspring.

My experiments have been confined to the fruit fly, Drosophila
ampelophila, a species kept for years ‘‘under cultivation’’ at
Columbia University. This species has proved to be very plas-
tic, throwing off great numbers of mutant forms. At the sug-
gestion of Dr. T. H. Morgan I crossed some of these mutants
with wild stock of the same species from widely separated locali-
ties in order to test whether through hybridization mutations
arise in greater numbers than in inbred stock.

The idea that new forms arise from crossing more or less
closely related species is an old one. One finds many references
in Darwin’s works to this conception. For instance, in the
** Origin of Species °? Darwin says:

576 THE AMERICAN NATURALIST [ Von. XLIX

When mongrels and the more fertile hybrids are propagated for
several generations, an extreme amount of variability in the offspring
in both eases is notorious; but some few instances of both hybrids and
mongrels long retaining a uniform character could be given. The vari-
ability, however, in the successive generations of mongrels is, perhaps,
greater than in hybrids.

One of the causes of ordinary variability ... is ... that the repro-
ductive system from being eminently sensitive to changed conditions of
life, fails under these circumstances to perform its proper function of
producing offspring closely similar in all respects to the parent form.

From ‘‘ Plants and Animals under Domestication ’’ we find
the following.

Crossing, like any other change in the conditions of life, seems to be
an element, probably a potent one, in causing variability.

A variation to be effective in species formation must reappear
in some of the descendants. That a variation could, through
selection within a pure strain be increased or decreased in the
direction of selection to form a stable species has been seriously
questioned since Johannsen’s classic experiments. It is well
understood, on the other hand, how selection in a mixed popula-
tion could cause the variation to move in the direction of selec-
tion up to a certain point.

The first mutant stock selected for the experiment was cherry
club vermilion. The factors for these three characters are
linked together and are also linked with sex; the second stock
was black pink bent, which has the three factors independent of
each other and none is linked with sex. These factors are sup-
posed to lie in the second, third and fourth chromosomes, re-
spectively. The third stock was black purple vestigial are speck,
which has the five factors linked together. They lie in the sec-
ond chromosome. A stock from France was crossed to the mu-
tant stock several months after the other crosses were made, and
eosin tan vermilion was substituted for the cherry club ver-
milion, and pink kidney sooty rough for the black purple ves-
tigial are speck stock because flies of these particular stocks
were not to be had at the time desired.

These forms were chosen because it was thought that if muta-
tions do arise from hybrid forms there would be more probability
of their origin from a mutant varying in several characters when
crossed to wild than if it varied in only one character. Also by

, using stock containing several recessive characters a check could

No.585] SHORTER ARTICLES AND DISCUSSION 577

be placed upon any variant from the expected classes due to
contamination; for the variant, if arising from the cross, would
give some offspring in the F, generation with some of the reces-
sive characters. However, extreme care was taken to avoid con-
tamination and at no time was there reason to suspect it in any
of the cultures.

The wild stocks used were from Arkansas, California, Massa-
chusetts, Illinois, Minnesota, Ohio, Wyoming, Porto Rico, Cuba,
Australia and France. The totals of the F, generations are as
follows:

Ch. Cl. Ver. Bi. Pk: B: Bl. P. Vg. Are. Sp.
pe eee ave ts 1,162 307 198
California S20) Deri e od 859 715 332
iois <2. 5 eaaa serene 211 287
Massachusetts ..........». 1,078 681 1,013
PEO Ss oes vues Fags s 771 274
Me eo ee pres eer 506 1,612 370
Wyoming o oo). PEERS Hes 925 150
Porka Aeteo sons ei a 151 207
CRBS se sean aon es Gena 819
PTE o ety « s 469 401 548
PPBNCG oe ce tes 814 ‘951 826
OO eas cee 6,946 “5,298 4,393

This gives a grand total of 16,637 flies. It should be noted
that these flies were examined with the greatest care under a
binocular microscope. Each fly was turned over separately and
every part carefully examined.

From the cherry club vermilion crosses the following ab-
normal forms were found; three gynandromorphs; twenty-four
flies with more or less beaded wings; two flies with three cross
veins on the wings; one truncate; and two flies with abnormal
abdomen.

The abnormal forms from the crosses with black purple ves-
tigial are speck were, sixty-three with more or less beaded wings;
one truncate; one abnormal abdomen; one fly with five legs; and
four flies with a projection from the posterior cross vein toward
the base of the wing.

From the black pink bent crosses were found two beaded; one
abnormal abdomen; three truncate; and one called furrowed be-
cause of the furrows in the eyes due to the foreshortening of
the head.

This gives a total of 109 abnormal forms or one abnormal in

578 THE AMERICAN NATURALIST [ Von. XLIX

every 152 flies. But 89 of these abnormals were flies with beaded
wings. This character is very variable; some of the flies had
only a few bristles missing from the margin of the wings, while
others had both the outer and inner margins of the wings ser-
rated. The character has been recurring in the stock so fre-
quently that it can scarcely be ascribed to outcrossing. Many of
these flies were mated, but they either did not leave offspring,
or the character did not reappear in the F, generation.

The three gynandromorphs are not to be considered as mu-
tants. The data here show that gynandromorphs occur once in
about five thousand five hundred times.

Flies with truncate wings are of occasional occurrence in the
laboratory stock, as are also those with abnormal abdomen; hence,
flies with these characters are not to be considered as due neces-
sarily to the outcrossing. The truncate would not breed and the
abnormal abdomen character did not reappear in the F, genera-
tion. If a character does not reappear in the F, generation it is
considered to be of somatic and not of germinal origin, unless
an environmental condition is necessary for the expression of
the changed character. -

The abnormality of the fly with five legs may have been the
result of accident, for the character did not reappear in the F,
generation.

Three characters were found to be inherited; the one called
‘‘furrowed,’’ which arose from the cross of black pink bent with
wild stock from Massachusetts; the one with a projection from
the posterior cross vein toward the base of the wing, called
‘* tau,’’ which arose from the cross of black purple vestigial are
speck with wild stock from Illinois; but since this stock had
just been received from Illinois, and since the character appeared
in four of the flies, it is suspected that the character was reces-
sive in the wild stock and not due solely to the cross. Also from
cherry club vermilion crossed to stock from Arkansas arose two
males with three cross veins on the wings and a disturbance of
the ommatidia of the eye. This character is called ‘‘ warty.’’

Pure stocks of flies with these characters have been bred for
many generations and each continues to breed true. ‘‘ Warty ”
has many other characters than the modification of the eyes,
e. g., beaded wing, spread wing, from two to five cross veins
on the wings, abnormal abdomen and disarranged hairs on
the thorax. The females are sterile and the race is maintained

No. 585] SHORTER ARTICLES AND DISCUSSION 579

by crossing the males to their heterozygous sisters. The char-
acter is not sex linked; it decreases the viability of the flies, but
more than this can not be said at present. Work is being con-
tinued on this character and on flies with the character ‘‘tau.”’

‘‘Furrowed’’ is characterized by having the head foreshort-
ened, which causes indentations or furrows in the eyes; also the
spines on the scutellum are stumpy. The last character is of
importance in determining some of the flies, as a female will
sometimes occur without any disturbance of the eyes.

This character arose in a male which was crossed to a wild
female. The F, generation gave normal females and half the
males were normal and half were furrowed. This established
the fact that the character followed the distribution of the sex
chromosome. The position of the gene in the chromosome was
next determined according to the theory that the genes in any
chromosome are arranged in a linear series.1_ Crosses were made
with eosin miniature, sable forked, and with vermilion barred.
Because of the low fertility of the furrowed females the cross
was always made with the furrowed males. Consequently, the
males alone are considered in the counts given below.

EosIN MINIATURE? By FURROWED ĝ

MOPE sS u aides ck why S POONA it a. esas 67

Eosin miniature furrowed. 1 Eosin miniature ........... 75
F, males.. : ps

Wom Tone sue. io SMA oe ek es 31

Miniature furrowed ..... 0 Eosin long furrowed ....... 28

In the first column are the cross-over classes between mini-
ature and furrowed and the per cent. of these to the whole
number is 3.4. Then the gene which determines the character
‘‘furrowed’’ is supposed to lie 3.4 points beyond miniature, or
at 39.6.

SABLE FORKED 9 BY FURROWED ¢

Furrowed sable forked... 1 Sable forked ...........:.. 61

F Í MOP eekan o TO o tein clea shes 105
2 males.. Worked N E a 3 Furrowed forked ........... 1
pisa Ma sieves O Gable (oo a et 16

In the first column are the cross-over classes between furrowed
and sable and these are 5.7 per cent. of the entire number.
Then furrowed lies at a point 5.7 to the left of sable, or at 37.3.

1 Sturtevant, Jour, Ex. Zool., 713.

580 THE AMERICAN NATURALIST [ Vou. XLIX

VERMILION BARRED 9 BY FURROWED ¢

f PET ee eek ee 6 Vermilion Bar. oer GG 86
Vermilion furrowed ..... Do Worrowed siok ae 102
F, males.. Mom i aa 0O Furrowed bar.. asina na 15
Vermilion furrowed bar... 0. Vermilion ................+ 28

The cross-over classes between vermilion and furrowed are
the bar and vermilion furrowed classes of which there are nine,
which is 3.75 per cent. of the entire number. Vermilion is at
33, hence the gene for furrowed lies at 36.75.

The cross-over classes between furrowed and bar are the fur-
rowed bar and the vermilion classes of which there are 43 which
is 18 per cent. Then furrowed lies/18 points to the left of bar
or at 39.

The discrepancy in these results is due to the low viability of
the furrowed flies, yet the results agree fairly well, varying from
36.75 to 39, giving an average of 38.1, which is considered as the
relative position of the gene for furrowed in the sex chromosome.

` H Br

The accompanying diagram will aid in understanding the
cross-over classes. The heavy straight lines represent the paired
sex chromosomes which a heterozygous female has received from
her parents. The upper one, which carries vermilion bar, was
received from the female parent and the lower, carrying fur-
rowed, was received from the male parent. Each of the sons
of this heterozygous female receives one of these chromosomes
which determines what it shall be with reference to these special
characters. In about 75 per cent. of the cases the sons receive

ese chromosomes without any interchange of substance be-
tween the two as is shown by the two straight lines which rep-
resent the non-cross-over classes. When there is an interchange

No. 585] SHORTER ARTICLES AND DISCUSSION 581

of material between the two chromosomes as indicated by the
crossed lines, then males arise with a different arrangement of
the characters from that which had appeared in the grand-
parents.

In the diagram v, f and Br stand for vermilion eye, furrowed
eye, and bar eye, respectively; while V, F and br stand for
the normal allelomorphs of these characters, i. e, red eye, not
furrowed and not bar. Reading from the left the top dotted
line includes v, F and br, but since F and br are normal the flies
will differ from normal forms in the one character alone, viz.,
vermilion. The dotted line below includes V, f and Br, hence
the males receiving this chromosome are furrowed bar. Re-
ferring to the table showing the cross between a vermilion bar
female with a furrowed male we see that there were 28 vermilion
and 15 furrowed bar flies. Reading from the left again and
omitting the normal allelomorphs, the upper dash line includes
vermilion and furrowed and the lower dash line includes bar
alone. The table shows that there were only three vermilion
furrowed and six bar males, hence the interchange of material
between vermilion and furrowed took place less frequently
than it did between furrowed and bar. Since the per cent.
of crossing over between any two genes is taken as the index of
the relative distance between those genes, then furrowed lies
much closer to vermilion than it does to bar.

The fine lines represent double crossing over, of which no
representatives were found in this cross.

SUMMARY AND CONCLUSIONS

Crosses were made with mutant stocks of Drosophila with
wild stock from many localities in the United States, from the
West Indies, France and Australia in order to discover, if pos-
sible, if hybridization is an essential factor in the formation of
mutant races. From 16,637 flies of the F, generation seven
flies arose which varied from the normal type and which bred
true. If we discard the four with the character ‘‘tau’’ for
reasons given above, then the result is narrowed to three flies
with two characters. This gives one mutant to every 5,545 flies.
Therefore, a mutation has occurred so seldom that we can
scarcely attribute hybridization as its cause. It is highly prob-
able that if the same number of wild flies had been reared under

582 THE AMERICAN NATURALIST [ Vor. XLIX

favorable conditions for the survival of any new forms that ap-
peared just as many mutations would have been found as in the
above experiment. In the light of these results we can attribute
the origin of mutations only to chance, since hybridization as a
causal agent does not occupy a privileged position relative to
the effect.

F. N. DUNCAN

COLUMBIA UNIVERSITY
wt i

LINKAGE AND SEMI-STERILITY

Tue Florida velvet bean (Stizolobium deeringianum) has nor-
mal pollen and embryo-sacs; it flowers (when sown in May)
early in September; and has pigmented (mottled) seed-coats.
The Yokohama bean (Stizolobium hassjoo) has also normal pol-
len and embryo-saes; it flowers in July; and has its seed-coats
unpigmented. The first-generation hybrids of Florida by Yoko-
hama had half their pollen and embryo-saes aborted (1, 2);
flowered at the end of August; and had more or less pigmented
seed-coats. In the second generation, half of the plants had
normal pollen and embryo-sacs, and half showed semi-sterility
(1, 2). These plants flowered from July to September, the ma-
jority being late. About three-quarters had pigmented seed-
coats; and one-quarter, colorless seed-coats.

Most of the semi-sterile plants, and also most of the plants with
pigmented seed-coats, were late in flowering. The semi-sterile
plants, however, were not later than the fertile, in the second
generation of the Florida by China cross. Hence there is no
necessary connection between semi-sterility and lateness. A ran-
dom sample of five second-generation plants of the Florida by
Yokohama cross gave one family with pigmented seed-coats, one
family with colorless seed-coats, and three families segregating
into pigmented and colorless in about the ratio 3:1. Hence the
pigmentation of the seed-coat is not a mere physiological conse-
quence of lateness, but is determined by a definite factor. If K
is the factor from the Florida concerned with semi-sterility ; P,
a factor concerned with pigmentation of seed-coat; and H, the
main factor for lateness; then K and H are strongly coupled in
the gametes of the first-generation plants, as are also P and H.
K and P show secondary coupling.

No.585] SHORTER ARTICLES AND DISCUSSION 583

The data follow.
SEMI-STERILITY AND LATENESS
Second generation of Florida by Yokohama
eeds n early in May
(Classes are approximately fortnights)

oe ee eee M
Fertile plants .............0-+- uio n]? Bing | u
Semi-sterile plants ...... 6 10 11 29 10 | ie

The average of the semi-sterile is about a fortnight later than
that of the fertile. If we divide the plants into those flowering
before and after August 11, we have:

| First Month | Second kud Third Mouths
Fertile | 43 38
Semi-sterile .... | 16 59

A calculation, based on the hypothesis used for semi-sterility
(1), shows that the crossing-over (3) between K and H is prob-
ably less than 17 per cent.

SECOND GENERATION OF FLORIDA BY YOKOHAMA

Seeds sown early in June

ajej a| | e | tek
Fertile tilaisi’ 8 oe
Semi-sterile A TiS ae ee 88

The fertile plants are earlier than the semi-sterile; though the
average difference is less than in the early planting, because, as
usual, the first-early plants are more affected by late planting
than are the later plants. .

PIGMENTATION OF SEED-COAT, AND LATENESS

Second Generation of Florida by Yokohama
arly sowing

eros Le | 5 | 6 | Totals
Unpigmented seed-coats .. | 16 | m6 | ol b | rio
Pigmented seed-coats........ 4 | 22 | 25 tR DI m

Thus most plants with unpigmented seed-coats are early. A
calculation again shows that the amount of crossing-over is prob-
ably under 23 per cent.

584 THE AMERICAN NATURALIST [ Vou. XLIX

SECOND GENERATION OF FLORIDA BY YOKOHAMA

Late sowing

6 | Totals

|
2
1

Ti
|
|
|

|
Unpigmented seed-coats ............ | 8 | 20 | 12 |
$1 eae |

1 43
Pigmento saisir itirirdi iraa | 17 3 HS

This confirms the results from the early sowing.

PIGMENTATION OF SEED-COAT AND SEMI-STERILITY

The coupling between K and P is given from the following:

SECOND GENERATION OF FLORIDA BY YOKOHAMA

Pigmented | Totals

| Unpigmented

| meae
Site o | me 110 | 155
Bemisterile 2 arare 39 120 | 159

The excess of pigmented semi-sterile and of unpigmented fer-
tile testifies to a slight coupling, and calculation shows that there
is probably about 35 per cent. of crossing-over.

According to the hypothesis (1), fertile second-generation
plants should be mainly homozygous for H (or h) and P (or p);
while semi-sterile plants should be mainly heterozygous for these
factors. This is being further tested.

JOHN BELLING

FLORIDA AGRICULTURAL EXPERIMENT STATION

REFERENCES
1. Belling, J.
1914. The mode of inheritance of semi-sterility in certain hybrid
plants. Zeitschr. f. ind. Abst.- u. Verebungslehre 12; 303-
342.

2. Belling, J.
1915. Inheritance of partial sterility. Report of Fla. Agr. Exp. Sta.
for 19 Pp. 96-105.

3. Sturtevant, A. H. -
1915. The Behavior of the Chromosomes as Studied through Linkage.
Zeitschr. f. ind. Abst.- u. Vererbungslehre 13: 234-287.

VOL. XLIX, NO. 586" OCTOBER, 1915

Ham

. Shorter Articles and EEE » Anticipatory Mutationist. Dr. R.

THE
AMERICAN
NATURALIST

A MONTHLY JOURNAL
Devoted to the Advancement of .the Biological Sciences with
Special Reference to the Factors of Evolution

CONTENTS

Early Portrayals ofthe Opossum. Dr. CHARLES R. EASTMAN. - —- - 586

Seventeen Years Selection of a Character. Dr. RAYMOND PEARL - - -595

Specific and Varietal Characters in Annual Sunflowers. Professor T. D. A.
omen -=

The Inh Dl in Matthiol d Petunia. I. The Hypotheses
HowaRD B. FROST - -=

The Coal Measure Amphibia, and EE Dr. Roy L. MOODIE - 637

RUGGLES GATES ~ 645

THE SCIENCE PRESS
LANCASTER, PA. GARRISON, N. F.
NEW YORK: SUB-STATION 84

The American

Naturalist

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

MSS
sent to the Editor of THE AM
Short articles containi

ERICAN NATURALIST, Garrison-on-Hudson, Ne
ing summari

w York.

es of research work bearing on the

problems of So as evolution are especially welcome, and will be given preference

in poran

rea reprints of mpfi am are supplied to authors free of charge.

pt
Further ‘reprints will be supplie

ns and advertisements should be sent to the publishers. The

aarin aries is four dollars
anadian postage
forty cents.

a yea
e twenty-five cents additional.
The advertising rates are Four Dollars for a pa

Foreign postage is fifty cents and
The cha hoe for single copies is

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered as second-class matter, ga 2, 1908, at the

Garrison, N. Y.

Post Office at Lancaster, Pa., under the Act ot

mgress of March 3, 1879.

FOR SALE
ARCTIC, ICELAND and GREENLAND
BIRDS’ SKINS,
Well Prepared Low Prices
Particulars of
- DINESEN, Bird Collector
Husavik, North Iceland, Via Leidle, England

JAPAN NATURAL HISTORY TE at:
Perfect Condition and Lowest Price
Specialty: Bird Skins, Oology, ae Marine
Animals and others. Catalogue free. Correspo ond-
ence solicite
oua É;
nosu, Saitama, Japan

T. FUKAI,

BOOKS FOR SALE

relating to all branches of

NATURAL HISTORY

Catalogue on application

JOHN D. SHERMAN, Jr.

403 Seneca Avenue MOUNT VERNON, N. Y.

The ory of Chicago

apg ya stig
s, Literature, Science,
Commerce and Admi nistra-

ion, Law, Med , Educa-
tion, and Divinity. Instruction
is given by regular of
University staff which isaugmented
in the r by appointment of
profess fr
Res asa Saan 15
st Term June 21 28
2d Term July ma

Marine Biological Laboratory
Woods Hole, Mass.

INV ESTIGATI Facilities for research in frie a
hs ON Embryology, Physiology
tire Year Seve

such a table is

we a

INSTRUCTION

July—August ology,
of

SUPPLY
DEPARTMENT
Open the Entire Year —

GEO. M. GRAY, Curator, Woods Hole, Mass

The annual announcement will be sent on ne
The Director, Marine Biological Laboratory, Woods

Mass-

THE
AMERICAN NATURALIST

VoL. XLIX. October, 1915 No. 586

EARLY PORTRAYALS OF THE OPOSSUM

DR. CHARLES R. EASTMAN

AMERICAN MUSEUM OF NATURAL History

THE quaint animal figures found in olden time works
on natural history are interesting not only as bearing
upon the contemporary state of zoological science and the
art of book-making, but also because many of the illus-
trations belong to a regular sequence or lineage which can
be traced back, like the textual descriptions, to primitive
sources. To a certain extent this has already been done,
or at least indicated, in the work by John Ashton, entitled
‘‘ Curious Creatures of Zoology.”

A subject deserving of the attention of naturalists but
which appears to have been neglected, is an historical and
systematic investigation of animal figures introduced in
early American cartography. Thanks to the magnificent
facsimile reproductions of sixteenth century maps which
have been published during recent years in this country
and abroad, abundant materials for this purpose are now
easily accessible. As for the ‘‘relaciones’’ of early voy-
agers and travelers in the western world, very few of
these have been published with scientific commentaries,
and among the really important seventeenth century writ-
ers on Central and South American natural history, only
the works of Hernandez (1628) and Maregrav' (1648)
have been systematically annotated. The first letter

1See the commentaries on these authors by Lichtenstein and Martius,
1827 and 1853, in the publications of the Berlin and Bavarian Academies
of Science.

£85

586 THE AMERICAN NATURALIST [Vou. XLIX

written from the newly discovered world, by Dr. Chanca,
companion of Columbus, was not adequately edited and
annotated until after four centuries had passed. Ves-
pucci’s letters also are deserving of mention in this con-
nection.”

In view of the fact that several communications have
appeared in Nature during the past year concerning the
first mention of the opossum in literature, it may not be
inopportune to trace the pedigree of some of the early
illustrations of this animal, both in maps and in printed
works. At the same time a few of the older printed de-
scriptions of American marsupials may be noticed. And
we will observe first of all that the earliest reference to
the common American opossum is found in the famous
collection of voyages published in 1504 by Angelo Trivi-
giano, under the caption of ‘‘Libretto de Tutta la Naviga-
tione de Re de Spagna, de le Isole et Terreni Novamente
Trovati.” In Chapter XXX of that work it is mentioned
that a live specimen, taken by the Pinzons in Brazil in
1500, was exhibited in Granada.

In Decas II of Peter Martyr’s ‘‘De Nove Orbe,’’ pub-
lished in 1511, occurs the first published description of
the American tapir; and immediately following this the
opossum is referred to in these words:

There is also an animal which lives in the trees, feeds upon fruits,
and carries its young in a pouch in the belly; no writer as far as I know
has seen it, but I have already sufficiently described it in the Decade
which has already reached Your Holiness before your elevation, as it
was then stolen from me to be printed.

In 1547 and 1548, and again from 1549 to 1555, Hans
Stade of Homburg, Hesse, passed some time in Brazil,
and wrote or dictated an account of his strange adven-
tures, which was published at Marburg in 1557. Under
the caption of ‘‘Servoy,’’? Chapter XXXII, we read:

2 See Fernandez de Ybarra in Journ. Amer. Med. Assoc. for September,
1906, and in Mise. Coll. Smithson. Inst. for the same year. Vespucci’s first
letter (1497) was republished in facsimile by Varnhagen in 1893, having
for frontispiece a design by Stradanus dating from about 1580, in which
various South American animals are well represented. Mention occurs 1
this letter of the iguana, puma and ocelot from the coast of Tampico.

No. 586] EARLY PORTRAYALS OF THE OPOSSUM 587

There is also a kind of game, called servoy, which is as large as a
eat, and has a tail like a cat; its fur is gray, and sometimes grayish
black. And when it breeds, it bears five or six young. It has a slit in
the belly about half a span in length. Within the slit there is yet

Dic Fugnr Datt, Cop. vevi

Tiaka Cap. pvei.

S bat auch eyn art Wiless peyffet Gerwoy/ift (o grof
E wie cyn Fane/weifaraw vobarensauch (d eps sive
bat enen (hwang wiseyntass. Unnd wann es gtberet/

Fic, 1. The “ Dattu” (tatou or armadillo) and “ Servoy” (opossum); after
Hans Stade, 1557.

another skin; for its belly is not open, and within this slit are the teats.
Wherever it goes, it carries its young in the pocket between the two
skins. I have often helped to catch them and have taken the young

ones from out of the slit.

588 THE AMERICAN NATURALIST [Vor. XLIX

In the original edition of the work just quoted wood-
euts are given of both the opossum and armadillo (Dasy-
pus novemcinctus Linn.) and these are reproduced in the
present article (Fig. 1) from a copy belonging to the New
York Public Library. The armadillo is thus described in
Stade’s ‘‘ Wahrhaftig Historia’’:

There is another sort of animal found in
savages call dattu; it stands about six inches high and is nine inches
long; its body is covered all over, except underneath, with a kind of
armor. This covering is horn-like, and the plates overlap one another
like those of chain armor. This animal has a very long snout, and 1s
usually found on rocks. It feeds on ants. Its flesh is sweet and I have
often eaten of it.

Two works published at about the same time as the
narrative of Stade also contain mention of the opossum,
the name of ‘‘Simivulpa’”’ or Fox-ape and ‘“‘Su” being

this country which the

applied to the creature. In the Italian edition (1558)
of Sebastian Miinster’s ‘‘Cosmographia’’ occurs this
passage, accompanied by an illustration which we have
reproduced in Fig. 2:

No. 586] EARLY PORTRAYALS OF THE OPOSSUM 589

Trovasi in quel luogo [Brazil] un animal prodigioso, le cui parti
davanti si rassomigliano a volpe & quella di dietro à Simia mai suoi
piedi sono como di huomo, ha le orecchi di civetta, & sotto le ventre
como una borsa, nella quale tien nascosti suoi figliuoli, finche crescono
di sorte che possino caminare securamente da lor stesi, & procurarsi il
cibo senza tutela della madre, ne mai escono di quella borsa se non
quando lattano. Quest’ animale mostruosa con tre suoi figliuoli fu
portato in Sibilía & indi in Granatá.”—p.

Münster’s illustration of the PE is evidently
derived from figures of the opossum appearing in several
editions of Ptolemy’s ‘‘Geography’’ from 1522 onward,

ae: npud of tesa
Sop

Cangila

Fig. 3. Earliest known figure of the opossum; from the Waldseemüller world-
map of 1516.

and other early maps of South America, all traceable in
the first instance to Waldseemiiller’s world-map of 1516,
where the same representation occurs (Fig. 3). It is
there accompanied by essentially the same legend as one
finds in the ‘‘ Tabula Terre nove” of the 1522 Ptolemy,
and in later maps and atlases,’ such as Cornelius de
Jode’s (1585), and van Linschoten’s (1598).

3 Modern reproductions of South American maps showing these figures
may be seen in Winsor’s ‘‘ Narrative and Critical History of America,’’
and in the magnificent na published by the Brazilian government
under the direction of Baron de Rio Branca. The representation of a Bra-
zilian landscape in the Ca sae no map of 1500, shown in our Fig. 4, is from
a photograph of Harrisse’s colored reproduction.

THE AMERICAN NATURALIST [Vou. XLIX

On
co
©

Jae ee

=

CY

¢
A

D)
1%

A

b
a
y

pt
y

Whos w er yp va v be
YW obs oP W2 :

Fic. 4. One of the earliest representations of an American landscape; from the
Cantino map of 1500

André Thevet, who sojourned for a short time in Brazil,
published his ‘‘Singularitez de la France Antarctique”’
in 1558. His description of the ‘‘Su,’’ in reality the
opossum, is paraphrased by Conrad Gesner, Edward
Topsell, J. E. Nieremberg and John Jonston under that
caption, and his grotesque caricature of the beast is re-
produced by these authors. It is also introduced in
sixteenth century cartography of the two Americas.
Blaeu, in his world-map of 1605, places the ‘‘Su’’ and
its descriptive legend in the region of Nova Francia;*
and in the La Plata region of the same map occurs still
another figure of the opossum, based upon the century-old
drawing which appears in the Waldseemiiller world-map.
Our Fig. 5 is taken from Thevet, and Fig. 6 from Nierem-
berg, whose ‘‘ Historia Nature’’ was published in 1635.

In Wolfe’s English edition of van Linschoten’s ‘‘ Voy-
ages,” figures of the sloth and ‘‘Simivulpa”’ are intro-

4See the new facsimile edition (1914) published = Dr. E. L. Stevenson
under the auspices of the Hispanic Society of Amer

No. 586] EARLY PORTRAYALS OF THE OPOSSUM 591

Fic. 5. The “Su” (common opossum) ; after André Thevet, 1555.

RN Wess y
SI BS N\ WSS
SS

Fig. 6. The “ Flaquatzin” (wooly opossum); from Topsell, after Nieremberg,
1635.

duced in the Brazilian and Argentine region of the map

of the South American continent, and at page 232 of this

work occurs the following description of one of these
asts:

592 THE AMERICAN NATURALIST [Vou XLIX

There is likewise another wonderful and strange beast of Gesnerus
called a Foxe ape, on the belly whereof Nature hath formed an other
belly, wherein when it goeth into any place, it hideth her young ones,
and so beareth them about with her. This beast hath a body and mem-
bers like a foxe, feete like mens hands, or like sea cattes feete, eares
like a batte. It is never seene that this beast letteth her young ones
come forth but when they sucke, or ease themselves, but are alwayes
therein, until they can gette-their own meate.

Passing now to the seventeenth century writers, we find
this account of Didelphis in Raphe Hamor’s ‘‘True Dis-
course of the Present Estate of Virginia’’ (London,
1615):

For true it is, that the Land is stored with plenty and variety of wild
beastes, Lions, Bears, Deere of all sorts. . . . Beavers, Otters, Foxes,
Racounes, almost as big as a Fox, as good meat as a lamb, Hares,
wild Cats, Muske rats, Squirrels flying, and other of three or foure
sorts, Apossumes, of the bignesse and likenesse of a Pigge, of a moneth
ould, a beast of as strange as incredible nature; she hath commonly
seauen young ones, sometimes more and sometimes lesse, which she
taketh vp into her belly, and putteth forth againe without hurt to her
selfe or them.

Of each of these beasts, the Lion excepted, my selfe have many
times eaten, and can testifie that they are not only tastefull, but also
wholesome and nourishing foode.

oF >?

Fic. 7. The opossum and young; after César de Rochefort, 1658.

About the same time Captain John Smith wrote the fol-
lowing brief characterization of the opossum, in his ‘‘De-
scription of Virginia” (1612):

An opossum hath a head like a Swine, and a taile like a Rat, and is
of the bignesse of a Cat. Under her belly she hath a bag, wherein she
lodgeth, carrieth, and suckleth her young.

After Nieremberg, a Jesuit professor at Madrid, whose
work on natural history (1635) is chiefly a compilation,
we come to George Marcgrav and Wilhelm Piso; and

No. 586] EARLY PORTRAYALS OF THE OPOSSUM 593

their contributions on Brazilian natural history, pub-
lished in 1648, are recognized as highly meritorious.

Ulysses Aldrovandi’s large posthumous folio on Quad-
rupeds (1637, p. 103) also contains a figure of the opos-
sum (otherwise interpreted, however) which is clearly
traceable to the early carto-
graphic designs. But it is unnec-
essary to pursue the subject fur-
ther, except to state that Fig. 7 is
copied after Charles César de
Rochefort’s engraving of an opos-
sum (‘‘ Histoire des Îles Antilles,”’
1658), and Fig. 8 shows the same
animal, acording to Eduard Sel-
er’s interpretation, as depicted in
one of the Maya Codices (Nuttall,
71). j Fic. 8. Maya representa-

Among other mammalian fig- tion of the opossum (?) From
ures in pre-Columbian Maya and Gwe
Mexican colored drawings® that
have been preserved are several that represent a spotted
dog, probably one of the varieties of ‘‘ Aleos’’ mentioned
by Hernandez. The occurrence of an indigenous spotted
dog in Central America is of interest in view of the fact
that a similar race is depicted in ancient Egyptian, As-
syrian and Pelasgian animal effigies and paintings, some
of the figures dating as far back as about 3000 s.c.

The oldest known representations of the hunting dog
of the ancient Egyptians, together with a number of
large African mammals, are inscribed in a palette dis-
covered a few years ago at Hierakonpolis.

| Fr
a UO
TE n e

5See Edward Seler, ‘‘Die Tierbilder der EE und Maya

Broin >’? Zeitschr. f. Ethnol., Jhrg. 41, 1909. A. M. Tozzer and

. Allen, ‘‘ Animal Figures in the Maya Codices,’’ papers of Peabody
Mosii iā; Ethnol., Vol. 4, No. 3, 1910. References to the literature
on ancient Egyptian and Assyrian animal effigies will be found in Amer.
Journ. Philol., Vol. XXX, 1909, pp. 322-331. The early history of the
rhinoceros is triood by B. Laufer in Publication 179 of the Field Museum,
and medieval ideas of the elephant are portrayed by E. D. Cuming in a
recent number of Field (April 3, 1915).

594 THE AMERICAN NATURALIST [Vor. XLIX

Concerning the several varieties of ancient Inca or
Ancon dog that are known from well-preserved Peruvian
mummies, Nehring® is of the opinion that their remote
ancestry is traceable to the North American wolf (Lupus
occidentalis var. mexicanus and rufus). The great an-
tiquity of domesticated dogs in South America is indi-
cated also by a canine skull which R. Lydekker has de-
scribed from the superficial deposits of Buenos Aires.
This dog, according to Dr. Lydekker,” ‘‘though appar-
ently contemporaneous with many of the wonderful ex-
tinct mammals of the Pampas, yet shows unmistakable
signs of affinity with domesticated breeds, although the
precise relationship has not been established.”’

Reference having already been made to animal figures
in early American cartography, we may call attention in
closing this sketch to a memoir by Anibal Cardoso in the
Anales of the Buenos Aires Museum for 1912 (Vol. XV),
on the origin of Argentine horses. The writer endeav-
ors to show from historical evidence that large numbers
of horses existed in the interior of the country prior to
the Spanish Conquest, and a figure of one of these ani-
mals drawn by Sebastian Cabot in his world-map of 1544
is interpreted as indicating that wild herds were seen by
that navigator in 1531. A portion of Cabot’s map is re-
produced in Sefior Cardoso’s memoir (p. 379), and also
in one by J. T. Medina on the voyage of Sebastian Cabot.

Nevertheless the conclusion appears unavoidable that,
had the horse actually persisted in the western hemisphere
down to the time of the advent of Europeans, some traces
of it would certainly appear in the culture of the primitive
inhabitants.

6 Sitzungsber. ges Naturf. Freunde, Berlin, 1884.

TR. Lydekker, ‘‘ Mostly Mammals,’’ London, 1903, p. 204.

8‘‘ Antigüedad del Caballo en el Plata.’’ On the horse in post-con-
quistorial times in North America see Clark Wissler, ‘‘ The Influence of the
Horse in the Development of Plains Culture,’’ in Amer, Anthropol., Vol.
XVI, 1914.,

SEVENTEEN YEARS SELECTION OF A CHAR-
ACTER SHOWING SEX-LINKED MENDELIAN
INHERITANCE!

RAYMOND PEARL

I

In 1898 there was begun at the Maine Agricultural
Experiment Station an experiment in breeding Barred
Plymouth Rock fowls, having for its purpose the improve-
ment by selection of the character winter egg production.
This investigation has continued to the present time. A
résumé of the results to date, considered with reference
to their bearing upon the general biological problem of
selection, may be of some interest.

The experiment has fallen into three divisions or pe-
riods: viz., (1) the period from 1898 to 1907, (2) the
period from 1908 to 1912, and finally (3) the period from
1912 to date. Detailed reports on the methods of breed-
ing in operation have been published elsewhere? For
purposes of clear orientation in the present discussion it
will be well here briefly to review the facts as to the
methods of breeding used in each of the periods. With
these facts definitely in mind we may then proceed to an
examination of the results.

1. The Period from 1898 to 1907.—During this period
the breeding followed the plan outlined at the beginning
by Woods and Gowell. Essentially it consisted of the
following elements.

1. Trap-nest record of the performance of each indi-
vidual female.

2. Selection as breeders of all females which laid more
than a definite number of eggs (150) in the first
laying year.

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

periment Station, No. 87.

2 Cf. particularly Woods, C. D., and Gowell, G. M., U. S. Dept. Agr.
Bur. Anim. Ind. Bulletin 90, 1906, pp. 42; Pearl, R., and Surface, F. M.,
Ibid. Bulletin 110, Part I, 1909, pp. 80; Pearl, Me. Agr. Expt. Stat. Ann,
Rept., 1911, pp. 113-176; and Pearl, Jour. Exp. Zool., Vol. 13, 1912, pp.
153-268.

595

596 THE AMERICAN NATURALIST (Von. XLIX

3. Selection as breeders of males whose dams had laid
more than another definite number of eggs (200).

4. The indiscriminate mass breeding, without individ-
ual pedigrees, of all individuals selected as de-
scribed under 2 and 3, and, in consequence,

5. No test of the progeny we particular matings with
respect to their laying ability.

This may be designated as the period of mass selection.

The following statement regarding the methods used
in this period was made by Woods and Gowell (loc. cit.,
p. 8):

The plans followed in this breeding work are based upon everyday,
practical common sense, and are the same as would be used in building
up a high-producing strain of dairy animals. Individual records of
performance are kept. The large producers are mated with sons of
large producers in the hope of obtaining a race of improved layers. In
the first year’s work three birds laid over 200 eggs each, and this fact
led to the adoption of that number of eggs as the minimum perform-
ance for a “registered” bird. Other than this there was no reason for
selecting 200 as the number of eggs necessary to entitle a bird to regis-
tration. Any other number, as 190 or 210, might have been taken with
equal propriety, just as horsemen might have selected some other time
than 2.30 by which to determine a standard horse.

2. The Period from 1908 to 1912.—For reasons which
have been fully set forth elsewhere? it was decided not to
continue the breeding along the same plan after 1907.

he new plant, put into operation first in the breeding
season of 1908, was calculated primarily to furnish defi-
nite information regarding the mode of inheritance of the
character winter egg production. It involved essentially
the following items:

1. Trap-nest record of the performance of each indi-
vidual female.

2. The selection of both males and females was made
on a double basis, including in addition to the
individual’s own performance as in the earlier
plan, also the idea of progeny performance. In
practice this worked out for hens in the following
way: Plans were made to see whether there could

3 Pearl and Surface, loc. cit.

No. 586] SELECTION OF A CHARACTER 597

be formed by selection and propagated three dis-
tinct strains of winter egg producers, namely, high,
mediocre and low. This involved, on the individ-
ual performance side, the separate selection in the
first years of three classes of females as breeders:
(a) good winter producers, with records before
March 1 of above 30 eggs; (b) mediocre winter
producers, with records below 30 eggs; and (c)
poor winter producers, which laid no eggs before
March 1. The division at 30 eggs was, after the
first year, merely a nominal one in the selection of
high producers. Actually only birds were used
in the a class whose records materially exceeded
30 eggs, running up to over 100 eggs in some
cases. :

The progeny performance idea was carried out in
two ways in the breeding. In the first place, no
female was selected for the high winter produc-
tion breeding pens, for example, unless, in addi-
tion to her own high winter record, all her sisters
and her dam were high producers. In the second
place, of all females fulfilling the above qualifica-
tion only those were bred a second time whose
progeny from the first year’s mating had proven
to be all high producers. Similar types of selec-
tion were followed by the mediocre and low lines,
except that segregating families were put in the
mediocre class.

3. The selection of males was along essentially the
same lines, with only such difference as is in-
volved in the fact that the male makes no per-
formance record himself. Males were put into
the breeding pen the first time on the basis of the
records of their dams, on the one hand, and of
their sisters, on the other hand. Those whose
progeny proved that they were transmitting the
character to which selection was being made were
used a second or even third time as breeders.

4. Complete individual pedigrees, whereby each off-

598 THE AMERICAN NATURALIST (Vón XLIX

spring individual’s parentage, both male and fe-
male, was known.

d. The records of production of the progeny of each

mating separàtely recorded and studied as a unit.

It will be noted that there are but two essential differ-
ences between the plan in this period and that followed
in the earlier one. These are: first and most important,
that in this second period the principle of progeny testing
was introduced into the scheme of breeding. The second
difference was that selection was carried on for low pro-
duction as well as for high, which had not been previously
done. A third difference is apparently found in the fact
that in this second period of selection the winter record
rather than the yearly record is made the basis of selec-
tion. This is in no way an essential difference. The
reasons for adopting the winter period have been set
forth in detail elsewheret and need not be repeated. It
suffices to say here that essentially the same results and
conclusions will be reached if one uses winter production
or annual production.

As a result of the studies made in this period on the
plan of breeding outlined the mode of inheritance of the
character winter production was definitely determined,
and has been confirmed by subsequent work.’ The char-
acter was shown to be Mendelian in its genetic behavior,
depending upon two factors, one of which is sex-linked.

3. The Period from 1912 to Date—The only difference
in the mode of breeding the stock of Barred Plymouth
Rocks in this period, as compared with the preceding one,
is found in the fact that during this last period all selec-
tions for low and mediocre production have been dropped.
The breeding for high production alone continues, with
only such differences in the details of manipulation of
the breeding stock as would naturally follow a definite
knowledge of the mode of inheritance of the character,

t Pearl, 1912, loc. cit., and Me. Agr. Exp. Stat. Ann. Rept., 1914, pp. 217-
ay Cf. also Wilson and Murphy, Jour. Dept. Agr. Ireland, Vol. XIV,

5 Pearl, 1912, loc. cit., also Amer. Nat., Vol. XLIX, 1915, pp. 306-317,
and Curtis and Pearl, Jour. Exp. Zoology, Vol. 19, 1915, pp. 45-59.

No. 586] SELECTION OF A CHARACTER 599

and of the gametic constitution of particular individuals
with reference to that character. As a matter of fact, all
low-producing lines were dropped at the end of the laying
year 1911-12. Certain of the mediocre lines were contin-
ued a year longer. In the laying flock of 1913-14 there
were no birds which had been bred for anything other
than high production, so far as the breeder’s deliberate
intention went.
II

The results of this seventeen year selection period are

set forth in Table I.
TABLE I

MEAN WINTER PRODUCTION PER BIRD OF THE BARRED PLYMOUTH Rock
FLOCKS FROM 1899 To 1915

Mean Winter Mean Winter
Mean Winter No. of Birds Production of Production of
Laying Year Production of |Making Winter| All Birds Se- All Birds Se-
All Birds Records lected for High | lected for cow
oduction Production
1899-1900 41.08 eggs 70 — ae
1900-1901 37.88 85 — —
1901-1902 45.293 “ 48 — —
1902-1903 26.01 = 147 — —
1903-1904 OO hb: ** 254 — —
1904-1905 35.04 “ 515 — —
1905-1906 40.65 ‘ 635 — —
1906-1907 yea * 653 — =
1907-1908 19.93 * 780 — —
1908-1909 26.69. “ 359 54.16 22.06
190 10 OLO rhe 247 47.57 25.05
1910-1911 30.49 ‘* 64 50.58 17.00
1911-1912 35.98: . ** 232 57.42 16.48
1912-1913 43.01 ** 182 52.61 —
1913-1914 52,2% *“ 192 52.20 —
1914-1915 45.89 ‘°° 179 45.89 '—
ToS rangs ;
eak 35.05 “ 4,842 51.49 20.14

The data of this table are shown graphically in Fig. 1.

From the table and diagrams the following results
appear:

1. The number of individuals involvedin this experiment,
on each one of which exact trap-nest records have been
kept, is large, amounting nearly to five thousand. This
number is large enough to lead to conclusions which are
trustworthy. It will be shown presently that wherever
it has been possible to compare the results on egg produc-

or)
Q
©

THE AMERICAN NATURALIST [Vou. XLIX

~<
™
> MEAN WINTER PRODUCTION
` N w A Q ®
S S S S S Ò
8 2
S /
X ers
g / k
S
Ş
& i —— — — a a cS
` an ARC A

ae “o
: ~X E
3 ` a
« y 8 Ha

a \
Š < :
i N

~ H
rnd á
X A 7
al re
Ñ /
ii °
& |
7
>
EN P
x| !
a

Fic. 1. — showing the course of mean winter egg Pearcy ate the
years 1899 and 1915. The solid lines and circles give the total flock m The
two straight aie. fitted by the method of least squares to the seas flock

ans, have ations = 5 — 2.1812, 17, 4 e
0 circles and broken (dash) line give the means of the lines selected for high
winter production betwe h 19 5. dotted line and

open cireles give the mean winter production of the lines selected for low produc-
tion between the years 1908 and 1912,

No. 586] SELECTION OF A CHARACTER 601

tion obtained in this experiment with independently de-
termined norms the general trustworthiness and normal-
ity of the present data have been demonstrated.

2. From the beginning of the experiment through the
laying year 1907-08 the general trend of mean produc-
tion was downward, with minor fluctuations from year to
year. In other words during the period in which the sys-
tem of breeding was mass selection for high production
without progeny test there was no change of the mean in
the direction of the selection. The fluctuations in mean
production during this period were, in the main, due prob-
ably to two sets of causes: (a) environmental differences
in different years acting at one point or another in the
life history of the birds; (b) random fluctuations in the
genetic constitution of the male birds used as breeders in
successive years, brought about because of the ignorance
of the breeder, in the absence of any individual progeny
testing plan, of the ability of any particular male to trans-
mit high fecundity to his daughters. The first of these
factors needs no special discussion and is relatively of
minor importance. The second will be dealt with in detail
farther on in the paper.

3. Since the laying year 1907—08 there has been a steady -
increase in mean winter production for the whole flock, ex-
cept for the years 1910-11 and 1914-15. Inthe former year
the decline in the mean is slight, and is probably due to
unfavorable environmental influences. In 1914-15 the
decline is certainly due to such causes. In the hatching
season of 1914 an inexperienced man was in charge of
the incubation and rearing. He had very poor success,
and the Barred Plymouth Rock pullets available for the
laying houses in the fall were relatively few in number,
of a relatively late average date of hatching, and poorly
grown. It is remarkable, not that the mean winter pro-
duction was lower in 1914-15 than in 1913-14, but rather
that it was so high as it was, taking all the circumstances
into account. It appears that the system of breeding
which made the selections on a progeny test basis was
immediately and, to date, continuously effective.

602 THE AMERICAN NATURALIST [Vou. XLIX

4. That selection on a progeny test basis was effective
is demonstrated not only by the general flock averages,
but also by the fact that it was possible to propagate sep-
arately high and low producing strains. The high pro-
ducing strains differed widely from the low producing in
mean winter production. Taking the average for seven
years in the case of the high, and four years in the case of
the low, it appears that the mean winter production of the
high producing strains was approximately two and a half
times that of the low producing strains. At the end of
the laying year 1911-12 the low producing lines were
dropped. In the next year (breeding season of 1913) no
birds were bred which were known to belong to segregat-
ing lines. Of course some were included which proved
afterwards to have been segregating, but this fact could
not, in any such case, have been told in advance from the
records in hand. The propagation of low producing
strains was attended with a great deal of practical diffi-
culty with the environmental conditions under which one
has to operate at this station. The growing season is
short. In order to grow properly a pullet for laying pur-
poses it is necessary that she be hatched after April 1 and
- before June 1 at the latest, and preferably before May 15.
If, however, one selects birds which produce no eggs what-
ever before March 1, and use up some valuable time be-
fore they get well started in the spring cycle of laying it
becomes perfectly clear that one is automatically pre-
vented from getting any considerable number of early
hatched chicks from such mothers. If late hatched chicks
are used the results obtained as to winter production
later will not be critical. These cireumstances make the
propagation of a low producing strain on a large scale
really a difficult proposition. There is of course no diffi-
culty in breeding birds which will not be winter layers.
One only needs to hatch in June, July or August. But
such birds will furnish no critical evidence regarding the
inheritance of winter production.

5. The mean winter production for whole flocks over
the entire period of the investigation is 35.05 eggs. In

No. 586] SELECTION OF A CHARACTER 603

the writer’s opinion, based upon rather extended expe-
rience in the study of egg records, approximately this fig-
ure may be taken as representing the general average
winter production of mixed flocks of Barred Plymouth
Rocks (or of American birds generally with the probable
exception of White Wyandottes), which have been hatched
at the proper time and well reared. As evidence on this
point the data presented in Table II have pertinent bear-
ing. These data give the mean winter production of birds
of the different American breeds obtained in the Fourth
Philadelphia North American International Egg-Laying
Competition, carried on at the Delaware Agricultural
Experiment Station in 1914-15. The conditions under
which these records are made are such as to safeguard
their essential accuracy. The figures here given are the
mean productions per bird up to the end of the eighteenth
and seventeenth week of the laying after November 1,
1914. Owing to the fact that the original records as pub-
lished are compiled by calendar weeks it its not possible
to give the exact production from November 1 to March
1. Eighteen weeks gives 5 days laying over this period,
and seventeen weeks gives 2 days under. Both sets of
means have therefore been tabled. It should be said that

the birds were kept in flocks of 5 birds each, thus tend-
ing to the most favorable condition for high individual
records.

TABLE II
MEAN WINTER PRODUCTION OF FOWLS OF THE AMERICAN BREEDS, CALCU-
LATED FROM RECORDS OF THE FOURTH INTERNATIONAL
AYING COMPETITION, 1914-15

Breed No. of Birds meee geadas Nov.1 Moa. a gg 1
Barred Plymouth Rocks caste 45 29.20 25.47
All Plymout y # rare o CONE N 80 32.39 28.39
All fja D TE E saaes! 55 48.20 44.00
IR 75 38.37 34.42
All peck ae Droóds.isscsrsse 210 38.67 34.63

From this table it is clear that the records presented in
Table I average about the same as those of the 210 birds
of the American breeds in the Delaware competition.

604 THE AMERICAN NATURALIST [Vow. XLIX

The Wyandottes alone give a distinctly higher mean,
and this the writer has also found to be true of Irish egg
records. The Barred Rocks in the competition this year
give a somewhat abnormally low winter mean. Irish
records® give a winter mean for 48 Barred Rocks, in flocks
of 6 each, of 36.54 eggs per bird which is more nearly
normal.
iit

Let us now turn to the question of the interpretation of
the data set forth in the preceding section. Broadly
speaking what the facts gleaned from this seventeen-year
experiment show is that mass selection for egg produc-
tion was not effective, while selection which was based
upon the performance of the progeny was extremely and
quickly effective. What is the explanation of the differ-
ence? The facts are purely objective realities, about
which dispute or question is idle. They are real and
obvious matters. Regarding the interpretation of such
facts as these there have been, and no doubt will continue
to be for some time to come, differences of opinion
amongst biologists. Under these circumstances, the
writer will, then, with all respect and consideration for
the differing opinions of others endeavor to make clear
the view of the meaning of these results which he has
come to hold.

In the first place we may definitely put aside any inter-
pretation which bases its explanation of the results on
environmental action. In an earlier paper the writer‘
has shown in detail the impossibility of this explanation
for the results during the period of mass selection (the
descending limb of the curve). The totality of the results
here presented make it still more apparent that such an
explanation can have no place here. We should have to
suppose that the environmental influences were adverse
throughout the period of mass selection, but suddenly
became favorable when the method of breeding was
changed, and have ameliorated at an ever increasing rate

6 Murphy, L., Jour. Dept. Agr. Ireland, Vol. XIV, pp. 8-30, 1913.
7 Pearl, R., Me. Agr. Expt. Stat. Ann. Rept., 1911, pp. 113-176.

No. 586] SELECTION OF A CHARACTER 605

as time has gone on. Nothing of the sort was, in fact,
the case. The only explanation which can satisfy the
case is one which is based upon or at least takes full
account of the changing genetic constitution of the flock.

It appears to the writer that the essential key-note to
the explanation of the results of this long experiment
is found in the fact that phenotypic variation of the char-
acter fecundity, in fowls, markedly transcends, in extent
and degree, genotypic variation. It is quite impossible
in the great majority of cases to determine with preci-
sion what is a hen’s genetic constitution with respect to
fecundity from an examination of her egg record alone.
In this case, as in so many others, but in an unusually
pronounced degree, where the phenotypic distributions
overlap, a sure diagnosis of genetic constitution can only
be made by means of the progeny test. Lacking this the
phenotypic performance becomes an always uncertain
and at times very misleading guide.

It can be shown that if, during the period of mass se-
lection, all the hens used as breeders had been, as they
were supposed in the theory of the originators of that
part of the experiment to be, either Type 1 or Type 2
females (fL L,- Fl, or fL,L,- Fi,L,) then the continued
mass selection must have resulted in improvement. The
only criterion of constitution which was used, however,
was the bird’s performance. But, taken by this criterion
alone, there would be constantly chosen a proportion of
birds whose genotype was for mediocre fecundity, but
which made a performance record (phenotypic) suffi-
ciently high to be selected. That this is what actually
happened is evident from the curve (Fig. 1), but the fact
was experimentally proved in 1909.8

Put in the fewest words, then, the reason why no effect
was produced during the first ten years of selection and
a marked effect was produced during the last seven, was,
in the writer’s opinion, because genotypically high pro-
ducers were uniformly selected (in the high lines) during

8 Pearl and Surface, Me. Agr. Expt. Stat. Ann. Rept., 1909, pp. 49-84,

606 THE AMERICAN NATURALIST [VoL. XLIX

the latter period, and were not uniformly selected in the
former. By the introduction of the progeny test as an
essential part of the selection the whole process of the
creation of a highly fecund race of hens was transferred
from the realm of blind chance to that of precise and defi-
nite control. And it becomes increasingly clear that
mass selection, on the basis of performance (phenotypic
appearance) alone, is in its essential nature a blind and
haphazard process, no matter with what precision and
stringency it is carried out, just so long as the correla-
tion between the gametic and somatic conditions of the
character selected is not perfect. And it is an outstand-
ing result of the Mendelian investigations of the last 15
years that the gamete-soma correlation is very rarely, if
ever, perfect.®

It appears on this view that selection for high egg pro-
duction in the fowl is effective when it is real. That is,
if one selects genetically high producers by means of the
trap-nest plus the progeny test, he succeeds very rapidly
in fixing a high producing strain. If on the other hand
he merely selects high layers by the trap-nest record
alone, he is not really selecting genetically high pro-
ducers except in a portion of the cases. Under these cir-
cumstances he makes no progress in building up a highly
fecund strain. To be effective in changing the average
productiveness of a flock of poultry selection must pick
out those birds as breeders which carry the factors for
high fecundity genetically, i. e., as an integral part of
their hereditary make-up, and not any other birds.

With the above interpretation of the results of seven-
teen years’ continuous selection of the character fecundity,
all the facts known to the writer are in complete accord.
No other interpretation of the results of this experiment
has yet been suggested which will meet all the facts.

® Complete citations on this point would make a tolerably full bibliog-
raphy of Mendelism. The methodological or the strictly quantitative aspects
of the problem have been but little dealt with. In this connection ef. W

F. R. Weldon, Biometrika, Vol. I, pp. 228-254, 1902, and R. Pearl, Biol.
Bulletin, Vol. XXI, pp. 339-366, 1911.

No. 586] SELECTION OF A CHARACTER 607

IV

What bearing have these results upon the general prob-
lem of the effectiveness of selection in modifying ger-
minal determiners? Let us at the outstart endeavor to
be quite clear as to the problem. It has been shown in
what has preceded that in the long experiment a change
in the average condition of the population has occurred,
and has been coincident with selection. The writer has
no hesitation in saying that the increase in average pro-
ductiveness since 1908 has been caused by the particular
kind of selective breeding practised. Furthermore the
average productiveness at the present time transcends
any average known in the previous history of the stock.

Granting all this, however, as plain matter of fact, it
does not, in the writer’s opinion, afford one iota of evi-
dence that through the process of selection the hereditary
determiners of fecundity either have been or can be
changed. All that the selection has done, so far as we
have any evidence, is to change the constitution of the
population in respect of fecundity genotypes. There is
no evidence that the genotypes themselves have been
changed. On the contrary, everything indicates that they
have not been changed. During the last seventeen years
we have merely sorted out from a mixed population, by a
systematic method of breeding, those individuals which
were alike in one respect, and have sold all the rest to the
butcher. That one respect was that each individual bore
progeny which were high producers.’® Those individuals
chosen as breeders in 1908 and 1914 were precisely alike
in this particular There were more of them available
in proportion to the whole flock in 1914 than in 1908, but,
as individuals, I am unable to discern any particular in
which they were different in 1914 from what they were in
1908.

The general point here involved is essentially the same
as one with which we are more familiar in demographic

10T am, of course, referring to the high line selections only, merely for

the sake of verbal economy. The same reasoning mutatis mutandis applies
to the low lines.

608 THE AMERICAN NATURALIST [Vor. XLIX

statistics. The constitution of a population does not di-
rectly affect the individual. My expectation of life will
not be materially increased if I chance to move into a
community in which all the other inhabitants are of ad-
vanced age. To this everyone will agree. But the ex-
treme selectionist appears to believe that in some mys-
terious way the act of continued selection, which means
concretely only the transference of each selected individ-
ual from one cage or pen to another to breed, will in and
of itself change the germ-plasm, so that after the act it is
different from what it was before! It does not seem that
the evidence that such is in fact the case is critically valid.
A careful study of the very interesting and valuable work
of Castle and Phillipsi! with rats leaves the writer with
the feeling that those experiments prove no more than do
the experiments here reported: namely that the compo-
sition of a population may be altered by selection. It
does not appear to be proven that selection has essentially
altered the constitution of the germ-plasm of any partic-
ular individual as compared with the germ-plasm of that
individual’s ancestors, making due allowance of course
for the phenomenon of segregation. That selection can
alter the composition of a population with respect to
genetic determiners, by a process of sorting over what is
already there and rejecting some portion, no one can
doubt. But it still appears to the writer to be true that:
‘< It has never yet been demonstrated, so far as I know,
that the absolute somatic value of a particular hereditary
factor or determinant (i. e., its power to cause a quanti-
tatively definite degree of somatic development of a char-
acter) can be changed by selection on a somatic basis,
however long continued.’’!?

11 Castle, W. E., and Phillips, J. C., ‘‘Piebald Rats and Selection,’’

Carnegie Institution of Washington Publication No. 195, 1914.
12 R. Pearl, Jour. Exp. Zool., Vol. 13, p. 264, 1912.

SPECIFIC AND VARIETAL CHARACTERS IN AN-
NUAL SUNFLOWERS

PROFESSOR T. D. A. COCKERELL

UNIVERSITY OF COLORADO

THe group of Helianthus annuus, the typical, annual
sunflowers of North America, is not a large one. The
annual habit seems to have been acquired independently
by several different Helianthine stocks, so that H. bolan-
dert Gray, H. exilis Gray, H. floridanus Gray and H.
tephrodes Gray are to be excluded from the H. annuus
group. The subgenus Helianthus s.str., or Euhelianthus,
contains the following:

1. H. annuus Linn. Based on the large cultivated form
(H. macrocarpus D. C.), Dr. A. H. Church of Oxford has
investigated the history of this plant, and I take the
liberty of quoting from a letter he wrote on March 4,
1915:

The published accounts of the giant sunflower in Europe in the six-
teenth century are so precise that it is interesting to remark that this is
in fact the oldest mutation known, which is still with us, quite unaffected,
though still never quite a pure strain, owing to insect pollination. The
facts are quite simple. The first description of the plant, by Dodonzus
(1567), tells us it grew in the Botanic Garden at Madrid, 24 feet. At
the Padua Garden, indoors, in a viridarium or orangery, 40 feet! The
usual height was 20 ft. The first English specimens, grown in London
by Gerard, were 14 ft.; and 15 ft. is the local record here. The giant
form is known by earrying one head, and having no trace of axillary
buds, Liar cadushalle strain, as opposed to reverting branching indi-
viduals. . The next point is, where did it come from? From Peru,
say the lls: but all Spanish things from America came via Peru,
because this was the last port of call. Hence Mexico is regarded as the
home. On the other hand Ximenes, who lived in Mexico several years,
and Hernandez after him, call it the Chimalacak del Peru; “acak” I
find means a reed, and thus refers to the long tall single stem of the cul-
tivated crop. The inference is that the plant as we know it was evolved
by ages of selection in Peru, by guano fed cultivation, possibly long be-
fore Inca rule, the plant having been taken by all migrating tribes from
the Mexican district. . Regarded as a product of Peruvian agricul-
ture the sunflower is aas parallel with the maize. . . . It was the
oil crop of ancient America.

609

610 THE AMERICAN NATURALIST [Vou. XLIX

The true H. annuus appears to be quite unknown in
the wild state, but nevertheless the monocephalic char-
acter may have arisen among wild plants. Dr. Church
makes the following suggestion:

If the monocephalic form is the giant of cultivation derived from the
Prairie form, it should be possible to repeat the history, by growing
Prairie forms in quantity, and selecting the suitable mutations when they
appear under the stimulus of excess manure (guano for choice). My
idea has been that, knowing what to look for, it might be possible to get
somewhere near it in say 10 years; though the Indians possibly took
2,000. General structural evidence alone suffices to show that the mono-
cephalic strain is the response to selection for close cultivation (about
two plants per square yard). The solitary heads are required for simul-
taneous harvesting.

2. H. lenticularis Douglas. The prairie sunflower,
much branched, and normally with dark disce. It has
been regarded as the wild type of H. annuus, but Ryd-
berg treats it as a distinct species. In crosses with
typical annuus, the F, is intermediate, often with a
tendency to fasciation. If annuus and lenticularis are
considered specifically distinct, we have to face the diff-
culty that the former is known only in cultivation, and its
one ‘‘ specific’’ character, the monocephalous habit, is
not constant.! The color of the dise is not a reliable
distinction, since yellow disces occur in wild plants.
Possibly the variation shown by H. annuus may be ex-
plained by contamination with lenticularis, since some
strains, at least, are constant in their characters. At
present, however, it seems probable that no wild species
ever existed with typical H. annuus characters; the actual
facts would probably be best represented by considering
lenticularis the species, and annuus a cultivated variety
derived therefrom. Since, however, the latter was first
named, the species-aggregate will have to be called H.
annuus, and the nomenclatural outcome will be as fol-
lows:

1 Shull, Botanical — 45, 105, figures a much branched form which is
not the wild lenticularis

No. 586] ANNUAL SUNFLOWERS 611

Helianthus annuus L.
(a) lenticularis (Dougl.)
v (b) macrocarpus (D.C.) = annuus L., s. str.
At the same time, for ordinary purposes, it may be per-
missible to simply write H. lenticularis when referring to
the wild plant.

3. H. aridus Rydberg. Like H. lenticularis, but
leaves lanceolate or narrowly deltoid, minutely toothed
or entire. Montana to New Mexico. Nelson calls this
a synonym of H. petiolaris, which it certainly is not. It
must be called H. lenticularis aridus (Rydb.) or H.
annuus lenticularis var. aridus, since it is a variable
form of lenticularis, which may possibly be due to cross-
ing with H. petiolaris, the hybrid having crossed back
with lenticularis. From the mode of its occurrence it is
nearly certain that it is not a simple lenticularis X peti-
olaris hybrid, petiolaris being often absent from the
immediate vicinity.

To give an idea of the actual condition of affairs
where H. aridus occurs in Colorado, I present a synopsis
of the forms found at Longmont, August 30, 1914:

(4) H. aridus type; smaller and more ra hte cuneate bases to leaves.
(a) seas with yellow disc; two plan

Dis m. diameter, light a aia entirely dull light

“alow ; rays ordinary; foliage unusually pale; base of

rather broad-cuneate, marginal teeth feeble.

(b) Dise ae (col lobes dark reddish). Leaves with cuneate base

subentire margins; typical aridus. Involucral bracts

sti broad and bristly. ay rather slender plants

have small discs (17-21 mm. diam.) and very ample

rays, which are not very numerous (10-13); color of

rays rich orange yellow; stems lightly speckled with

purplish.
(i) Rays longer, about 38 mm. long and 15 broad.
(ii) Rays shorter, about 28 mm. long and 14 broad.
(This difference in size of rays is probably environmental.)
(B) H. neal type; bases of leaves truncate or cordate; plants usually
robust; dise dark.

(a) ii of aridus, being rather slender, with small (diam. 23.5
mm.) dise and long rays; but leaves broadly truncate at
base and rather strongly toothed, quite lenticularis style.
This is a very pretty form, with long rays (about 40
mm, long and 11 broad), more or less twisted at end, and
rather narrow. The rays number about 15.

612 THE AMERICAN NATURALIST [Vou. XLIX
(b) Aspect a o ; more or less robust, rays rather short and

(i) “ng comparatively short and broad (about 22 mm. long
and 1 oad on a small head), the middle third
= with its apical half variably light brown-
ish-red. Leaves thick, with broad petioles.
(ii) Rays normal.

(a) Upper leaves ovate, scarcely at all dentate, inequi-
lateral. Dise small (22 mm. diam.), rather paler
than usual, the corolla lobes showing less red.
Rays 14, about 31 mm. long and 11.5 broad.

(B) he sis ae broad at base, but somewhat cuneate,

ly dentate. Disc 20-23 mm. diameter.
i ra had (dise 20 mm.) with many (21) quite
short rays, about 17 mm. long and 7 broad.

(y) Typical lenticularis, with broad-based strongly den-
tate leaves. Disc 37 mm. diameter; rays 37 mm.
long, numerous (about 33).

It would of course be possible to maintain that H.
aridus was originally a distinct or isolated species, which
has now lost its purity by crossing with lenticularis. We
ean at least say this, that if annuus, lenticularis and
aridus, in their pure forms, inhabited three different
aiaga, few would hesitate to regard them as perfectly

‘good species.” Also, if they grow mixed for any
length of time, they are sure to suffer from ‘‘ vicinism ’’
to such an extent as to lose their supposed original dis-
tinctness. At present, however, we have no assurance
that H. aridus has ever constituted a distinct species, in
the sense of occupying any considerable area in its pure
form. On the other hand, it is manifestly not a ‘‘ fluctuat-
ing variation,’’ due to mere environmental conditions.

4. H. petiolaris Nuttall. Described by Nuttall in
1821, from ‘‘the sandy shores of the Arkansa,’’ and
recommended as ‘‘an ornamental annual of easy cul-
ture.” It extends from British America to the State of
Chihuahua. It differs from H. lenticularis by (a)
smaller stature, (b) leaves differently shaped, lanceolate
or broad-lanceolate, not dentate, more or less shiny above,
those of lenticularis being quite dull, (c) bracts of invo-
lucre lanceolate, with margin very short-ciliate. Stem
rough, with a little purplish color; basal third of rays
deeper orange than the rest.

No. 586] ANNUAL SUNFLOWERS 613

This is a good species in the ordinary sense; in Colo-
rado it is often found abundantly in the cafions of the
foothills, growing without admixture of other species.
Lower down, it frequently oceurs with lenticularis.

The variety patens (Lehm.) Rydb. is said to differ by
having the heads larger, long-peduncled, the peduncles
fleshy toward the top; leaves large, long-petioled. Nut-
tall described his original petiolaris as having the pedun-
cles ‘‘of great length,’’ and the petioles ‘‘of an ex-
traordinary length,’’? though the leaves were ‘‘rather
small.’? Probably patens is not far from the original
petiolaris. Gray considered patens a synonym. Ac-
cording to Rydberg, the leaves of patens are broadly
ovate or subcordate, much in the style of lenticularis,
while the bracts are those of petiolaris, thus reversing
the condition of aridus. It is possible that aridus and
patens are both remote results of the lenticularis X petio-
laris cross, but in the vicinity of Boulder, when aridus is
common, I have not found patens.

5. H. canus (Britton) Wooton and Standley. A
species of New Mexico, Chihuahua, and adjacent regions,
close to petiolaris, but with abundant white pubescence
on leaves and stems. The involucral bracts are of the
petiolaris type. This is to petiolaris much as H.
argophyllus is to lenticularis, but the pubescence is long
and spreading, not subappressed and silky.

6. H. argophyllus Torrey and Gray. Discovered by
Drummond in dry soil in Texas. This has the form and
leaves of lenticularis, but is very remarkable for the long
subappressed silky white hairs, totally different from
those of any other Helianthus known to me. Gray re-
marks that it ‘‘ degenerates in cultivation apparently into
H. annuus,” which merely means that it suffers from
vicinism. Old cultivated stocks, kept pure, are quite con-
stant. A remarkable feature of H. argophyllus is the
extremely slow growth, at least until near flowering time.
This peculiarity is dominant in a cross with H. annuus X
lenticularis.

614 THE AMERICAN NATURALIST [Vou XLIX

7. H. debilis Nuttall. Florida to Texas.

8. H. praecox Engelm. and Gray. Florida to Texas,
near coast. Differs from debilis by being strongly
hirsute.

9. H. cucumerifolius Torrey andGray. Texas. Differs
from debilis and precox by having the branches mottled
with purple.

The last three were eventually reduced by Gray to a
single species, but Small keeps them separate.- My wife
and I have grown H. cucwmerifolius for several years,
and have crossed it with annuus X lenticularis. The
first cross is quite fertile, but it is impossible to get any
quantity of F, seed. Mr. Leonard Sutton in England
has had the same experience; he writes (April 3, 1915)

We are arranging for a large breadth of the cucumerifolius crosses

this season, but we have found as you mention that very little seed is
produced, and we are hoping that the plant will improve in this respect
if grown for a few years, and the best seeding plants are selected for
stock.
These hybrids are of considerable horticultural value,
especially those derived from crosses with the red sun-
flower, so it is desirable to secure fertile strains if pos-
sible. Something may be attained by crossing back with
the parent species.

The H. cucumerifolius type is dwarf, freely branching,
with broad bright green leaves, shiny on both sides. The
involucral bracts are long and narrow. The bulb or
swelling of the dise corollas is minutely puberulent,
whereas that of the lenticularis forms is long hairy. In
the F, hybrid the bulb is long-hairy as in lenticularis, the
character being dominant. Although H. cucumerifolius
is very unlike the other species (except debilis and
precox) in apperance, its constant structural differ-
ences are very few. The base of the leaves, as in the
annuus forms, may be auriculate or truncate. The dise
bracts may be long-ciliate, or with the margins merely
appearing scurfy. It is proper to state that my material
belonged to cultivated strains; possibly the wild plant is
less variable.

No. 586] ANNUAL SUNFLOWERS 615

Thus we have at the most nine species, which can prob-
ably be reduced to five. They belong to the region which
used to be marked in old geography books as the ‘‘ Great
American Desert,’’ though members of the debilis group
extend along the Gulf States to Florida. The dominant,
widely distributed form is lenticularis, a plant of sandy
river-bottoms and similar places, which has spread as a
weed in cultivated areas. Prior to the era of cultivation.
it is probable that H. petiolaris occupied a greater area,
at least in acreage. At the present time H. lenticularis
is common in California, but I suspect that it has been
introduced into the Pacific coast region by man.

In order to give an idea of the cultivated forms of our
group, I have made a table from Sutton’s Catalogue for
1915.

Silver-leaved (argophyllus- dai ; rays yellow;

iso blaok: G Th cornea Silver- FE

Not DAVOS ORE a air Vona E Oa a a Mie er araa ‘

1. Cucumerifolius-type, none over 4 ft. high.........-.-+-eeeee reer ees i
AMNU YDE Mostly tall osese resas es Conners tai iN ay 6.

my: Only 12 mehes hiph: (Onpa ioe erre Dwarf Miniature.
Four Coot kgh -oeer rae Rine e ra nh LES a 3.

S: munya tolled, ‘Tike the eactie dahla s: isc. sce suas i eas n esa Orion.
FORTE MOG POU oo a Fo es oa a a Eb Pe he ek hoc pas Ceeepes

4, Bays pale primrose, dise dirk ....0.00 6s eecec cies ss Primrose Stella.

Kays pripit FOHOW oso in ees ook bi ei sce laws eck e ay peru eee 5.

5: Heads email, with dark Qine sissu- oiera senstepevcsssss Miniature.

Heads larger, rays long ...........-+. acne T E S ees Stella.

6. Bays Wholly OF partly chootant red i... oioi ioe ae et Red.

Rays Toa OF DATUY VINORS ioei fh aes ees oe 28 Langley Gem

Haye primos e as sos as ho 0s 8 oe ks ee sk Na, T

Rays kiai pi u a A E a 9.

7. Double (i. e., dise florets Aglio) E E A eee Double Primrose.
Single; iad Pale E es We es Os Se ee

8. Tall, 6 POG esi es Vika kes es A R Primrose Datodinn

STORIE, Sarre Re cio cv ong iek he cee Single Dwarf Primrose.

9. Double (i. e., disc florets ligulate). ...........eesesroseseeserese. 10.

Singlo (i 0., normal heads). 06.5 ees eee eevee eria 11.

10. Flowers orango; 6 ft: ioii ouir nce os Double (also a pa 5 ft. high).

Compact habit: a ft. 6 cone iia sche 05 <i tied on wh 8s arf Double.

11. Heads extremely large; height 8-10 ft. ............0cseeeeeee Giant.
Heads ordinary or smallish; dise dark; height 6 ft. ............... 12.

12. Heads medium size ..............+ Aisthetic al (J. Veitch & Sons).

PURVES Owen sis oi ds rossii aes ssa ess Earliest of All.

616 THE AMERICAN NATURALIST [Vou. XLIX

Of the above Red is coronatus, and Langley Gem is
vinosus, both derived from our Boulder cultures. The
seed offered as Langley Gem was grown in Boulder. The
Primrose variety I have called primulinus.

An old cultivated variety is H. annuus var. indicus
(Helianthus indicus Linn., Mant. I, 117), peculiar for the
foliar expansion of the involucral bracts. It did not come
from India, but from Egypt. Tithonia speciosa, once
regarded by Hooker as a Helianthus, has the bracts nor-
mally foliaceous. In 1913 I witnessed the appearance of
foliaceous bracts in the F, generation from primulinus X
coronatus. The plant in question was a very abnormal
dwarf, wholly unlike the rest of its generation, or any
known parents. It was described as follows:

Dwarf, about 28 mm. high; slender, fasciated at top of
stem; rays vinous, but on nearly all the heads a very
dilute and dingy color; dise dark, stigmatic branches dark
red; apical part of dise corollas dark greenish, tipped
with red, and very hairy; anthers not projecting, but not
shrivelled, almost wholly without pollen, and what there
is probably no good; achenes hairy, usually with super-
numerary pappus scales; pappus scales stained with pink;
involucral bracts long and tapering, strongly hirsute,
curled over, one or two outer ones long and foliaceous;
stem hirsute; leaves long and narrow, narrowly cuneate
at base; margins irregularly, sharply dentate, entire on
small very narrow leaves; sometimes one of the large
lateral veins of the leaves, and its supporting tissue,
absent.

Such a plant may result from some unwonted combina-
tion of genes, whereby the normal constitution is broken
down and in the resulting disruption characters usually
suppressed appear. Such monstrosities quickly perish,
but during their transient existence may reveal, like a
drunken man, matters which in the well behaved would
never reach the surface.

One of the most remarkable of cultivated varieties is
the Chrysanthemum-flowered, of which we obtained a

No. 586] ANNUAL SUNFLOWERS 617

perfectly constant and uniform strain from Dreer. It
may be named var. chrysanthemoides; plants of the same
general type have passed in horticulture as var. califor-
nicus (not H. californicus D.C.)

Helianthus annuus var. chrysanthemoides

Manner of growth.—(Tested in two seasons). Grows
much more slowly than the other forms (except argophyl-
lus), but is very robust. Nine plants studied were 15-17
inches high July 14, about 36 inches July 30, and coming
into flower at about 5 ft., 6 inches, August 15.

Foliage.—At first (June 8) leaves are narrow and long;
very uniform. Later, the upper (small) leaves are con-
spicuously pallid. At time of flowering the leaves are
broad, cordate, with auriculate base; surface very
strongly crinkled; margin moderately dentate.

Pubescence.—Leaves soft with very scanty pubescence;
petioles somewhat scabrous; stems, especially toward
the top and under the heads, with abundant and con-
spicuous soft white pubescence.

Heads.—Stalks greatly broadened under heads, diam-
eter about 27 mm. just under bracts; involucral bracts
hairy, the marginal hairs not longer than those covering
the backs of the bracts, five strong veins, and others
weak; basal half of bracts about 15mm. broad, gradually,
not abruptly, tapering to acuminate ends; bracts extend-
ing about 18 mm. beyond outer florets, which are like the
inner ones; heads entirely double (i. e., corollas ligulate),
rays very bright orange or saffron, dises light green
before they come into flower; immature achenes with

‘much silvery hair. -

This plant is so distinct, structurally and physiologic-
ally, that if it were not known to have originated in cultiva-
tion, it might well pass as a distinct species. Although I
have no information concerning its history, I can only
suppose that it is part argophyllus. Mr. Leonard Sutton
writes me that the similar Double catalogued by him,
which is of continental origin, does not grow more slowly

618 THE AMERICAN NATURALIST [Vou. XLIX

Fic, 1, above, Double

Red Sunflower.

Fig. 2, below, Helianthus von tortuosus.

No. 586] ANNUAL SUNFLOWERS 619

than other sunflowers in its early stages. The var.
chrysanthemoides was found by us to cross freely with
the annuus X lenticularis varieties, producing a series of
semi-doubles. The double and semi-double forms ex-
tracted from this cross and from crosses with Sutton’s
various double forms need not be described here, but in
order to illustrate the double type, I give a figure of a
full double with chestnut color, the red being derived
from the variety coronatus (Fig. 1).

MODIFICATIONS OF THE Rays

Number.—Halsted (Rept. Bot. Dept. N. J. Agric. Exp.
Sta. for 1911, pp. 335-337) has given elaborate data from
the branched form of the cultivated H. annuus, showing
that the terminal heads have most rays, and when there
are many lateral branches, the rays on these are com-
paratively few. Nevertheless, there are inherited differ-
ences in the number of rays, not depending on conditions
of nutrition. I observed a striking case by the roadside
in Boulder, where three wild lenticularis plants, growing
close together, differed thus:

(a) Rays normal, at right angles to axis; number of rays
in well-formed heads, 21, 21, 21, 21,.20, 21, 21.

(b) Rays normal, elevated, their plane oblique in rela-
tion to the axis of head; rays in well-formed heads,
14, 13, 14, 14, 14, 11, 15.

(c) Rays set obliquely, but less so than in b; rays in well-
formed heads, 18, 18, 18, 19. In this plant many of
the rays were modified by quilling and splitting, some
being completely quilled, i. e., hollow and tubular.
The normal rays were very obtuse, and distinctly
emarginate at end. Some showed a little red color on
apical part of middle third beneath.

Length.—The length also differs, the differences due
sometimes to race, sometimes to illumination or nutri-
tion. In our cultures mutational forms have arisen with
unusually short rays, thus:

(a) Var. vinosus with dise 55 mm. broad, rays only 35
mm. long; dise unusually convex.

620 THE AMERICAN NATURALIST [Vot. XLIX

(b) Var. bicolor with disc 64 mm. diameter, rays only 29
mm. (Next to it, in the same lot, grew a plant with
dise diameter 38, rays 47 mm.)

These measurements represent average heads from the
respective plants. A quite analogous variation was seen
in two plants of Ratibida columnifera, growing at Boulder
along with the typical form (var. nov. breviradiata, rays
yellow, only about 10 mm. long, about half the normal
size).

Torsion.—A peculiar form which appeared in our cul-
tures is the variety tortuosus, in which the ends of the
rays are twisted, as though in curl papers. This is
wholly unattractive, but other variations have the long
rays moderately curled or twisted, promising the develop-
ment of a series of forms analogous to the cactus dahlias.
As with the cactus dahlias, the rays may be rolled instead
of twisted; a wild form of this type may be described
thus:

Helianthus lenticularis var. n. angustus. Rays about
20, narrow, rolled, so that they are separated by wide in-
tervals. The rays were 36 mm. long and 5 wide (a normal
-lenticularis ray 30 mm. long is 9 wide). Disc 26 mm.
‘diameter. Goodview, Colorado, July 28, 1913.

Tubular Rays.—Under the heading ‘‘Number’’ above,
‘a case of completely quilled rays in a wild sunflower is
recorded. This peculiar modification indicates some
deep-seated tendency in the Composite, since it appears
in several genera, e. g.:

(a) Ratibida columnifera var. n. tubularis. Rays of the
usual orange color, about 25 mm. long and 3.5 broad,
completely quilled, being hollow cylinders. Flag-
staff Hill, Boulder, Colorado, July 19, 1914.

(b) Rudbeckia hirta var. tubuliforme S. H. Burnham
Amer. Botanist, Feb., 1914.

(c) Gaillardia pulchella var. fistulosa (G. fistulosa Hort.).

Emarginate and Cleft Rays.—This is another common
modification, also observed in other genera, as Ratibida
` {R. columnifera var. n. incisa; rays with one or two deep

No. 586] ANNUAL SUNFLOWERS 621

incisions, and also some narrow supplementary rays;
Boulder, Colorado, August 8, W. P. Cockerell).

Double Rows——The dise remaining normal, the rays
may be in two rows, indicating an approach to a type re-
sembling the star dahlias.

Color.—The yellow may be of various shades from
deep orange to very pale, approaching white. This has
already been discussed in Science, August 21, 1914, pp.
283-285. It may be possible eventually to get a pure
white. Dr. Church (in litt.) refers to a white form as
having been mentioned long ago by Hernandez. There is
also the development of the soluble (anthocyanin) red
pigment, giving us the chestnut red and wine red vari-
eties.

CONCLUSIONS

It is impossible at the present time to give all the evi-
dence on which opinions have been formed, but such facts
as are reported above, and others, seem to suggest the
following generalizations :

1. The number of genes or determiners in Helianthus is
not infinitely great; it is probably very much less than
exists in most animals, and the study of the processes of
heredity is relatively simple.

2. In the history of the sunflowers of the H. annuus
group, there have been few really new developments.
Species which seem very distinct prove on examination
to have few special characters of their own.

3. It is quite common for variations to arise, in wild
and cultivated plants, which appear to break the type,
and initiate something altogether new. When, however,
we begin to gather data on the variation of the Compositæ,
we find that practically all these ‘‘ new ”’ variations re-
peat themseles in various species, and at various times,
indicating that they represent deep-seated common tend-
encies. Their occurrence among wild plants shows that
they are not necessarily connected in any way with culti-
vation, and it is equally evident that they need not indi-

622 THE AMERICAN NATURALIST [Vou XLIX

cate any sort of hybridization. For example, Ratibida
columnifera presents many variations parallel with those
of Helianthus, in localities where it is the only species of
its genus.

4. We are led, then, to think of the annual sunflowers as
plants representing a certain complex of potentialities or
genes (of which we may hope at length to make a reason-
ably complete catalogue), offering these in different com-
binations at different times, usually failing to register
any permanent advance, but once in a long while reaching
a new position of stability, suited to a particular en-
vironment. These positions of stability represent what
we call the species. As with the dahlia, the horticulturist
may expect to be able to produce many interesting varie-
ties by selecting and saving the various possible combina-
tions, but analysis shows that the genes going into these
are the old ones, the effects of which may be seen from
time to time even in wild plants.

The perennial sunflowers appear to offer a more com-
plex problem. Mr. S. Alexander has found hundreds of
what are considered ‘‘ elementary species ” in Michigan.
He has been good enough to send me a large number of
these, and I ean testify that they are appreciably differ-
ent; yet they seem to represent recombinations of old
characters, already known to exist in the species of the
manuals. Some would dismiss them, along with the mul-
titudes of Crategus, as hybrids; but it does not seem justi-
fiable to assume hybridization without better evidence.
We have sufficient proof, I think, that all sorts of new
combinations of characters may arise within a type, with-
out hybridization.

Undoubtedly new determiners are formed (how, we
need not here speculate) from time to time, but the oc-
currence must be so rare and so difficult to demonstrate
that we can hardly hope to obtain satisfactory evidence
concerning it.

THE INHERITANCE OF DOUBLENESS IN MAT-
THIOLA AND PETUNIA. I. THE HYPOTHESES*

HOWARD B. FROST

Citrus EXPERIMENT STATION, UNIVERSITY OF CALIFORNIA

THE peculiar inheritance of ‘‘doubleness’’ in stocks
(Matthiola) has long been a matter of special interest.
Some races produce only single-flowering plants. A pure
double-flowering race, on the other hand, is an impossihil-
ity ; the doubles are absolutely sterile, stamens and pistils

Matthiola plants, unselected progeny of two porene showing the

s i
for photographing. The singles are plants 1, 2, 3, and 11 (mis aa). from the
left side, in the upper row, and plants 1 to 5 in the lower sary

being entirely absent. Certain races, however, consist of
both singles and doubles, in nearly equal numbers,’ each
generation being descended from the singles of the pre-
* Paper No. 17, Citrus Experiment Station, College of Agriculture, Univer-
versity of California, Riverside, California.
1 The usual proportion of doubles in large cultures seems to be near 53
per cent., or perhaps slightly higher in some cases,

623

624 THE AMERICAN NATURALIS1 [Vou. XLIX

ceding generation (see Fig. 1); the following diagram
shows the mode of inheritance in such races:

Single
|

i |
Double Single
(sterile) i

ouble Bingi
(sterile)

Miss Saunders (1911; 1913; Bateson, 1909, pp. 201-204)
has done a great amount of work on heredity in Matthiola,
and has developed an ingenious hypothesis to explain the
peculiar behavior of doubleness. Goldschmidt (1913) has
given another explanation, which has been vigorously
criticized by Miss Saunders. Several years ago (per-
haps in 1909), largely on the basis of Miss Saunders’s evi-
dence, I formulated a hypothesis somewhat simpler than
either of those just mentioned.

In view of the special interest of the case at present,
and the fact that one or both of the essential points of
my explanation have been suggested incidentally by an-
other writer (Belling, 1915, 1915a), it seems desirable to
give a general review of the hypotheses at this time.?

As Miss Saunders’s (1911; Bateson, 1909, p. 201-204)
crosses have shown, the EEE E singles are
heterozygous, the approximately 1:1 ratio being due to
the fact that the functional pollen is all double-carrying.
This is shown by Miss Saunders’s crosses between double-
throwers and pure singles. When the double-thrower
is the seed-parent, about half the F, progeny are hetero-
zygous, the rest being pure singles ; about half the double-
thrower eggs, then, are ‘‘double-carrying.’? On the other

2 My own data bearing on the problem have largely been published (Frost,
1911) or will be published in two papers (Frost, unpublished) soon to ap-
pear; some further evidence, relating to the proportions secured with some
8,000 plants of. one variety, together with a summary of my other data, is to
be presented in a paper to follow the present one. Aside, however, from one
important general feature of these results, to be briefly stated below, the

view given here is dependent on Miss Saunders’s evidence and that cited
by her.

No. 586] THE INHERITANCE OF DOUBLENESS 625

hand, when the double-thrower is the pollen-parent, all
the F, progeny are heterozygous; hence all the double-
thrower-pollen is double-carrying. These facts are il-
lustrated by the two following diagrams (adapted from
Goldschmidt) :

P: Double-thrower ?

Pure single d

| Í
F: Single (3) Single (4)
l
] J l
F- Single Single (4) pi (4) Double (4)
| | l
F: Single Single Single (4) Single (4) Double (3)
(pure) (pure) (pure (heterozygous) (sterile)
Pi Pure single Ẹ Double-thrower ¢
F: Single
l
| | |
F: Single (4) Single (3) Double (4)
g I |
F: Single Single (4) Single (4) Double (4)
(pure) (pure) (heterozygous) (sterile)

In these two crosses, where the ‘‘ singleness ”’ in the F,
(or later) heterozygotes comes entirely from the pure
single parent, we get what seems to be an ordinary Men-
delian® result in F,; the pollen of these heterozygotes
must carry both ‘‘singleness’’ and ‘‘doubleness.’’ The
absence of singleness from the double-thrower pollen is
taken by Bateson (1914, p. 292, foot-note) as almost con-
clusive evidence of somatic segregation of factors, occur-
ring in such a way that the pollen-mother-cells receive.
only doubleness. Neither he nor Miss Saunders, how-
ever, gives any reason why singleness, rather than double-
ness, should be thus eliminated. Goldschmidt (1913)
and Belling (1915, p. 126) have stated that selective
sterility of pollen will also explain the case, and definite
evidence for this view is presented below.

3In some crosses the proportion of doubles is smaller, possibly 1/16
instead of 1/4.

626 THE AMERICAN NATURALIST (Vor. XLIX

To explain the slight but constant excess of doubles

over singles, Miss Saunders assumes that two comple-
mentary linked factors, X and Y, are essential to single-
ness, and that these factors cannot be carried by the male
gametes of the double-throwers, which are all xy. X and
Y are supposed to be so linked in the ovules that the four
kinds of eggs are produced, not in equal numbers, but in
the ratio 7XY:1Xy:1xY:7xy—or else in the ratio 15:1:
1:15. Fertilization by xy pollen will give, in the former
case, TXY -xy +1Xy-xy+1xY-xy-+7xy-xy; if only
zygotes having both X and Y are single-flowering, only
the first class will consist of singles, and the doubles will
constitute 9/16, or 564 per cent., of the total. Linkage
on the 15:1 plan would give 17/32, or 534 per cent., of
doubles.
- For certain cases where crosses with pure singles have
given much less than 25 per cent. of doubles in F,, Miss
Saunders assumes the presence of a second set of two
linked factors, X’ and Y’; then any zygote receiving X or
X’ together with Y or Y’ is a single, and the proportion
of doubles is correspondingly reduced.

Von Tschermak (1912) favors Miss Saunders’s hypoth-
esis; he suggests the possibility of selective elimination
(in a dihybrid scheme), but does not consider this expla-
nation probable. It would seem, however, in view of con-
siderations stated below, that any dihybrid scheme to ex-
plain the usual slight deviation of the double-throwing
races from a 1:1 ratio is unnecessarily complex.

Goldschmidt’s (1913) hypothesis assumes selective
degeneration or sterility of pollen in the double-throw-
ers, and considers the case to be one of sex-linkage, class-
ing the slightly aberrant ratio with the known cases of
slight deviation in the sex-ratio in animals. He supposes
that this ‘‘ hermaphroditic ’’ plant is homozygous for a
distinct factor for femaleness (F), producing eggs all of
which carry this factor. He assumes that singleness is

4 Miss Saunders (1911) rather favors the latter gametic ratio, which also
corresponds closely to my own data.

No. 586] THE INHERITANCE OF DOUBLENESS 627

determined by one dominant factor, S; the eggs of the
double-thrower, then, are SF (‘‘single’’) and sF
(‘‘double’’). He assumes, also, that half the pollen-
grains or microspores, in all races of Matthiola, lack F,
probably because of elimination of part of an X-chromo-
some, and that these pollen-grains degenerate or at least
are non-functional, so that no staminate plants are pro-
duced. It is necessary to assume, then, that in the
double-throwing races S (or s) and F are carried by the
same chromosome, and that the S-carrying chromosome
is always the one to eliminate F. The S-carrying chro-
mosomes will then be the ones destined to degenerate.
The pollen resulting is of two kinds, Sf (‘‘ single,’’ non-
functional), and sF (‘‘ double,’’ functional).

The double-throwing plant, then, is SFsF ; its eggs are
SF and sF, while its pollen-grains are Sf (non-fune-
tional) and sF. Self-pollination gives, then, SFsF
(heterozygous or double-throwing singles) and sFsF
(homozygous sterile doubles).

The factor S, however, can not in itself, in general, in-
sure pollen-degeneration, since homozygous singles (SS)
produce fertile pollen. Nor can the case be one of degen-
eration of all pollen-grains receiving a maternal X-chro-
mosome, as is proved by the results of crossing SS and Ss
races. Heterozygous singles (Ss) which get the S factor
from a pure single (SS) parent, either through egg or
through sperm, produce good S pollen, as is shown by the
ordinary Mendelian ratio among their progeny (1 homo-
zygous single (SS) :2 heterozygous singles (Ss) :1 homo-
zygous double (ss). Goldschmidt is driven to assume,
therefore, that the singleness factor (S,) in the double-
throwers differs from that in the pure singles (S)—or
else to suppose that another factor interferes in the
former type.

It will be seen that Goldschmidt gives, at most, only an
indefinite implied explanation of the deviation of the
double-single ratio from equality in the double-throwing
races. And it is hard to see what advantage is secured

628 THE AMERICAN NATURALIST [Vot. XLIX

by introducing sex-factors into the discussion at all, when
all actual individuals have both stamens and pistils, or
else neither. When we assume—as Goldschmidt does—
that the factor S is so modified in the double-throwing
races as to insure the sterility of pollen-grains receiving
it, the known facts must follow; it seems wholly superflu-
ous to refer the sterility to linkage of S, with a sex-factor.
The hypothesis seems quite unnecessarily complex; there
is no real evidence here for the existence of a distinctly
heritable femaleness factor, or for any elimination of sex-
factors in pollen-formation, or for the occurrence of non-
functional pollen in ordinary pure single (SS) races.

In a reply to Goldschmidt, Miss Saunders (1913) gives
a very clear presentation of both her formulation and
that of Professor Goldschmidt, urging most of the ob-
jections to the latter scheme which are stated above, but
especially emphasizing its failure to explain the excess
of doubles over 50 per cent. She also objects to the as-
sumption of the existence of non-functional pollen, but I
can not agree with her on this point.

I have sectioned anthers prepared for cytological
study, and have frequently observed stages subsequent to
the reduction divisions. The spore-tetrads appear nor-
mal, and there seems to be no early and conspicuous evi-
dence of later degeneration. The ‘‘single’’ pollen, how-
ever, might even germinate and yet be strictly non-func-
tional because of weak growth; and, as is shown below,
the singles are actually inferior to the doubles in vigor.
Selective partial sterility seems to be a rather common
phenomenon, and it very probably occurs here.

The only other recourse seems to be the hypothesis of
somatic segregation mentioned above, and somatic segre-
gation, except as a rare accident of abnormal cell-divi-
sion, has no decisive evidence in its favor® and an over-
whelming convergence of probabilities against it. Bel-
ling (1915) calls attention to decisive evidence against

. 5 Bateson’s (1914, p. 292) positiveness in its favor seems to depend on
just such cases as that here in question.

No. 586] THE INHERITANCE OF DOUBLENESS 629

it in five genera representing as many distinct orders.
Bateson himself (1909, chap. 9) reports a fact which
seems to exclude it in the sweet pea, although his redu-
plication hypothesis? (Bateson and Punnett, 1911) would
require it there if anywhere.

This phenomenon is one to which East (1915, p. 87) has
recently referred, the ‘‘ zygotic ’’ nature of certain pol-
len-grain characters. In the sweet pea, for instance, F,
hybrids between certain races with long (dominant) and
round pollen have the pollen all long, although segrega-
tion, on any hypothesis, must have already occurred
before the shaping of the pollen-grains. If segregation
takes place as a result of chromosome-reduction, in the
formation of the spore-tetrads, it is not strange that the
cytoplasm of the young pollen-grain still retains the im-
press of the diploid maternal set of chromosomes, so that
the pollen-grains give no evidence in their shape of the
segregation that has just taken place. On the other
hand, if segregation takes place early enough to permit
of extensive ‘‘ reduplications ’’ of the cells carrying cer-
tain combinations of factors, it is very strange that the
cytoplasm of the pollen-grain should be essentially
maternal in nature. Especially does this evidence nega-
tive any hypothesis of cytoplasmic segregation—and if
segregation is nuclear, surely we have reasons enough for
connecting it with the reduction of the chromosomes.

It is due to Goldsckhmidt’s hypothesis to note that a
factor ‘‘ completely coupled ’’ with S, completely lethal
for pollen and only slightly so for the embryo-sac, would
explain the peculiarities of the case in Matthiola, both the
non-functioning of the S-ecarrying pollen and the excess
of doubles over 50 per cent. This is an amplification of
his suggestion, in a passing reference (1913, p. 81), of a

6 Bateson and Punnett explain linkage of genetic factors by means of the
hypothesis of somatic segregation. They assume that a period of cell-divi-
sion intervenes between the segregation of Mendelian factors and the
formation of the germ-cells, and that the cells bearing certain sets of factors
divide more often than the rest. This would result in making some classes
of germ-cells more numerous than others.

630 THE AMERICAN NATURALIST [Vou. XLIX

possible ‘‘ further distinct hereditary factor ’’; I merely
omit the sex-factor, and suppose the other factor to be
lethal in itself. He evidently does not notice that this
sort of factor might well explain more than the sterility
of the pollen. It amounts to the same thing, however,
to suppose the double-thrower S (or §,) to be itself the
lethal factor. The introduction of sex-factors seems en-
tirely unnecessary here, and the supposed lethal elimina-
tion of an F tactor can not be general in hermaphroditic
plants, since it would involve the universal occurrence of
sterile microspores or pollen.

Our case appears to be merely one of a hybrid showing
selective sterility of pollen-grains, a sterility due to the S
factor or to a lethal factor linked with S. Further, if
there is also a slight tendency to selective elimination of
S-carrying eggs, we have a simple and direct explanation
of the excess of doubles over the expected 50 per cent.
Or, if the s-carrying eggs are more often fertilized, the
excess of doubles is explained. Once more, selective
elimination of single (Ss) embryos might produce the
same result.

There are several facts which are extremely suggestive
in relation to all these possible forms of selective elimina-
tion. First, it is known that, in a double-throwing race,
the doubles are longer-lived than the singles in the seed
stage; Miss Saunders (1911, p. 362) has definitely con-
firmed the common belief that the proportion of doubles
tends to increase with the age of the seed. Second, Miss
Saunders (1911, p. 364) has obtained a higher proportion
of doubles from seed of lower viability, even with fresh
seed. Third, some seed-growers (deVries, 1906, p. 335)
regularly ‘‘ starve ’’ the seed-bearing plants, in the belief
that they thus increase the percentage of doubles among
the progeny. Fourth, the writer (Frost, 1911) has found,
with one variety, that inhibition of flowering by high tem-
perature is much more marked with singles than with
doubles; in field cultures, in many cases, hot weather
greatly delayed or entirely prevented flowering, the

No. 586] THE INHERITANCE OF DOUBLENESS 631

difference there being very much greater than that shown
in the tables in the paper cited.” Fifth, in the cultures
just mentioned the doubles had larger leaves than the
singles and evidently were decidedly larger as young
plants. It seems that the double form (ss) is superior to
the heterozygous single (Ss) of this double-throwing race
in general vegetative vigor, and a similar difference may
exist between s and S gametes; on these facts probably
depend the peculiarities of the observed ratio.

In order to make the case of doubleness in Matthiola
as clear as possible, let us consider a brief summary of
the formulations that have been proposed. There are
two essential points to be explained, namely: (1) the fact
that the singleness factor or set of factors of the double-
throwing races can not be carried by functional pollen,
although the corresponding factor or factor-group of the
pure single races so far tested is normal in relation to
pollen, even in single-double hybrids; (2) the fact that
the double-throwing races show a small but feini con-
stant excess of doubles over 50 per cent.

Miss Saunders gives a formally adequate but rather
complex factorial hypothesis for (2). She leaves (1),
however, essentially unexplained; she evidently relegates
it to the realm of somatic segregation, and in any case
makes no suggestion as to the real cause of the uniform
elimination of singleness.

Goldschmidt, on the other hand, gives for (1) a hypoth-
esis of selective sterility which is adequate, though of
obviously unnecessary complexity, but fails with (2)
about as completely as Miss Saunders does with (1).

It is here maintained that an extension of the general
idea of selective elimination or viability, in any one of
the several forms consistent with the evidence, complies
with all the requirements, adequately explaining both (1)
and (2). It might seem, at first thought, that the as-
sumption of a difference between S and §,, or of the exist-

7 This evidence is to be published mainly in my forthcoming paper on
‘‘ Mutation in Matthiola. ”?

632 THE AMERICAN NATURALIST [Vou XLIX

ence of a distinct linked lethal factor, makes this scheme
as complex as that of Miss Saunders; this is not the case,
however, since Miss Saunders’s scheme simply omits any
attempt at real explanation of the peculiarity of the
double-thrower pollen; her formulation imperatively re-
quires the addition of the hypothesis of selective via-
bility, or of some definite equivalent for it.®

The real puzzle of the case lies in the fact that the
double-throwers plainly differ from the pure singles so
far tested in at least two respects—heterozygosity for
singleness (ability to form sporophylls) and the associa-
tion of some peculiarity with the remaining singleness.
This, however, is essentially a problem of the origin of
the double-throwing races, and is, in any case, nowhere
simpler than with the hypothesis here suggested. Miss
Saunders’s scheme really implies four factorial or linkage
differences between pure singles and double-throwing
singles, and for certain cases six such differences, in place
of the two or three required by the hypothesis here
favored. That is, the double-thrower is supposed to
differ from the pure single in the following points: (1)
that it is heterozygous for two complementary factors
(X and Y) for which the pure single is pure, and in some
cases also for a second set of such factors (X’ and Y’);
(2) that its ‘‘ singleness ’’ can not be carried by func-
tional pollen; (3) that X and Y are partially instead of
completely linked. It is here proposed to drop half the
factors of (1), and this makes (3) superfiuous.

8 She supposes (Saunders, 1911, p. 334) that X and Y are completely
linked in the pure singles, but only partially so in the double-throwers; this
explains why, with self-pollination, the latter give 50 + per cent. of doubles
rather than 50 per cent., but not why they approximate 50 per cent. instead
of 25 per cent. It would seem, however, that the Ss hybrid between pure
single (SS) and eian thrower (S,s) may usually give double progeny
approximating, not the 25 per cent. assumed, but a slightly lower ratio.
A possible general slight deficiency of doubles in this cross is not provided
for in Miss Saunders’s hypothesis, complete linkage of X and Y explaining
why there is not an excess of doubles; whether the viability-hypothesis is
adequate depends on the general viability-relations of the S factor (as dis-
tinguished from S,), which quite possibly is even superior to s in respect to
vigo

No. 586] THE INHERITANCE OF DOUBLENESS 633

It must be admitted, however, that (3) is not in itself
improbable if (1) is true, in view of Miss Saunders’s
evidence. A similar difference between races, with re-
spect to linkage, occurs with ‘‘cream’’ flower-color,
which is partially linked with doubleness in the sulfur-
white races, but completely linked in the pure-cream
races. The essential difference of the viability-hypoth-
esis, as here presented, relates to (1); the demon-
strated lower viability of the singles, evidently the basis
of (2), makes possible the simplification of (1).

If we accept this viability-hypothesis, there seem to be
two general possibilities as to the origin of the double-
throwing races. One is that the mutation by which Ss
(double-throwing) races arise from SS (pure single)
races involves a simultaneous or consequent alteration
in the remaining S factor (or the production of a lethal
factor completely linked with S), by which the presence
of S becomes incompatible with pollen-formation.
Second, it may be that the particular race or races in
which our double-throwing forms originated had an S
factor originally different from that of the pure single
races which have been used in crossing with double-
throwers—that is to say, an S factor originally incompat-
ible with the formation of good pollen in an Ss plant—
or else that they originally possessed the lethal factor
suggested. If the second supposition is correct, such
pure single races may be found,—races which in crossing
with double-throwers never give the F, ratio 3 singles:1
double, but only approximately 1 single: 1 double.

With Petunia, if we ignore the new seed-producing
double (Francis, 1913), which has a distinct type of
flower, the general case would seem to be similarly
simple. Here, as is well known, the doubles are pro-
duced only when singles are pollinated by doubles, the
ordinary doubles having stamens but not pistils, or, at
most, non-functional rudiments of pistils. In this case
the doubleness factor (D) is plainly dominant, and is
perhaps to be considered an inhibitor; the single, then, is

634 THE AMERICAN NATURALIST — [Vov. XLIX

dd, and the double Dd, cross-pollination giving 1dd:1Dd.
There is usually (Hamüdort, 1910) an excess of singles;
here, as in Matthiola, the heterozy gous form is the one
deficient in numbers, and it is also the one which appears
inferior in vegetative vigor.? Probably the deviation
from the 1:1 ratio is due in Petunia to selective elimina-
tion of doubleness.

We have, then, in Matthiola and Petunia, hybrids
evidently due, not to the crossing of widely different
forms, but to mutation within the race,” and yet they
are partially sterile, and perhaps even lacking in vegeta-
tive vigor because of their hybridity! In connection with
the vigorous discussion of mutation now going on, it
seems worth while to ask whether, in a case like that of
(Enothera, hybridization is the cause of mutation or
mutation one great cause of hybridity; apparently both
views may be in part correct.

Miss Saunders favors a dihybrid scheme for Petunia,
evidently supposing the difference here also to depend on
two complementary factors, both necessary for single-
ness. Her assumption that singleness is dominant, as in
Matthiola, seems absolutely untenable. In considering
the last point, we may ignore the dihybrid feature, since
this evidently concerns only the deviation of the ratio
from 50 per cent.

Her formulation, as thus simplified, makes the singles
Ss and the doubles necessarily ss; the data then indicate
that the functional single pollen is all S-carrying (the
reverse of the case in Matthiola), since self-pollinated
singles produce no doubles. Then, either the single eggs
are S + s, and the double pollen s + s, or the single eggs
are s-+s, and the double pollen S+s. The latter
assumption is obviously impossible, since it not only con-
tradicts the assumption that singleness is dominant, but

® Theodore Payne, a seedsman of Los Angeles, California, says in his ~
1914 seed-catalogue, ‘‘The weaker seedlings should be carefully saved, as
these invariably produce the double flowers.’’

.10 This is not to assume that some disturbance due to crossing of two
single-flowering forms might not have led to the ‘‘mutation.’?

No. 586] THE INHERITANCE OF DOUBLENESS 635.

makes both singles and doubles heterozygous (Ss); the
former assumption, however, is also excluded, as Miss
Saunders shows, by the fact that all singles tested pro-
duce some doubles when pollinated by doubles—that is,
the expected class of pure singles (SS) does not occur.
Evidently, as both Goldschmidt (1913) and Belling (1915)
assume,!? doubleness is dominant in Petunia, and selective
viability probably completes the explanation.

BIBLIOGRAPHY
Bateson, Willia
1909 indat Principles i Heredity. 14+ 396 pp. Cambridge,
Univ. Press. 6 pl. a
1914. The Address of the President of the British Association for the
Adva slope of Science. [Heredity.] I. Science, N. S.,
Bateson, blest m, and 3 R. C.
911 n Gametie Scam Involving Reduplication of Certain Terms.
Jour. of Genetics, 1: 293-302, 1 pl. and 1 fig.
Belling, John.
1915. On the Time of Segregation of Genetic Factors in Plants. AM.
Nar., 49: 125-126, Bibliog.
1915a. "Coniitions of Mendelian Inheritance. Jour. of Heredity, 6: 108..
East, Edward M.
. The Phenomenon of Self-sterility. Am. Nart., 49: 76-87, Bibliog.
Francis, ede Shepherd.
191 A New Creation in Floriculture. The Rural Californian, 37:
397-399, 410, 411

11 In the dihybrid scheme, here, if linkage is to be invoked, as in Matthiola,
to explain the deviation from 50 per cent. of doubles, both the singles and
the doubles must carry both factors—since the singles possess both by
hypothesis, and we are supposing the pollen of the doubles to show linkage!
Further, doubles, not singles, would be expected to be in excess of 50 per
cen eam e fact, no self-consistent dihibrid scheme seems to be possible with
Petu

12 Gádsihaidi ’s statement (1913, p. 84) seems to segues that Miss
Saunders herself made this change, especially as he refers to ‘‘ Journal of
Genetics, 1, 1911’; apparently, however, he intends the alone article, pub-
lished in 1910, and I have failed to find any reference to the matter in the
part of the volume published in 1911. The explanation peat above was
stated in a letter sent to her in May, 1914, but no reply has been received.
Bateson (1913) considers singleness to be dominant, admitting the necessary
conclusion that all cultivated singles appear to be heterozygous. How this
universal heterozygosity could be maintained in self-pollination of the singles
(S +s eggs X S pollen, since the opposite assumption is untenable), he does
not explain.

636 THE AMERICAN NATURALIST [Vou. XLIX

Frost, Howard B.

1911, Variation as Related to the pager Seite Am.
Breed. Assoc, R

ept., 6: 384-395, 4 tables, arts.
[In Ms.] The Relation of Pemperatare to valves: in Matthiola.
[In Ms.] Mutation in Matthiola.
Goldschmidt, Richard
Der Vererbungmodus der gefüllten Levkojenrassen als Fall
ee oe Vererbung. Zeitsch, f. indukt. Abstam.-
. Vererbungsl., 10: 74-98. Diagr.
a Maith R.
91

Studies in the Inheritance pi poseg. in Flowers. I. Petunia.
Jour. of Geneties, 1: 57-69, 5 tables,
1911, i Re HI on F p AA of Doubleness and Other
acters in Stocks. Jour. of Genetics, 1: 303-376, 8 tables.
1913. pe pene Mode

of sed Si of Certain Characters in Double-

throwing jag A Reply. Zeitsch. f. indukt. Abst
aa 06s : 297-310.

Tschermax, Erich v

stam.- U.

1912. fis ida Miatleieneeeckc an Levkojen, Erbsen und Bohnen mit
iicksicht auf die Faktorenlehre. Zeitsch. f. indukt. Abstam.-
u. Vererbungsl., 7: 81-234, tables.
de Vries, Hugo
190 tieeaion and Varieties: Their Origin by Mutation. 2d ed.,
18 + 847 pp. Chicago, Open Court Pub. Co.

THE COAL MEASURES AMPHIBIA AND THE
CROSSOPTERYGIA

DR. ROY L. MOODIE

DEPARTMENT OF ANATOMY, UNIVERSITY OF ILLINOIS, CHICAGO

Ir has been assumed for many years that the crossop-
terygian ganoids are more nearly in the direct line of
descent of the amphibia than any other known group of
fishes. Recent work along this line adds considerable
evidence to support this assumption and it is fast becom-
ing accepted as practically proven that such was the line
of descent of this group of vertebrates. Watson (1),
Broom (2), Gregory (3-4), Pollard (5), Klaatsch (6),
Budgett (7), the writer (8), and others have added to our
knowledge of this relationship, which is based on the
structure of the skull, the limbs and the mandible, so far
as these anatomical features are known. Our knowledge
of the osteology of neither group is satisfactory and it is
to be hoped that additional material will do much toward
a solution of this problem.

It is our purpose here to state in a brief manner what
the Amphibia from the Coal Measures add toward the
solution of the problem of the derivation of the amphib-
ians from the crossopterygians.

The fish characters of the larval stages of the Amphibia
have been often cited as evidence of this relationship.
Budgett (7) says:

It has been admitted by the most competent paleontologists that the
structure of the dermal bones of the head and shoulder girdle of Poly-
pterus is so like that of certain Stegocephali, that it must be regarded
as more than a mere resemblance while there are many points in the
development of the skeleton that distinctly approach the condition of
the Amphibia. The only possible interpretation of these facts appears
to me to be that the living crossopterygians form a central group among
recent forms, having some characters in common with most of the great
groups.

The writer (8) has called attention to the similarity of
arrangement of the lateral line canals in Amphibia and
Crossopterygia and has attempted a correlation of the
cranial elements. In that essay it is stated:

637

638 THE AMERICAN NATURALIST [Vou XLIX

H ali

Sguali Paji
Mi T e

Cycjostomata

Osira rE

Aimphioxus tike ancestor
Tunicata

Posible invertebrtles ancestral frn

A phylogenetic scheme illustrating in a tentative way the possible relationships
“of the Coal Measures Amphibia to the other Chordata. The Crossopterygia may
have had more immediate ancestral relations than the diagram idicates.

No. 586] THE COAL MEASURES AMPHIBIA 639

The following elements of the stegocephalan cranium are homologous
with the same elements in (the crossopterygian) fishes: premaxille,
maxille, nasals, frontals, prefontals, parietals, squamosals, and post-
frontals. The epiotics and supraoccipitals of the Stegocephala are
homologous with the supratemporal elements of fishes. The quadrato-
jugal is homologous with the suboperecular of fishes. The supratem-
poral is homologous with the preopereulum (8).

Wilder has commented on the close relationship of
these two groups and even goes so far as to say: ‘‘that
terrestrial vertebrates were originally. derived from a
single form, perhaps a single species’’ (i. e., of the Cros-
sopterygia) (9). It will be interesting in this connection
to give a brief résumé of the geological history of the two
groups of vertebrates. The history of the Amphibia is
briefly this:

Devonian: Thinopus antiquus Marsh, footprint from
Pennsylvania.

Mississippian: Footprints from eastern North America.

Pennsylvanian: Five orders of Amphibians represented
by hundreds of more or less complete skeletons from
Europe and North America.

Permian: Four orders of Amphibia, known from an abun-
dance of material from Europe, Africa, Asia, and
North America.

Triassic: Two orders of Amphibia, the species of which
may be compared very favorably with modern spe-
cies; North America and Europe.

Comanchean: None known in North America.

Cretaceous: Caudata known from imperfect fragments,
North America.

Eocene to Recent: Frogs and salamanders as in modern
times. |

Recent: Salientia, Caudata, Gymnophiona.

The geological history of the Crossopterygia is briefly
summarized by Huxley! in the following:

The group of Crossopterygide as thus established appears to me to
have many remarkable and interesting zoological and paleontological
relations. Of the six families which compose it, four are not only
Paleozoic, but are, some exclusively and all chiefly, confined to rocks of

1 Scientifie Memoirs, IT, p. 445.

640 THE AMERICAN NATURALIST [Vou. XLIX

the Devonian age,—an epoch in which, so far as our present knowledge
goes, no fish belonging to the suborders Amiadz or Lepidosteide (unless
Cheirolepis is one of the latter) makes its appearance. Rapidly dimin-
ishing in numbers the Crossopterygide seem to have had several repre-
sentatives during the Carboniferous epoch, but after this period...
they are continued through the Mesozoic age only by a thin, though
continuous line of Celacanthini, and terminate, at the present day, in
the two or three known species of the single genus Polypterus, now
recognized under two genera: Polypterus and Erpetoichthys (Calam-
oichthys).

Such, in brief, is the geological history of this peculiar
order of fishes. At the time of the publication of Hux-
ley’s essay the fact of the geological distribution of the
group was of the greatest interest, since it proved the
transition of a vertebrate group from the Paleozoic to
later times in practically unchanged form.

Tt will thus be seen that in the Devonian we have repre-
sentatives of these two groups; one of which has been
supposed to have given origin to the other. The oldest
known Crossopterygian coexisted, in the Devonian, with
well-developed Amphibia-like forms if we may trust the
evidence of the single imprint of Thinopus antiquus
Marsh from the Devonian of Pennsylvania. That the
impression described by Professor Marsh is in reality a
footprint no one, who consults his figures, can doubt.
Its geological horizon is vouched for by the late Dr.
Charles E. Beecher, whose interpretations of geological
facts have never been impeached. So that we may say
with perfect assurance that the oldest known Crossop-
terygia existed side by side, geologically, with well-devel-
oped air-breathing, quadrupedal vertebrates. Whether
the latter were Amphibia or not is a matter which no one,
in the light of our present knowledge, can decide. The
inference, however, is that they were such.

In order that all the facts of the case in regard to the
origin of the Amphibia may be given, the following defi-
nition of Crossopterygia is given. It is based on the
first definition by Huxley and the definition contained in
Zittel’s ‘‘Paleontology’’ (ed. by Eastman).

Dorsal fins two, or if single multifid and very long; the pectoral and
Waly the ventral fins lobate; no branchiostegal rays, but two prin-

No. 586] THE COAL MEASURES AMPHIBIA 641

cipal with sometimes lateral and median jugular plates situated be-
tween the rami of the mandibles; caudal fin diphycereal or hetero-
cereal; scales cycloid or rhomboid, smooth or sculptured.

These characters may be supplemented by the follow-
ing: Notochord persistent or vertebre slightly ossified,
infraclavicle present. Teeth dendritic in a few forms.

The following list contains brief statements concerning
the structures in these two groups which are usually re-
garded as of great taxonomic importance in all vertebrates.

4. COMPARATIVE TABLE OF STRUCTURES IN AMPHIBIA AND
CROSSOPTERYGIA

Amphibia
I. Typically quadrupedal tetra-
and pentadactyl, aquatic or terres-
trial vertebrates. Body usually
provided with ventral armor of
ossified, calcified or chondrified

Body completely sealed in
one species. Seales present in
other species but incompletely

own.

II. Teeth labyrinthine in some
species.

II. Vertebre always ossified,
sometimes highly developed; as-
suming various forms. Notochord
incompletely persistent as inter-
_ central masses.

IV. Skull covered with dermal
bones, which are at times variously
sculptured and grooved by pits
and canals.

V. Lateral line canals present
as distinct impressions in the der-
mal elements.

VI. Parasphenoid largely de-
veloped.

VII. The following cranial ele-
ments correlated wit of the
Crossopterygia: Supraoccipital,
epiotie, parietal, frontal, prefron-

tal, nasal, premaxilla, maxilla,

Crossopterygia

Aquatie, fring-fined, completely
sealed, fishes with dorsal, ventral
and caudal fins

Dentine of the teeth dendritic
in a few forms

Vertebre usually unossified or
incompletely. Notochord largely
persistent and but slightly con-
stricted.

Skull covered by dermal bones
which are seldom sculptured al-
though possessing a similar type
of lateral line system contained
within the s

Lateral foe ee not always
impressions in bones, but similar
in arrangement to the Amphibia.

Parasphenoid largely developed.

The following elements absent
in Amphibia but present in the
Crossopterygia: Supratemporal os-
sicles, ethmoid, hyomandibular,
jugular plates, opercular appara-
tus consisting of opereulum and
subopereulum.

Pineal opening present in at
least one genus.

Air bladder present, with at
times calcified walls

Pelvie girdle composed of rhom-
boid plate of two parts which may
correspond to the ilia.

642

quadrate, pterygoid,
supratemporal, para-
Ethmoid

squamosal,
palantine,
sphenoid, postorbital.
present .in Gymnophiona.
ineal opening present
in most extinet species.
X. No air bladder. Lung an-
lagen on ventral surface of phar-

ynx

XI. Pelvic girdle composed of
osseous ilium, ischium and car-
tilaginous or osseous pubis.

m consisting of hu-
merus, ios ulna, carpals (some-
times cartilaginous), metacarpals
and phalanges.

XIII. Leg composed of femur,

yore fibula, tarsus (sometimes
artilaginous), metatarsus and
el

XIV. Ribs single, long or
short, heavy or slender, inter- or
intracentral.

Form of body not fish-

like.

XVI. Mesonephros functional
in adult

XVII. External gills slender

and thread-like in the young of
all living and some of the ancient
forms.

THE AMERICAN NATURALIST

(Vou. XLIX

Pectoral fin composed of pro-,
meta—, and mesopterygium, acti-
nosts and fin rays. Broom (2) ha
correlated the elements in the arm
of Sauripteris taylori with those
of the amphibian arm down to the
actinosts, which he regards as car-
pals. The fin rays would repre-
sent phalanges

Pelvic fin composed of basiost
which represents the femur, and
meta- and mesoptergyia and fin

rays.

Ribs double; dorsal, arising in-
dependently of transverse proc-
esses; ventral one an exogenous
process of vertebra, which in the
tail becomes the hemal arches by
fusion of the tips.

Form of body always typically
fish-like.
Mesonephros functional in adult.

xternal gills of young, slender,
thread-like, recalling those of lar-
val salamander.

It will be seen from the above list of comparative struc- .
tures that there are a variety of instances in which the two
groups approach each other. It is, however, to be clearly
kept in mind that the oldest known Amphibia, as indi-
cated by our present knowledge of these forms, are more
like the modern forms of Amphibia than they are like
the ancient types of fishes from which they supposedly
have been derived. In other words, a more complete
knowledge of the Coal Measures Amphibia has not served
to simplify our ideas of amphibian descent in the least,
but rather to confuse them.

All of the early Amphibia have well-developed ambula-
tory or natatory limbs and none of them are fish-like in

No. 586] THE COAL MEASURES AMPHIBIA 643

external form. Many of the representatives of the Am-
phibia in the Coal Measures of North America are highly
specialized and adapted for a variety of modes of life.
One of the most significant factors in the derivation of the
early land vertebrates from the fishes is the question of
the origin of limbs from fins, on which much has been
written from a theoretical standpoint, but nothing has
been seen in the nature of material supporting and defin-
ing the details in the process of evolution. It must be
remembered, however, that the Coal Measure forms are
Amphibia in a high stage of development and when new |
discoveries show us the anatomy of the forms from the
Mississippian, Devonian and possibly the Silurian, then
we shall be in better shape to discuss the question of the
origin of tetra- and pentadactyl limbs.

The evolutional status of the Coal Measures Amphibia
may be briefly stated. They were an assemblage of
forms, highly developed and highly specialized, with few
primitive characters which would tend to ally them di-
rectly with any known group of more primitive verte-
brates. We should expect to find among these Paleozoic
Amphibia some evidence of a transitional type of limb
(11) structure between that of Eusthenopteron (12), or
allied Crossopterygian, on the one hand, and the penta-
dactyl terrestrial vertebrate, on the other. But such evi-
dence is not forthcoming among the material at present
available. Evolutional forces had brought about a wide
diversion of faunas in the Coal Measures, so that the spe-
cies are easily separable into distinct geographic groups,
which are more distinct than are the species of Amphibia
inhabiting the same regions to-day.

The living Amphibia are much more commonly iden-
tical in eastern Ohio and northern Illinois than were the
species of the same order during the Coal Measures.
Species of Necturus, Amblystoma, Rana and Bufo are not
widely different in the two localities referred to; yet the
species of Amphibia during the Coal Measures at Mazon
Creek, Illinois, and Linton, Ohio, are widely distinct.
They are, in fact, more widely distinct than the species of

644 THE AMERICAN NATURALIST (Vor. XLIX

modern Amphibia existing in New York and in Cali-
fornia; for in these states we find modern genera in com-
mon. This wide diversion of structure between the faunas
of Linton and Mazon Creek is not due to difference in age,
since it is almost assured that eit were nearly contem-
-poraneous geologically.

The high degree of specialization attained by various
members of the Coal Measures Amphibia is remarkable.
They had become adapted to nearly every condition of
vertebrate existence, which animals of later times have
adopted. There were strictly aquatic, fossorial, terres-
trial, climbing, crawling, worm-like, snake-like, lizard-like,
crocodile-like, all living ecntomporaneously. The absence
of fish-like forms is noteworthy.

1. Watson, D. M. S. 1912. The Larger Coal Measure es seme Mem.

and Proc, Manchester Lit. and Philos. Soc., LVII, No. 1, pp. pl.
ch R. 1913. On the Structure of the Mandible in a ria

2
alia. Anat. Anz., Bd. 45, No. 2/3, pp. 73-78, figs

3. Gregory, W. K. 1911. The Limbs of Eryops ed the ig of Paiređ
Limbs from Fins. Science, N. S., 33, No. 848, pp. 508-9.

4. Gregory, yeh K. 1913. The Orois Ancestry 5 the Amphibia.
Science, N. 8., A No. 960, pp. 806-8.

5. Pollard, H. B. 1. On the Anatomy and Phylogenetic Position of
Polypterus, Fak. Po Bd. VI, p. 338; Zool. Jahrb. (Morphol.
Abth.), V, oe 7m 387-428, with pl.

6. Klaatsch. Herm 896, Die Brustflosse der Crossoptyeygiaer. Fest-
schr. fiir ei Bd. I, pp. 259-391, Taf. I-IV.

T

. Budgett, J. S. 1902. On the Structure of the Larval Polypterus.
Trans. Zool. Soc., London, XVI, Pt. VII, October; Works, pp. 154-
176, pls. X-XIII 1 907.

8. diate: R. L. 1908. The Lateral Line System in Extinct Amphibia.
Journ. Morph., XIX, No. 2, 511-540, 17 figs.

9. Wilder, H. H. 1909. ‘His story of the tities Body. New York, Henry

Holt and Co. Amphibia, pp. 541-2
10. Moodie, R. L. 1915. The Sealed Amphibia of the Coal Measures.

- , pp. 463-464, 1915.

ahi si W. K. 1913. Crossopterygian ape p the Amphibia.
Science, N. S., XXXVII, No. 960, pp. 806-808.

12. Patien. Wan. 1912, Evolution of the Vertebrates and their Kin, p,
391, Fig. 265.

SHORTER ARTICLES AND DISCUSSION
AN ANTICIPATORY MUTATIONIST

WHENEVER any new view gains acceptance it is usually found
to have been partially anticipated in the writings of various au-
thors. The mutation theory is no exception to this rule, and the
purpose of this note is to direct wider attention to the anticipa-
tion of mutationist views by Thomas Meehan. While it is known
to some that Meehan held such views, it is not, I think, generally
realized how consistently and persistently he advocated them
throughout the course of his life.

Thomas Meehan was born near London in 1826, was trained
as a gardener at Kew and afterwards came to America. He
settled in Philadelphia as a horticulturist, became a prolific
writer for agricultural and horticultural journals and finally, in
1891, established Meehan’s Monthly, a journal devoted to garden-
ing. He traveled as far as the Rocky Mountains and Alaska,
Was appointed state botanist for Pennsylvania, and was in 1875
elected a fellow of the American Association for the Advance-
ment of Science. His publications included ‘‘Native Flowers
and Ferns of the United States,’’ in four volumes, and the cur-
rent of his work continued until after his death in November,
1901.

But the phases of his active life which I wish to emphasize here
were (1) his keenness and accuracy as an observer, and (2) his
constant advocacy of discontinuity in the variations of species on
the basis of his own observations, at a time when such views were
by no means popular. Meehan was particularly active in the
Philadelphia Academy of Science, and he contributed in all no
less than 257 papers and notes to the Proceedings of that society
between the years 1862 and 1901.

Although Meehan accepted evolution with his contemporaries
and with Darwin, yet he never lost an opportunity to emphasize
the probable significance of the wide variations which he fre-
quently observed in nature, as opposed to the insensible changes
which were believed to furnish the material for evolution. The
character of his observations as well as the trend of his views,
may be indicated by a few quotations from his writings.

645

646 THE AMERICAN NATURALIST [Vor. XLIX

In a paper entitled ‘‘Change by Gradual Modification not the
Universal Law,’’? he begins as follows:

Natura non facit saltum has been accepted as a grand canon by most
naturalists, and the evident absence of connecting links has been thought
fatal to theories of evolution. My studies in plant life lead me to the
belief that one form will spring from another essentially different, and
without any gradual or insensible modifications uniting them.

He then describes a variation in Halesia tetraptera.2, The new
form had undergone a change in leaf shape and the veins were
rugose. The fiowers, instead of having a narrow tube at the base,
were open, cup-shaped, the pistil wholly enclosed and not ex-
serted. This form produced good seeds and if found wild would
be considered a new species.* He then refers to a variation in
Yucca filamentosa. One plant in hundreds threw out a more
branching panicle which opened two weeks earlier. Its charac-
ters remained and were continued in the progeny. After citing
other eases, Meehan says:

Not only do strikingly distinct forms come suddenly into existence but
once born they reproduce themselves from seed, and act in every respect
as acknowledged species.

He states that a ‘‘weeping’’ variety of the peach came into
existence ‘‘about 30 years ago,’’ and also ‘‘ten years ago a deep
blood-leaved variety appeared.’’ The following quotations from
the same paper will serve further to illustrate his views:

In over a quarter of a century of experience among living plants, I
have rarely known any striking form to have originated by gradual
modification, but always by one great leap. The slight changes are
generally in efforts backwards; as when we sow purple beech seed, some
few are a trifle paler than their parents; there is little or no hesitation
in the forward leap... .

Forms are not only called into existence suddenly, widely different
from their parents, and can reproduce themselves from seed, but they
come into existence without seed agency, and the same or similar form
in widely separated localities, and not all necessarily by seed from one
individual.

Flowers of Viola pedata were sent to him from five localities
in Pennsylvania, New York, Illinois and Indiana, having ‘‘the
two upper petals a beautiful maroon color as in the pansy.’’

1 Proc. Amer, Assoc, Adv. Sci., 1874, B. 7-12, 1875.

2 Now known as Halesia carolina L

3 This fuller account is taken from his later paper, ‘‘ Variation in
Halesia,’’ Proc. Phila. Acad., 1884, 32-33, Figs. 4, 1885.

No. 586] SHORTER ARTICLES AND DISCUSSION 647

Among the conclusions of this paper, which pretty well sums
up his views, we find,

Morphological changes in individual plants are by no means by grad-
ual modification,
and

New and widely distinet species may be suddenly evolved from pre-
existing forms without the intervention of connecting links.

A discussion followed in which Professor Morse, C. V. Riley,
Professor Gill and Asa Gray took part, but although some agreed
that there was no reason why marked changes and gradual modi-
fications should not both play a part in evolution, yet the tend-
ency was rather to look upon the former as sports which were of
little evolutionary significance. Meehan afterwards referred to
this paper as showing that:

New forms “ jumped ” into existence, and frequently these new forms
were diverse from each other, under precisely the same “ environment ”
as far as human knowledge had yet reached, as had been the surround-
ing circumstances of the parent form.

Quotations of a few of Meehan’s other papers, with notes upon
them, will serve to show the range of his ideas and the accuracy
of his observations.

“‘On the Agency of Insects in Obstructing Evolution,’’ Proc.

Phila. Acad., 1872, 235-37, 1872.

‘On Rapid Changes in the History of Species.” Proc. Phila.

Acad., 1884, 142-43, 1885.

‘‘ Persistence in Variations Suddenly Introduced,” Proc. Phila.

Acad., 1885, 116, 1886.

We see that identical forms may appear simultaneously in loealities
widely separated; and, the circles meeting, cover a district in a compara-
tively short time.

‘‘Spicate Inflorescence in Cypripedium insigne,’’ Proc. Phila.

Acad., 1885, 30-32, 1886.

Such a belief [in jumps] would tend materially to remove difficulties
in the way of theories of evolution, that now prevented a full acceptance
thereof. :

‘‘On a White-seeded Variety of the Honey Locust,” Proc. Phila.

Acad., 1885, 404, 1886.

In this paper he describes a tree of Gleditschia triacanthos
growing near Germantown, Pa., which had seeds white instead of
dark olive-brown. They also differed in shape, being nearly

648 THE AMERICAN NATURALIST [Vou. XLIX

orbicular, instead of narrowly ovate, and twice as long as broad.
He remarks in this paper:

When variations occur it is difficult for some to believe that cross-fer-
tilization, a return to some characteristic of an ancient parent, or some
accident of climate or soil had not [been] an agency in the change.

This type of difficulty is still formidable in the minds of some.
In the case he describes such explanations are excluded as in-
applicable,

‘*On Parallelism in Distinct Lines of Evolution,’’ Proc. Phila.

Acad., 1886, 294-95, 1887.

Meehan refers* to a paper given by him at the Troy meeting
of the American Association (1870), ‘‘On the Introduction of
Species by Sudden Leaps.’’ But there is no such paper in the
report, although he gave three other papers dealing respectively
with fasciation, pollination by insects, and the influence of nu-
trition on sex. His last paper, published posthumously, in Proc.
Phila. Acad., 54, 33-36 (1902), is in two parts, dealing with
“‘The Bartram Oak, in Connection with Variation and Hybrid-
ism’’ and ‘‘Observations on the Flowering of Lobelia cardinalis
and Lobelia syphilitica.’’ He may well be described with justice
and accuracy as an anticipator of the mutation theory, not on
theoretical grounds but on the basis of his own keen observations.

Meehan also contributed to the earlier volumes of the AMERI-
CAN NATURALIST, the Botanical Gazette, Torrey Bulletin and
other journals during his active life. One of these® character-
istically sets forth his mutationist views.

; R. RUGGLES @ATES

4 Proc. Phila. Acad., 1885, 30.

5 On the Relation Between Insects and the Forms and Character of Flow-
ers,’’ Bot. Gazette, 16, 176-77, 1891.

“
VOL. XLIX, NO. 587 NOVEMBER, 1915

THE
AMERICAN
NATURALIST

A MONTHLY JOURNAL
Devoted to the Advancement of the Biological Sciences with
Special Reference to the Factors of Evolution

CONTENTS
Page
I. Variability and Amphimixis. Professor L. B. WALTON - - - - - 641
TI. Genetic Studies of Several 2 atest Races of California Dear Mice: Dr.
Francis B. SUMME - - -= - -688
I. Shorter Articles and Discussion: Additional Evid f Mutation in Oenot
Dr. BRADLEY M. Davis; The Value of Inter-annual Correlations: Dr.
J. ARTHUR HARRIS - - ~ - -702

THE SCIENCE PRESS
LANCASTER, PA. GAREISON, N. Y.

NEW YORK: SUB-STATION 84

The American

Naturalist

intended for publicati.> and books, etec., intended for review should be

MSS
sent age Editor of THE AMER

ICAN NATURALIST, Garrison-on- Huoson, N

ew York,

articles containing summaries of research work bearing on the
preblems of mente evolution are especially welcome, and will be given preference

in peui

idrea reprints of iran are supplied to authors free of charge.

e hu
Further 1 reprints will be supplied a
S ions and a
subscription price is at dollars a
anadian postage twenty-five
forty cents.

Lancaster, Pa.

soriant should be sent to the publishers. The

ear.
s additional.
The advertising sioe ae Four Dollars for a

THE SCIENCE PRESS

Foreign postage is fifty cents and
The yas for single copies is
page.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Aet of
Congress of March 3, 1879.

FOR SALE
ARCTIC, Ree teat GREENLAND
BIRDS’ SK
Well Prepared ice Prices
Particulars of
- DINESEN, Bird Collect
Husavik, “North Iceland, Via Leidle, England

JAPAN NATURAL HISTORY SPECIMENS
Perfect Condition and Lowest Prices.
Specialty: Bird Skins, Oology, Entomology, Marine
Animals and others. Catalogue free. Correspond-

enoe solici
T. FUKAI, Naturalist
Konosu, Saitainn, Japan

BOOKS FOR SALE

relating to all branches of

NATURAL HISTORY

Catalogue on application

JOHN D. SHERMAN, Jr.
403 Seneca Avenue MOUNT VERNON, N. Y.

The University of Chicago]

on
during the other of
The ad y coll the
uate eges,
graduate schools, and profes-
sional sch courses in

provide c
Arta, icone Scie

Si toes tobe aT

Marine Biological Laboratory
Woods Hole, Mass.
INVESTIGATION Facilities for research in Zool

En Year Be

pe be $100 each for ae over three
months, Thirty tables are —<_
ie inners research

INSTRUCTION
July—August ology, d
of

SUPPLY
DEPARTMENT
Open the Entire Year

GEO. M. GRAY, Curator, Woods Hole, Mass

nouncement will be sent on application to

The Director, Marine Biological Laboratory, Woods Hole
Mass-

THE
AMERICAN NATURALIST

VoL. X LIX. November, 1915 No. 587

VARIABILITY AND AMPHIMIXIS

PROFESSOR L. B. WALTON

KENYON COLLEGE

A Comparative Stupy oF THE VARIABILITY IN ZyGo-
SPORES OF Spirogyra inflata (VAUCH.) FORMED BY LATERAL
(CLOSE BREEDING) AND BY ScALARIFORM (CROSS BREEDING)
CONJUGATION, AND ITS BEARING ON THE THEORY OF AMPHI-
MIXIS AND CoRRELATED PROBLEMS

Rt Prenminity onilin a. eas i io cod ae ae ERP ER es en cae 650

1. Introduction
2. Historical b
3. Material `
4. Methods

Ii Conmdorntion Of Touha iocis Aa r ea 658
1. Comparative variability in length of totaal
2. Comparative variability in diameter of zygospor
3. Comparative correlation praia length and a
4. Comparative size of zygospor

Reds UIA OL TOON ooe i ee a keel eee s cea Chenu ee es « 668
1. Comparative oe
2. Comparative
3. Comparative pep RA
4. Origin of amphimixis and of death

LV. A working hypothesis of OVOIULION s. essersi si oron irera 680
Ws I A E T E bop E REE Sass 682
AA I a E a aE O a A A E E O Oh E 80's oes a ae 684

650 THE AMERICAN NATURALIST [ Vou. XLIX
I. PRELIMINARY OUTLINE

1. Introduction

Comparative studies along statistical lines of the results
produced by cross breeding and close breeding afford
data of value bearing on the problem of evolution as well
as the subsidiary problem of the origin of amphimixis.
It has long been assumed (Weismann, ’76) that sex existed
primarily to increase variability and with the further as-
sumption that the variations thus produced were heritable
and accumulated, the differentiation of organisms was
logically explained. As a corollary to such a conclusion
the belief has long been prevalent that the offspring of
organisms produced by cross breeding were as a group
more variable than those produced by close breeding, an
idea which gained further acceptance in connection with
the investigations of Castle (’06), Jennings (’08, ’09, 712,
713) and others interested in problems of genetics. That
there was excellent evidence for exactly an opposite view
and that an analysis of the results presented by the in-
vestigators mentioned above did not bear out the conclu-
sion that variability was increased by cross breeding has
been pointed out by the writer (Walton, ’08, 712, 714) in
some earlier papers.

The importance of arriving at a correct conclusion con-
cerning the part played by hybridization and cross breed-
ing in evolution can not be overestimated. If units are
merely redistributed and form characters resulting in no
actual evolutionary progress, work along Mendelian lines
tends rather to obscure the facts of value toward solving
the problem of the origin of species as well as that of evo-
lutionary control in animal and plant breeding. It is
therefore well to obtain data from as many sources as
possible bearing on the question.

Among the species of Spirogyra, a group of alge be-
longing to the class Conjugate, there are several which
reproduce both by lateral conjugation (Fig. 1, 4) where

No. 587} VARIABILITY AND AMPHIMIXIS 651

the adjacent cells of a single filament unite to form the
zygospore, itself a young individual, and at the same time
by sealariform conjugation (Fig. 1, B) where the cells of
two distinct filaments unite to form the zygospore. Thus

1. FORMATION OF pegs gen IN Spirogyra inflata (Vauch.) by lateral
conjugation (A) close bred from the same filament, and by scalariform conjuga-
n (B) cross bred from two distinct proren z = zygospore.
there is an example of a population producing under
the same environment two groups of individuals, one
by close breeding (lateral conjugation) and the other
by cross breeding (scalariform conjugation), and a com-
parison of the variability by statistical methods should
afford evidence toward the solution of the problem pre-
sented where the offspring have arisen from a common
ancestor as indicated in the material studied.

2. Historical
Much has been published concerning hybridization,
cross and close breeding, amphimixis and parthenogene-
sis, all of which are distinguishable from one another
merely by degree, nevertheless so far as the subject under

652 THE AMERICAN NATURALIST [ Vou. XLIX

discussion is concerned, the conclusions in general have
largely been assumptions based on little or no evidence.

It was Weismann (’76) who was evidently the first to
definitely express the importance of sex in producing
variations, an idea to which he consistently held in his
subsequent papers, while Nägeli (’84), Strasburger
(’84), Hatscheck (’87), Hayeraft (’95), ete., believed like-
wise on theoretical grounds that variability was reduced
by amphimixis.

The first paper presenting tangible evidence upon the
subject was that of Warren (’99) who found that par-
thenogenetically produced Daphnia magna were slightly
more variable as measured by the ‘‘Standard Deviation’’
which had a value of 2.95, than the mothers whose ‘‘Stand-
ard Deviation” was 2.22. The small number utilized,
96 in the first instance and 23 in the second instance, to-
gether with the fact that the mothers represented a se-
lected class, only those Daphnia producing young being
included, did not allow placing much reliance in the re-
sults. Warren (’02) compared 60 parental aphids (Hya-
lopterus trirhodus) and their 368 offspring as well as a
series from 30 aphid grandparents and their 291 grand-
children. The variability was found in a comparison of
grandparents and grandchildren (parthenogenetic) to
have slightly decreased in respect to frontal breadth and
considerably increased in respect to length of right
antenna, but again objections similar to those in the pre-
ceding paper render the conclusion of little value, as
Warren himself observed.

Casteel and Phillips (’03) measured drones and workers
of Apis mellifica, the honey bee, selecting individuals at
random from different colonies, and tabulating classes
and frequencies without, however, a further application
of biometrical methods. The ‘‘range of variability” was
found to be greater in the drones than in the workers.
Lutz (’04) criticized the methods utilized in the paper,
nevertheless.variation as measured by the standard devia-

No. 587] VARIABILITY AND AMPHIMIXIS 653

tion upon calculation by Wright, Lee and Pearson (’07) »
was found greater in the drones by a difference ranging
from 0.22 to 2.63 in respect to all five characters studied
in the single group of 50 Italian workers and 50 drones
of real value for comparative purposes.

Kellogg (’06), in a preliminary paper dealing with
drones and workers of bees and also with female aphids,
concluded that not only was there no evidence that amphi-
mixis produced increased variability, but that it was an
unnecessary factor in the production of Darwinian varia-
tion. The results were summarized as follows:

(a) In all but one of the characteristics studied, the amount of varia-
tion both quantitative and qualitative, is markedly larger among the
drone bees than among the workers, and in the one exceptional char-
acteristic it is no less; (b) no more variation in wing characters is
apparent among drones or workers that have not been exposed in
imaginal condition to the rigors of personal selection than exists among
bees, drones or workers, that have been so exposed; (c) the variation
in wing characters in drone bees reared in worker cells is no greater
than that among individuals reared among drone cells; (d) the varia-
tion among drones hatched from worker laid eggs is "o larger
than that among drones hatched from queen laid eggs.

Eleven ‘‘lots’’ were studied with a small number (No.
3, 48; No. 7, 54; No. 8, 75; No. 9, 26; No. 11, 60) in many
of the ‘‘lots.’’? Even though the probable errors would
have been large and while the material was heterogene-
ous, the facts brought out are of extreme interest, par-
ticularly when considered with the results obtained by
Casteel and Phillips (703).

Wright, Lee and Pearson (’07) made a comparative
biometrical study of 129 queens, 130 drones, and 129
workers taken from a nest of the common wasp Vespa
vulgaris in Charterhouse, England. In connection with
the wing dimensions, the coefficient of variation was found
to be greatest in the worker, less in the drone, and least
in the queen, differing from the bee as noted above where
drones were more variable than workers. The conclu-
sion here of interest was:

654 THE AMERICAN NATURALIST [Von. XLIX

There is no evidence in favor of parthenogenesis resulting in a smaller
variability than sexual reproduction, for if the workers be more, the
queens are less, variable than the drones.

It was suggested by the writers that the large variabilities
of the workers might have resulted from subclasses
among them due to differentiated functions or nurtures.

Castle, Carpenter, Clark, Mast and Barrows (’06) made
observations on the variability and fertility of Drosophila
ampelophila Loew, the small fruit fly, as modified by in-
breeding and cross breeding. They found that ‘‘inbreed-
ing does not affect the variability in number of teeth on
the sex comb of the male, nor the variability in size.’’
While the conclusion is not in accord with an earlier ob-
servation (p. 780) that variability would seem to have
been increased by inbreeding so far as a comparison of
the sixth inbred generation with the sixty-first genera-
tion, the small number utilized in the sixth generation
(40 males in series A-6, B-6, C-6 each) was ground for
the opinion that such a conclusion had little value in com-
parison with data pointing in the reverse direction. If
however we calculate the coefficient of variation for the
length of the tibia, an unfortunate omission on the part
of the writers, it may be noted that the flies produced by
inbreeding are decidedly more variable than those pro-
duced by cross breeding. Data for this conclusion are
given in a subsequent part of the present paper.

Walton (’08) noted that the results of measuring zygo-
spores of Spirogyra indicated that the close-bred indi-
viduals were more variable than the cross-bred individuals
and furthermore that the data went far toward confirm-
ing the theory that sex existed for the purpose of limiting
instead of augmenting variability.

Emerson (710) found that crosses between races of
plants (maize, squash, beans, gourds) differing in size and
shape had the variability of the second (F,) generation
approximately twice as great as the variability of either
parental form or of the first (F,) generation. This he

No. 587] VARIABILITY AND AMPHIMIXIS 655

explained on the basis of the segregation of size and shape
characters. Similar results were obtained by East (711)
for maize and Hayes (’12) for tobacco.

Jennings (’11) extending and summarizing his breed-
ing experiments on Paramecium concluded that

The progeny of conjugants are more variable, in size and in certain
other respects, than the progeny of the equivalent non-conjugants.
Thus conjugation increases variation.

Later (’13) continuing his investigations he stated that
conjugation increased the variability in the rate of repro-
duction. In a subsequent part of the present paper a
somewhat critical review of the data and conclusions
therein noted is presented.

3. Material

In obtaining material early one April for the labora-
tory work of a class in biology, the collection being made
in a small pool resulting from the overflow of a rivulet, a
peculiar species of Spirogyra was noticed in which both
lateral and scalariform conjugation was taking place often
in the same filament. It was at once suggestive that a
comparison of the variability in the two groups of zygo-
spores would present facts of interest in connection with
the effect of close breeding and cross breeding on varia-
bility as well as affording evidence as to the theories of
amphimixis.

The species was first determined as Spirogyra quadrata
(Hass.) but subsequent examination indicated that it
should be classified as Spirogyra inflata (Vauch.).

The material utilized for the measurements was all pro-
cured at one time from a restricted area one or two inches
square on the surface of the pool and included only the
one form of Spirogyra, that alone being present as a mass
3 or 4 inches in diameter. Inasmuch as both lateral and
scalariform conjugation occasionally took place in the
same filament (Fig. 2) a suggestion that two species were

656 THE AMERICAN NATURALIST [ Vou. XLIX

represented can not be made for the filaments are alike in
every characteristic. Of the 500 zygospores observed 45
per cent. were produced by lateral conjugation.

G. 2. Spirogyra inflata (Vauch.) x 800, with both scalariform and lateral
conjugation in the same filament. (a) Zygospore formed by lateral conjugation.
‘ male

lateral conjugation. (e) Cell from which the protoplasm has passed to form the
zygospore in (b). Obj. 1/12, Ocul. 2. Camera lucida drawing.

4. Methods

In considering the variability of large numbers of
microscopic forms, rapid and accurate measurements are
a necessity. Pearl and Dunbar (’03) in measuring Ar-
cella used a camera lucida, marking the dimensions by
means of a needle point, and reducing to microns. Pearl
(°06) adopted a similar method for Chilomonas, using a
magnification of 689.7. Pearl (’07) in measuring Para-
mecium used a 2/3-inch objective and a No.1 ocular. By
means of a camera lucida the points to be measured were
projected on cards, marked, and measured with a vernier
calipers to tenths of millimeters. Multiplying the meas-
urements so obtained by the proper reduction factor found
by calibrating with a stage micrometer, they were re-

uced to microns and recorded. Jennings (’11) at first
measured Paramecium from a slide with an ocular mi-
crometer. Later an Edinger drawing and projection ap-
paratus was used, the projected images of the specimens
on a slide in a flat drop of 25 per cent. glycerine, with-

No. 587] VARIABILITY AND AMPHIMIXIS 657

out a cover glass which by pressure would have caused
distortion, were enlarged to 500 diameters and measured
with a milimeter ruler.

In the present study, the material was preserved in 2
per cent. formalin, the first series of measurements! being
made April 2, while measurements of 358 were completed
before May 16, and the remaining 42 finished Aug. 21 of
the same year. Swelling of the zygospores did not occur
to an appreciable extent, a possible error considered
in a subsequent part of the paper. Using a B. and L.
BB-6 microscope with a No. 1 ocular and a 1/12 oil immer-
sion, a slide with a couple of drops of fluid containing the
material was covered with a No. 2 coverglass, the super-
fluous liquid drawn off by means of a pipette, and the
preparation placed on the mechanical stage. Beginning
at the lower right-hand corner the slide was moved from
left to right and each zygospore presented in the field in
a uniformly horizontal condition, was measured. On
reaching the left margin of the preparation, the slide was
returned to the first position, moved sufficiently toward
the observer so that a new path would be traversed, and
the operation repeated. . Thus the selection was at random
and no zygospore measured twice. The dimensions were
marked on note paper by means of a camera lucida at a
magnification of 1,460 diameters, the two cross lines
representing length (a) and diameter (y) having at the
point of juncture an ‘‘S’’ or an ‘‘L’’ for scalariform or
lateral conjugation. Only those zygospores having defi-
nitely formed membranes were considered.

In the reduction of data, so soon as the projections of
the apparent dimensions were completed, the length of the
lines æ and y were measured with proportional dividers
(Keuffel and Esser No. 441 special) adjusted at the ratio
1,460 to 1,000, thus giving a reading in tenths of microns.
Accurate adjustment was made possible by means of a

1I am indebted to Dr. C. C. W. Judd, of Baltimore, Md., at that time a
senior in Kenyon College, for work in part in obtaining the first series of
measurements.

658 THE AMERICAN NATURALIST [ Vou. XLIX

micrometer screw, on the basis of the equation for similar
triangles;

1,460 mm. : 1,000 mm.— 160 mm.— x mm.: x mm.
where 160 represented the total length of the dividers,
and aw or 65.04 mm. the point of adjustment. Having
checked the adjustment, it only became necessary to note
the size of a given zygospore with the longer legs of the
instrument, then by applying the shorter legs to a milli-
meter scale, to read the result. The various constants
were then computed on the basis of the work of Pearson
and of Elderton by means of a Brunsviga calculating ma-
chine. I am indebted to Dr. H. H. Mitchell of the Uni-
versity of Illinois for checking the mathematical data.

II. ConsmERATION oF RESULTS

The direct results obtained by the statistical methods
employed are here presented. These furnish the basis
for the general discussion and conclusions which follow.
The problems of biology relating to evolution need the
application of statistical methods to studies in genetics.
In no other way will it be possible to clearly demonstrate
the relative efficiency of the different types of variation—
fluctuation, amphimutation, cumulation, ete.—in originat-
ing and maintaining the diverse forms of life that exist.
Similarly the importance or unimportance of small varia-
tions in animal and plant breeding may only thus be ex-
plained. The refinements of curve fitting are by no
means necessary, nevertheless values are thus exhibited
which are presentable in no other way.

1. Comparative Variability in Length of Zygospores

In the frequency distribution for lengths of the two
groups of zygospores (Table I) the class range adopted
was two microns as compared with a range of one micron
in the distribution of diameters. The more extended as
well as the more irregular distribution of lengths of the
lateral zygospores when compared with the scalariform

No. 587] VARIABILITY AND AMPHIMIXIS 659

zygospores is at once suggestive that the group thus
close bred, is the more variable one. It is also of some
interest to note that the empirical range of variation for
the laterally formed zygospores,—with length from 49 m.
to m.,—is considerably greater than in the scalari-
form zygospores with lengths from 47 m. to 79m. While
this is not a measure of statistical variability, it un-
doubtedly has a genetic value.

TABLE I

LENGTH OF 400 ZyGOSPORES FROM Spirogyra o (VaucH.), 200 PRO-
DUCED BY LATERAL CONJUGATION AND 200 PRODUCED BY SCALARIFORM
CONJUGATION, ARRANGED IN CLASSES ACCORDING TO FREQUENCIES.

MAGNITUDES IN 1/10m

Length of zygospores in microns.

L 1 Conjugati Scalariform Conjugation
Frequency Frequency
Class ius
Observed Calculated Observed Calculated

38.0-39.9 1 0.21 0 0.

40.0-41.9 0 0.43 0 0.

42.0—43.9 1 0.84 0 0.

44.0-45.9 8 1.49 0 0.
6.0—47 2. 2.56 3 1.17
8.0-49. 0 4.18 3 3.14
50.0—51.9 4 6.44 6 6.76
52.0-53.9 16 9.34 16 12.65
24.0-55.9 10 42.71 20 19.82
56.0—57.9 26 16.16 25 1
58.0-59.9 11 19.17 25 29.40
60.0-61.9 19 21.15 26 28.41

62.0-63.9 22 21.67 25

-0-65.9 20 21.18 18 18.30
66.0—-67.9 21 18.10 11 12.66
8.0-69.9 13 14.72 7 8.12
70.0-71.9 16 11.08 6 4.90
72.0-73.9 6 2.83
74.0-75.9 1 4.98 1 1.57
76.0-77.9 5 1 .85
78.0-79.9 2 1.66 1 45

80.0-81.9 1 .86 0 0.

82.0-83.9 1 42 0 0.
ToM aa 200 200.06 200 201.04

The general constants for the variability in the length
of the zygospore of the two groups are shown below
(Table II). It may be noted that the mean (M.) or aver-

660 THE AMERICAN NATURALIST [ Vou. XLIX

TABLE II

GENERAL CONSTANTS FOR VARIATION IN LENGTHS OF ZYGOSPORES BASED ON
200 FORMED BY LATERAL AND 200 FORMED By SCALARIFORM CONJUGA-
TION WITH A CLASS RANGE oF 2 MICRONS. STANDARD
DEVIATION IN MICRONS

Constant ase meaner oe

Name Symbol Value Prob. Error Value Prob, Error

PUI OR eh i N. PO ee es yd E ee a ae
Mean ee M. 62.38 +0.1776 60.44 +0.1345
Modes. aaa Mo. ORDES lo orae SOGLO i;e
peer deviation........ e. 7.4460 | +0.1304 5.7474 | +0.1104
at t of variation.... CV. 11.9364 | +0.0330 9.5093 | +0.0330
Bkewness. 65] 3 oe Sk. —.0356 | +0.0468 .1589 .0464

age length of the zygospores produced by lateral con-
jugation exceeds the mean of the scalariform conjugants
by 1.94 microns, while the probable error for the first
constant is +.1776 and for the second constant + .1345.
The difference is therefore a significant one so far as the
present material is concerned.

It is in the comparison of the standard deviations (c)
and the coefficients of variation (C. V.) that the results of
most interest appear, however. The former constant in
lateral conjugation has a value of 1.6986 in excess of the
same constant in scalariform conjugation, or relatively
29 per cent. This is more than thirteen times the prob-
able error. In the coefficient of variation, an abstract
number permitting comparison with similar constants in
other organisms, the results indicate that the variability
in lateral conjugation exceeds that occurring in scalari-
form conjugation by 2.4271 or relatively 26 per cent., a
result corroborated by the distribution of the diameters.
The probable errors are sufficiently small in comparison
with the differences noted, that they may be considered
negligible.

Skewness is negative in the curve for lateral conjuga-
tion, the mean being on the left side of the mode, but its
value is less than the probable error. In the curve for
scalariform conjugation skewness is positive with a value

No. 587] VARIABILITY AND AMPHIMIXIS 661

slightly more than three times the probable error. There-
fore the differences of the two constants appear to have
no particular value so far as the present material is
concerned.

The analytical constants (Table III) necessary for the

TABLE III

ANALYTICAL CONSTANTS FOR VARIATION IN LENGTH OF ZYGOSPORES FORMED
BY LATERAL SCALARIFORM CONJUGATION,

Scalariform

Constant Conjugation Conjugation
Wee ee T 13.8606 8.2583
Whe oo 8 ey eto tee — 3.9024 9.2989
Meee ver eeks wis Ces T 606.8690 245.0223
De ae ol epi NaS FeS .0057 1535
MS Po oe wei ee Ge aks .0756 3918
Bee ee O 3.1068 3.5928
Meals E des .0219 .1654

fitting of the curves indicate that type IV curves may be
used for each method of conjugation. In lateral con-
jugation the equation is

a” en 9.5995 tan —1(z/29,2687)
= : TAE SRA OR -—-9.« n —1( 2/29.
= 10.842 (1 + aage) Xe
and in scalariform conjugation similarly the equation is
—11,0945

ial a —8.9889 tan —1(«/11.50)
y= 5.0014( 1+ 555 ) Xe

while the frequency polygons and the fitted curves (Figs.
3, 4, and 5) illustrate the conditions diagramatically.

2 The following formule as the basis of the probable errors, may be noted:

PEs. = 67449 £ ha Vi + 3(8k.)2.

or)
or)
bo

THE AMERICAN NATURALIST [ Vou. XLIX

a
>

EU SRRY

© o
r—
N
p
ra

F
B
a
-x
Pa

P<]
e—
A p
AT

3
r Pa \
4 A fi
: Z AL
\ ~
o
37 39 W 43 YS 47 49 50 53 55 57 59 6i 6&3 OF 67 OF V B 75 717 17 BI 3 bs 7 89
a 3. FREQUENCY POLYGON AND FITTED CURVE R VARIATION IN LENGTH

ZYG ORES PRODUCED BY LATERAL CONJUGATION peirar IN gree
pet (Vauch.).

a
c

JAM
3 ELS
4 I SESE

by
a

E F A
~

p
bae

w | HE N

: J \

6 f à

is N

2 JH |

o A

N B u B YS 47 4q SI 5> 55 57 59 GI 63 GS 67 CF U 13 75 77 79 81 83 £5

Fic. 4. FREQUENCY POLYGON AND FITTED CURVE FOR VARIATION IN LENGTH OF
ZYGOSPORES PRODUCED BY SCALARIFORM CONJUGATION (CROSS BREEDING) IN Spiro-
gyra inflata (Vauch.).

2. Comparative Variability in the Diameter of the

Zygospores

The class range adopted in the frequency distribution

for diameters (Table IV) of the two groups of zygo-

No. 587] VARIABILITY AND AMPHIMIXIS 663

74 a8
7 é N
PY N
Fe a N
to j fi N \
7 x
fie \
ww ig K
f Z i A
`
3 "i N
T L 4 2 %
£ É £ a
N
(5 A d w
+ N,
: 47 TL
a My
> ae Fi A SN -
e d S ne
Pre a — ere
317 34 W u3 y5 yt Ye St 53 55 57 59 68 be 65°60 8 ENMANE GA
Fie. 5. CE a OF FITTED CURVES FOR Ks conan IN LENGTH OF ZyGo-
SPORES PRODUCED BY LATERAL CONJUGATION (CLOSE BREEDING) AND BY SCALARI-
FOR anne oxsuaamioN (CROSS an G) IN spirogvra iniata ea, Lateral

eonin
J

Scalarifor

spores was one micron, measurement being made at the
maximum diameter. An inspection of the distribution
shows at once the greater concentration of the variates

TABLE IV
DIAMETERS OF 400 ZYGOSPORES FROM Spirogyra quadrata (Haas.) 200 PRO-
DUCED BY LATERAL CONJUGATION AND 200 PRODUCED BY SCALARIFORM
CONJUGATION, ARRANGED IN CLASSES ACCORDING TO FREQUENCIES.
MAGNITUDES IN

DIAMETER OF ZYGOSPORES IN MICRONS.

Lateral Conjugation Scalariform Conjugation

Class Frequency Class Frequency
23.0-—23.9 0 3.0-23 x
24.0-24.9 3 24.0-24.9 0
25.0—25.9 4 25.0-25.9 3
26.0-26.9 12 26.0-26.9 9
27.0—27.9 22 27.0-27.9 13
28.0-28.9 44 28.0-28.9 38
29.0-—29.9 40 29.0-29.9 46
30.0-30.9 25 30.0-30.9 46
31.0-31.9 16 31.0-31.9 30
32.0-32.9 15 2.0-32.
33.0-33.9 10 33.0-33.9 4

0-34 9 34.0-34.9 1

To 200 ToM. es 200

664

TABLE V

THE AMERICAN NATURALIST

[Vou. XLIX

GENERAL CONSTANTS FOR VARIATION IN re eect: OF ZYGOSPORES BASED
200 FORM By SCALARIFORM

ON FORMED BY

2 LATERAL AND
CONJUGATION WITH A CLASS RANGE OF 1 MICRON

Lateral Conjugation,

Scalariform Conjugation,
Cross Bı reedi ng

POORINON ~ e opo Breeding |
Name Symbol Value | Prov. Error | Value Prob. Error
Taa ac A PEP URN N. C e eaan | Fao s
on TEE Sie aka oe BE Ce e M. | 29.66 +.1049 29.725 +.0801
bas ee ye Seas, Dts BOL EOS fi kins Ga | > eR IOO be a ec es ok
adad deviation........ o. | 2.1980 | =.0688 | 1.6796 | +.0583
icy of variation. CY | 7.5376| +.0339 5.7471 | +.0338
a egies Waa aes Sk. | .2285 | +.0505 | —.1480| +.0566
ug
mamma core
uy a
ry |
i
o zk >
4 r 7
I \
36 > ‘
Se
32 h À
i | Ree
\
2S f N \
I \
l o
24 j k
d \
90 a | X \
i | N \
z \
\6 f- y
f | Yh
12 bf ae
i E | y%
N
+ P=
© a A a `
7 A
4 s \
E E ‘
o Pog ‘
a a Ae ee

GRAM SHOW

7 WING FREQU
F ZYGOSPORES IN Spirogyra subse (Vauch.
sera FORM Gouieaies:

=L

THE

an
DIAMETER
Coxsvoatiox (INBREEDING) AND
NG e uni

Sth Pore breeding),

BY
abscissa is one mikro

cy POLYGON FOR THE VARIATION IN
) PRODU LATERAL

conjugation (cross breeding)

No. 587] VARIABILITY AND AMPHIMIXIS 665

in sealariform conjugation, and thus their smaller varia-

ility. In considering the general constants of varia-
bility (Table V) based on the diameters it is to be noted
that the means (M.) do not differ, as was found when
considering length. The standard deviation (e) and
the coefficient of variation (C. V.) once more demonstrate
the greater variability of the laterally formed zygospores.
The values of the constants for skewness (Sk.) are not
sufficient, however, when considered with the probable
error, to be of importance. The frequency polygons
(Fig. 6) illustrate conditions, although no curves have
been fitted.

3. Comparative Correlation of Length and Diameter

In view of the results obtained in a consideration of the

variability, it will be of some interest to ascertain whether
TABLE VI

CORRELATION BETWEEN LENGTH AND DIAMETER OF 200 ZyGOSPORES OF
Spirogyra inflata ei ) PRODUCED BY LATERAL CONJUGATION

1 oO |
Pam gg] 3] 2) 8/8] 2] 8) ai el 3] 3
oo Se: 21 ejej jd dj 4] dj d| d| 4] | Totas
Length ~ | 8) 8) 8) 8) R R| 3 5/8) 8) 8
| |

38.0-39.9 Lie ES Bien Bias Gone OSE 1
40.0-41.9 Ber tie ee Ree ee o 0
42.0-43.9 eee) ee bb ede dee. 1
44,0-45.9 ee e eg eed ae a 3
46.0-47.9 He oe a a a Cea eo ee 2
48.0-49.9 | BREN oes ae aN Bel ot BS 0
50.0-51. Petri. 1 TE eats FO ee
52.0-53.9 1| 2) S| 21 8/2. 2 T.J] 16
0-55.9 ple a a 10
COSTS E R (eRe é 1} 4| 6| 5| 3| 2| eS eer
58.0-59.9 akaa Cel Si 402) Sina i ek
.0-61.9 ee pat ilei 6) ailal
62.0-63.9 fe ten eg 7.6 4l.. Bl Rd 2
64.0-65.9 1 Bike 2| 5| 3| 1) 3) 2| 2/1] 20
66.0-67.9 ee Creer ee ea i aTa ii a
68.0-69.9 S a 44242
WoT e ae eg f 2] 1) 8] 1] 3] 1]...1 2] 16
72.0-73.9 geste gee Eoi 87 iaia’ &
74.0-75.9 ee Leads de a Cs Oe
6.0-77.9 Ppa Ree ids LAL Se BL les: 5
78.0-79.9 Baer cide! Theses Fe ae Cee 2
.0-81.9 ceefece cesleseens tee | do pid zà
82.0-83.9 eepepebepepepepep Heere] 1
— tele oo | 3 | 4 [12/22 |44|40|25|16|15|10| 9 | 200

666 THE AMERICAN NATURALIST [ Vou. XLIX

the inbred zygospores produced by lateral conjugation
will be more or less correlated than the cross bred zygo-
spores produced by scalariform conjugation so far as
length and diameter are concerned.

The value of perfect correlation as measured by the
constant (7) is unity, while absence of correlation allows
the value to become zero. Length is taken as the subject
class (y) and diameter (x) as the relative class in the
accompanying tables (VI and VII).

TABLE VII e

CORRELATION BETWEEN LENGTH AND DIAMETER OF 200 ZYGOSPORES OF
Spirogyra inflata (VAUCH.) PRODUCED BY SCALARIFORM CONJUGATION

Diameter (2/Sigieigigigig¢iaiaisl¢

v N N nN N AN o> | N r] c] oO | V oD
m AEE hE ra i a k PIPE E

Length E E E EE AE E
38.0-39.9 Pas o ai f 0
40.0-41.9 ae es r E A ES 0
42.0-43.9 e As ee L. 0
44.0-45.9 E Cee, oe a ae BO ee 0
46.0-47.9 Pe ee o ge eke sles ee 3
0-49.9 Wot Se A ee 3
0-51. tel eb at ge 6
52.0-53.9 1 et 2| 5| 4| 2... 16
0-55. EO) ol 6) @) ab is i 20
56.0-57.9 Piet a6) 8) Bl al ok 25
58.0-59.9 Jal Jal blis |.. J1 2
.0-61.9 Meh ry er eee 4 26
62.0-63.9 1 siell 3| 6...) 1 25
.0-65.9 LiOt 8) 4 eal gis 18
66.0-67.9 are aera eee 11
68.0-69.9 1L Sl et 1) abe 7
70.0-71.9 iali. ees ae 6
72.0-73.9 ALT AS be. 6
74.0-75.9 fen eek eL 1
76.0-77.9 es T 1
78.0-79.9 1 ne 1
Woas <. s: 1/0/13) 9 |13'38|46/46/30| 9 | 4| 1! 200

While one might infer that the longer a zygospore the
greater the diameter, such a condition is not apparent by
mere inspection of the tables in either case. Conse-
quently on solving the equations we are prepared to find
that the coefficients have an extremely low value in each
group.

No. 587] VARIABILITY AND AMPHIMIXIS 667

Lateral Conjugation ... 6.0.0.0... 60005 r = .1894 + .0460
Scalariform Conjugation ........... r= .0934 + .0473

Although in lateral conjugation the value is more than
four times the probable error, one is scarcely prepared to
state that there is greater correlation between characters
in close breeding than in cross breeding on the basis of
the data noted above. When considered with the results
presented in Table XII, the conclusion seems fully estab-
lished, however.

4. Comparative Size of Zygospores

The term ‘‘size’’ as noted in the subsequent discussion
is open to various interpretations dependent as to
whether length, diameter or volume is being considered,
a condition which to some extent complicates the inter-
pretation of size characters among multicellular organ-
isms which are in general dependent on the number
rather than the dimensions of the component cells.

Those zygospores produced by lateral conjugation
(close bred), so far as the present material is concerned,
have an average length considerably exceeding those
produced by scalariform conjugation (cross bred) while
the diameter is approximately the same. This is illus-
trated in Table VIII.

TABLE VIII
COMPARATIVE LENGTH, DIAMETER AND VOLUME OF ZYGOSPORES PRODUCED
)

BY LATERAL (CLOSE BRED) AND By SCALARIFORM (Cross BRED
CONJUGATION

Method Produced | Mean Length | Mean Diameter | Mean Volume

62.380 m. +.178) 29.660 m. +.105| 28,733 cub. m.

Lateral conjugation........
60.440 m. +. 135, 29.725 m. +.080 27,972 cub. m.

Scalariform conjugation. ....

Differences RENNY lateral |
CODIUGAOR asaos erin +1.940 m. | —.070 m. | +771 cub. m.

Consequently, here, the average zygospore produced by
lateral conjugation has a greater volume than that pro-
duced by scalariform conjugation. Utilizing the formula

668 THE AMERICAN NATURALIST [ Vou. XLIX

for computing the volume of a prolate spheroid (V =
1/6zld?) the difference is 771 cubic m. in favor of the
former, although relatively this approximates only 3 per
cent.

A question of some interest is at once suggested,
namely, the possibilities for nourishment and develop-
ment in cells of large and of small volume, inasmuch as
one with a maximum volume has relatively less surface
through which nourishment may be obtained. Thus
growth may be retarded.

III. Discussion or RESULTS

The close bred forms on the basis of the characters
studied in the given population have been found more
variable as to both length and diameter, more highly
correlated, and larger taking into consideration length
and volume. The value of the conclusions in their ap-
plication to the solution of problems of evolution is de-
pendent on the logical application of cause and result as
well as the methods of the investigation.

That the two groups of zygospores are comparatively
close bred and cross bred will scarcely be denied, par-
ticularly when it is remembered that in lateral conjuga-
tion nearly all adjacent pairs of cells in a filament had
united in the process, each pair producing a zygospore,
all pairs having originated from the same cell. With
the material taken from a part of a mass a few cen-
timeters square, a sample of a whole population has been
utilized, and from what is known of the reproduction of
Spirogyra, it may be assumed with reasonable certainty
that the entire mass had its origin from zygospores pro-
duced in a few filaments the preceding year. With prac-
tically all zygospores measured in each filament, the
eriticism that isolated zygospores of mixed descent were
studied, and that greater variability would be expected in
those produced by lateral conjugation, loses its force.
Furthermore it is believed that all investigations thus far

No. 587] VARIABILITY AND AMPHIMIXIS 669

made, upon analysis support the direct conclusions which
follow.

It may be objected that cells of mature filaments
originating from the zygospores should have been studied.
While this would have been of interest, the zygospores
themselves are individuals in the cycle of development,
and the differences as represented in the groups chosen
can not be said to have less value than data from another
part of the life cycle.

The possibility of the results being affected by the
swelling of zygospores due to the 2 per cent. formalin
used in preservation, became apparent when other duties
prevented measurements within the anticipated time.
The first series of 358 zygospores was measured between
April 2 and May 16, while the remaining 42 were meas-
ured between August 17 and 21. The question seemed
an important one, and in order to test the extent of such
an error if present, the average diameter of the last lot
was compared with that of the first lot, the values being
29.15 m. and 29.08 m., the difference of 0.07 m. being well
within the limits of the probable error. The 42 zygo-
spores measured August 17—21 happened to consist of an
equal number of lateral and scalariform individuals,
which would thus tend to eliminate an error should it
have occurred. Consequently the use of the formalin
does not appear to have affected the results.

Some evidence has been presented that new phylo-
genetic characters are more variable than older char-
acters. Thus if lateral conjugation was a recent acquisi-
tion the greater variabitiy might have been expected.
Pearl and Clawson (’07) found a higher variation in the
great chela of the crayfish, Camburus propinquus Girard,
than in the protopodites of the 2 and 3 legs, nevertheless
they preferred to attribute the result to ontogenetic
rather than to phylogenetic factors. MacDougall, Vail
and Shull (’07) stated that

the greater variability of P new characters as compared
with older ones .. . is confirme ip

670 THE AMERICAN NATURALIST [ Vou. XLIR

The conclusion is open to objection inasmuch as they
were comparing a hybrid with a single parental type and
in general the greater variability would be expected.
Consequently even admitting that lateral conjugation has
been a more recent development than scalariform con-
jugation, it would not be demonstrated that an error had
thus arisen.
1. Comparative Variability

Within the limits of the characters studied so far as
the present material is concerned, it is evident that the
zygospores produced by close breeding are more variable
than those produced by cross breeding. While it tis
another proposition to extend the conclusion and insist
that organisms produced asexually, by pure lines, or by
close breeding, are more variable than those produced
sexually or by cross breeding, it would seem that the facts
strongly support such a conclusion and in connection with
the evidence afforded by the investigations of Warren,
Casteel and Phillips, Kellogg, and Wright, Lee and Pear-
son, it certainly may be denied that amphimixis or cross
breeding as compared with other types actually produces
variations, as has long been the prevalent belief.

The question here of particular interest, however, is
that of the excess type of variability represented in
Spirogyra. Inasmuch as the material was homogeneous
in every way, it may be asserted that the greater vari-
ability exhibited by the close-bred forms is not fluctu-
ability due to environment. It is also evident that, theo-
retically, cross breeding produces a greater number of
combinations than inbreeding, nevertheless that the vari-
ability thus resulting is overwhelmed by that of another
type in nature, is clear from the results noted in the pre-
ceding pages. An excellent demonstration of such con-
dition is obtained by recalculating constants obtained by
Hayes (712) as shown in the accompanying table (Table
TX) based on data obtained in connection with the breed-
ing of Nicotiana tabacum.

No. 587] VARIABILITY AND AMPHIMIXIS 671

TABLE IX
COMPARISON OF a ois oF Nicotiana tabacum IN COMBINED PARENTAL
Types (No. 3 8) WITH VARIABILITY IN SEPARATE PARENTAL TYPES

(No. 1, 2, 6 AND a IN THE First HYBRID oe. (No. 4 AND 9)

AND IN THE SECOND HYBRID GENERATIONS (No. 5 AND 10). No. 3 AND 8
CALCULATED FROM DATA By HAYES IN TABLES NOTED. OTHER CONSTANTS
S GIVEN BY HAYES

No. Table Type | Character Ss. D; | E | Cc. Vv. E

1 AV 401 | Number of leaves 0.96 | +.037| 5.00 |+.189
21 XVI 403 | 1.49 |+.058| 5.27 | +.578
BRR pe | “ . 4.70 |+.129| 19.55 | 4.129
di 2 4 n 1.30 | +.056| 5.51 |+.215
& i XVIII ede Oey la os 2.24 | +.103| 9.40 |+.551
6 XV pet | Height of plant | 3.85 |+.150! 7.00 |+.150
7 XVI es ™ .55 |4.177| 5.98 |+.177
8 | XV-XVI 1014-408 s x 11.31 14.311] 17.35 |+.312
9| XVII 403 X ale oe 4.54 |+.177! 6.41 | +.249
10 | XVIII “403x401 — 1-F,| e a 7.22 | +.333 13.60 | +.333

Here the constants of No. 3 and No. 8 have been ob-
tained by combining the two parental types (401 and 403)
both for the number of leaves and the height of the plant,
and it may be noted that the coefficient of variation has
dropped from 19.55 to 9.40 in the one case and from 17.35
to 13.60 in the other case. Thus variability as measured
statistically has decreased. Those who have advo-
cated an increased variability as the result of hybridiza-
tion are correct when comparison is made of the F, gen-
eration with the F, generation or with a single parental
generation. They are not correct, however, in making a
general statement that cross breeding increases varia-
bility since the variability of the group composed of both
parental types must be considered and upon so doing, it
may normally be found that there has actually been a
decrease in variability.

The possibility exists however that the variability will
appear to have been increased when forms having the
same phenotype but different genotypes are bred together.
Such a condition may be illustrated by the two strains of
white sweet peas crossed by Bateson which produced
purple flowers in the first (F,) generation, and purple,

™~

672 THE AMERICAN NATURALIST [Vou. XLIX

pink, mixed and white flowers in the second (F,) genera-
tion. New combinations had arisen, but only as an ex-
pression of that which already existed in the phenotypes,
for there is no evidence of an increase in unit characters
nor was there an actual increase in variability.

There are only three papers of a statistical nature in
which it has seriously been asserted that cross-bred forms
or conjugating forms produced greater variability than
resulted in close-bred forms or non-conjugating forms.

The first is that of Castle, Carpenter, Clark, Mast and
Barrows (’06) based on a series of observations as to the
effect of cross breeding and close breeding on the varia-
bility and fertility of the small fruit fly Drosophila
ampelophila Loew. In conclusion it was stated that
‘inbreeding did not affect the variability in the number
of teeth of the sex comb of the male, nor the variability
in size,’’ the first opinion resulting from the value of the
coefficient of variation in the number of tibial spines, the
second from the standard deviation in the length of the
tibia. In the former case the data certainly do not
permit a clear conclusion one way or the other. In the
second case, however, if the value of the coefficient of
variation is computed for the length of the tibia—which,
strange to say, was not done in the original investiga-
tion—and thus allowance made for the greater length of

TABLE X
ILLUSTRATING COMPARATIVE VARIABILITY OF CROSS BRED AND INBRED FORMS
OF Drosophila AFTER COMPUTING THE VALUE OF THE COEFFICIENT OF
VARIATION FOR THE LENGTH OF TIBIA FROM DATA BY CASTLE
AND OTHERS , i
S.D. = Standard Deviation. C.V. = Coefficient of Variation.

Group Spines of Sex Comb Length of Tibia
Generation — cs Qv so ee a
Cross bred (X-1)...... 100 | 1.749.083 | 16.23 | 1.461+.070 | 3.531 +.168
Inbred (M-31)........ 100 i 568. plied | 15.51 | 1.723 +.082 | 4.452 +.212
Inbred (N-30)......... 100 | 1.684+.0 80 | 17.38 | 2.842 +.136 | 8.167 +.389
Inbred (A-61)......... 100 | 1.857+.089 17.60 | 2.041 = 097 | 5.245 +.250

No. 587] VARIABILITY AND AMPHIMIXIS 673

tibia in the cross-bred forms (Table X) the average
variability of the three inbred groups is 68 per cent.
greater than that of the cross-bred group. Consequently,
the results decidedly support the facts in the present
paper.

The remaining papers are those of Jennings (711 and
713) in a study of Paramecium. In the first paper the
breeding experiments are summarized as follows:

The progeny of conjugants are more variable, in size and in certain
other respects, than the progeny of the equivalent non-conjugants.
Thus conjugation increases variation.

It seems difficult to account for this conclusion if one
subjects the data to a critical review. So far as a ‘‘pure
race’’ is concerned the non-conjugants and their progeny
were decidedly more variable than the conjugants and
their progeny (Table 28, p. 94), although the small number
utilized March 31 for the statistical work (42 and 34) is
not sufficient to justify a conclusion in either direction.
Even in a ‘‘wild eulture’’ (Table 32, p. 99) the evidence
is too conflicting to justify a definite expression of
opinion. Of the seven comparisons here made among
the progeny, five showed an excess variability for the
conjugants, but in only one case did the difference exceed
three times the probable error, while two cases showed
an excess variability for the non-conjugants, the differ-
ence in one case exceeding twice the probable error.
Data from numbers so small (22-95) can scarcely be con-
sidered reliable. The comparison of the variability of
‘all pairs’’ and ‘‘all unpairs’’ on June 22 and June 23
denotes an excess variability for those completing con-
jugation at the beginning of the experiment.

In the second paper Jennings concluded (p. 363) that
conjugation increased the variation in the rate of repro-
duction. The variation was increased, but the explana-
tion of such increase seems comparatively simple when
it is noted that among the conjugants there were many
with a low rate of fission with death occurring. As com-

674 THE AMERICAN NATURALIST [Vou. XLIX

pared with the more normal rate of fission among non-
conjugants, this could result in nothing but an increased
variability, having, however, no bearing on the question at
issue.

At the present time, therefore, it would seem that the
preponderance of evidence demonstrates that variability
is decreased in cross breeding.

2. Comparative Size

The zygospores produced by close breeding have a
mean length of 62.38 ». + .18 ». with a mean diameter
of 29.66 ». + 10 p. and those produced by cross breeding
have a mean length of 60.44 » + .13 m. with a mean
diameter of 29.725 ». + .08 a. Thus so far as length is
concerned the close bred zygospores are relatively 3.2
per cent. larger and although slightly smaller in diameter,
when volume is considered by utilizing the formula
(V =4Anld*) the close bred forms are also 2.8 per cent.
larger. Since these results are not in accord with the
general belief that cross fertilization increases size and
vigor, terms which have a diverse usage, however, it will
be well to consider other evidence bearing on the problem
with a view of attempting an explanation which may meet
the conditions imposed.

Pearl (’07) in studying the conjugation of Paramecium
with particular reference to assortative mating, notes
that ‘‘the conjugant individuals when compared with the
non-conjugant, are shorter and narrower’’ and stated in
accordance with Calkins (’02) that the reduction in size
was quite probably dependent on functional changes con-
nected with reproduction. In Spirogyra, however, both
the close-bred and the cross-bred zygospores go through
similar reproductive processes in consequence of which
one may question the theory that the method of conjuga-
tion is the decisive factor in bringing about the result
even in Paramecium.

Jennings (711) in comparing the size of conjugant and
non-conjugant Paramecium stated that

No. 587] VARIABILITY AND AMPHIMIXIS 675

The progeny of conjugants . . . were a little larger than the progeny
of non-conjugants and the difference appears to be significant.

This conclusion was based on measurements of length and
diameter, the volume not being computed. When this is
done as shown in the accompanying table (Table XI) by

TABLE XI

COMPARISON IN SIZE OF CONJUGANT AND NON-CONJUGANT ForMs oF Para-
mecium aurelia AND THEIR PROGENY BASED ON VOLUME
(V =1/6,ld2)r FROM LENGTH AND DIAMETER MEASURE-
MENTS BY JENNINGS, 1911

| |
Experiment Non-conjugants and Progeny | Conjugants and Progeny | Non R
| -|

Diam. | Volume | | Volume | Exceed the
Date, — Length | Diam. | Conjugants
Culture No. Mi- Cub. | No. Cub. 1
1908 rons | crons | Micro ons | panos RS Microns ume

| Nf» |Mar. 31| 34| 144.59|34 87, sil 48 136.95) 35.52 | 90,471 — 2,953
Table |Apr. 10| 65) 137.97/44 139,859 61 148.20 42.30 |138,844 + 1,015
28 |Apr. 20/103) 156.48) 43.82|157, 327 108, | 160. 85 42.04 148 „849 + 8,478

Q

C: |Sept. 16/110} 132.18] ? ? h 38 121.91 ? C hey
Table |Sept. 18| 70| 116.17|31.20| 59,211, 15 1 31.20 | 65,241 — 6,030
29 |Sept. 26| 52| 122.15/34.81) 77,500 ay 112.36| 29.50 51,198 + 26,302
g (Sept. 27/118) 135.35] ? ? F 174| 118.28| ? rot
Table (Sept. 29| 10, 156.40/49.60/201,465| 6 135.33 36 | 91,833 +109,633
30 | re

œo

kz]

per rou on ee
e a)

k Sept. 12/100/ 140.20 Re ? =|336) 129. | ?
d | Table |Oct. 28| 10 136 7.60 100 673 39 131.38) 35.49 | 86,644| + 14,029
31 jOct. 30| 28 123.71| Hi 14, 75 °497| 25| 128.16) 36.32 88,520| — 13,023

utilizing the formula V = 1/6z/d?, thus allowing for slight
decreases in diameters, the facts present a different in-
terpretation.

Three (a, b, c) of the four experiments dealing with a
‘‘pure race’’ of P. aurelia indicate that the progeny of
the non-conjugants become larger, even when as a group
they are smaller (a, b?) at the beginning of the experi-
ment. While the fourth (d) indicates a reverse condi-
tion so far as the measurements of October 30 are con-
cerned, the measurements of the sixth and seventh genera-
tions immediately preceding, demonstrate that the non-
conjugants were larger. The result on October 30, where
the non-conjugants became smaller, may have depended

676 THE AMERICAN NATURALIST [ Vou. XLIX

on the elimination suggested by ‘‘all existing progeny.’’
The extraordinary diminution in length (140 ». to 123.71
p.) Suggests some disturbing factor of metabolism.

The results of the experiment with a ‘‘wild culture’?
where progeny of ‘‘unpaired’’ and ‘‘paired’’ forms of P.
caudatum (?) were considered, again suggested to Jen-
nings the greater size of the progeny of the paired indi-
viduals (conjugants), a condition which was particularly
evident in the first generation. But it must be noted that
the disturbance of the function of conjugation in ‘‘un-
pairing’’ may have produced the result. The progeny of
the ‘‘unpairs’’ were relatively becoming larger from the
first to the seventh generation. These facts taken to-
gether with the absence of measurements of mean diam-
eters by which to caleulate the mean volumes, suggest
that such a conclusion based on that part of the work
could not be accepted, and that the data strongly support
the proposition directly contrary to Jennings that the
progeny of conjugants tend to become smaller than the
progeny of non-conjugants although the latter may be
larger directly after conjugation as a result of slower
fission. Thus the evidence from various sources,
although incomplete, suggests that cross-bred unicellular
organisms are smaller than close-bred forms.

Among multicellular organisms however it has long
been recognized that hybrids usually grew to a larger
size than either parental form, as has been observed by
Kohlreuter (’63), Knight (’99), Gartner (’49), as well
as Darwin, Mendel and others, although the cause of the
increased growth has been purely conjectural. It is quite
evident that the result is due to either the increased num-
ber of cells, a suggestion made by East, to the increased
size of the cells, or to the combination of both conditions.

The question immediately arises as to the cause of the
increased size and vigor among cross-bred multicellular
organisms when the evidence indicates that cross-bred
unicellular organisms are smaller instead of larger.

No. 587] VARIABILITY AND AMPHIMIXIS 677

Some investigations in progress? suggest an answer
meeting the conditions, although more than a provisional
opinion may as yet not be ventured. This is to the effect
that the cells of cross-bred multicellular organisms are
actually smaller than the cells of pure line or inbred
organisms, and that the more rapid division is a function
of the greater ratio surface has to volume in a small cell
with the better opportunity this afforded for an increased
metabolism.

The increase of size in plant and animal forms to the
physiological limit has great importance for the future
of agriculture and stock breeding, but many subsidiary
problems must be solved before practical results are at-
tained in this direction. The relative rate of growth,
number and size of the constituent cells of pure line and
of hybrid individuals is one of the problems.

3. Comparative Correlation Resulting from Close
Breeding and Cross Breeding

‘The close-bred zygospores are more correlated as to
length and diameter than the cross-bred zygospores, but
since the difference only slightly exceeds twice the prob-
able error, the value of the result here is questionable.
Considering other investigations (Table XII), it may be
noted that the group containing close-bred, asexual or
non-conjugating organisms, is more highly correlated in
respect to characters than the group consisting of cross-
bred, sexual, or conjugating organisms, although two
exceptions, No. 12 and No. 14, are presented. An in-
teresting fact, although possibly only a coincidence, is
that cross bred zygospores of Spirogyra and of conjugat-
ing Paramecium have approximately only one half the
correlation exhibited by close bred er ECeporee of Spiro-
gyra and by non-conjugating Parameen

The explanation of the conclusion here’ reached, that
the value of a character ‘‘x’’ in cross-bred forms does
not have the same tendency t change that the value of a

8 Walton (’14).

[ Vou, XLIX

678 THE AMERICAN NATURALIST

TABLE XII

COMPARATIVE CORRELATION OF CHARACTERS IN CROSS-BRED AND CLOSE-BRED

ORGANISMS INCLUDING CONJUGANT AND NON-CONJUGANT Paramecium,
N

PARTHENOGENETIC AND SEXUALLY PRODUCED WASPS

Organism Authority pes naa Type of Development ‘oo
Paramecium | Pearl, ’07 Length and oa Ser. A| .589-+.03
e , a diameter. .278 +.04
Drosophila Barrows, ’06 | Number of Close bred, fer. A-61. 469 +.05
= _ s spines and Ser. M-31. 8+.05
Sa P "a length of F ie r. N-30. | .708+.03
aa s: " tibia. Cross bred, Ser. X-1. 41+.07
Nicotiana Hayes, ’12 Number of Close bred, aia 2 368 +.05
ph = T leaves and s 631 +.03
e i ps height. Cross bred, 403 x01. 406 +.05

ts 5 Length and Close bred, No. 4 +.0
? A pa breadth of No. os 497 +.04
1 A = ie eaf. — — 403 X401. | .818+.02
Vespa vulgaris ime Lee, | Length and | Dro 772 +.02
4 ai isi erm of iter tee 912+.01
15 E Pearson, pa aie 8+.04
Spirogyra | Walton, n, Length ‘and Close bred (Lat. C.) .189 +.05
7 fe | diameter. | Cross bred (Seal. C.) |.093+.05

related character has in eclose-bred forms, appar-
ently rests on a Mendelian basis. Its importance in
evolution, beyond the idea that more pronounced tempo-
rary combinations are thus allowed in the trial and error
plan of nature, is conjectural.

6.27
y

4. Amphimixis and Death

With the assumption that the results obtained in the
preceding investigation, together with the data presented
by other writers, when correctly analyzed, strongly sup-
ports the view that asexually produced organisms tend to
be more variable than those produced by the union of two
gametes, there is furnished evidence for the interpreta-
tion of the origin of sex—amphimixis and also for the
origin of death that would seem to rest upon a much more
secure basis than the purely speculative theories of Weis-
mann, Nägeli, Hatscheck, Metschnikoff, Minot, ete., which
have previously been advano

The chief advantage wéinied 3 in the reduction of varia-
bility, while somewhat conjectural, would appear to be

No. 587] VARIABILITY AND AMPHIMIXIS 679

that of holding organisms within limited bounds, or in
other words, asexually produced organisms in general
tend by their variability to exceed the limits of their
environment and thus perish, while organisms produced
by the mingling of two diverse lines of germ plasm with
their lessened variability meet the conditions of the com-
paratively slowly changing environment and their race
persists. This idea was proposed entirely upon specula-
tive grounds by Hatscheck (’87) who suggested that
variation would run riot if not controlled by the union of
germ cells, and it would now appear that the facts sup-
port such a proposition. While it has been suggested
that the chief function of amphimixis was that of re-
juvenation, a consideration of the discussion on ‘‘Com-
parative Size’’ as well as the recent experimental results
obtained in the production of Paramecium do not support
such an opinion to the exclusion of the hypothesis here
put forward. East and Hayes (’12) have advanced the
theory that recombinations in accordance with Mendelian
principles were the chief purpose of amphimixis. While
new combinations are thus brought about, apparently
there exists a real difficulty in understanding how transi-
tory heterozygotic forms could become of selective value
in originating and maintaining such a process.

The acceptance of the conclusion that asexually pro-
duced organisms are more variable than those produced
by amphimixis, and that thus some of the units are more
readily subject to the eliminating influences of the en-
vironment, affords a comparatively simple explanation
of the origin of death in multicellular forms which are
built up of such units—the cell. Consequently the infer-
ence is that* death occurs as the result of the continually
forming body cells becoming so variable through the
absence of control by amphimizis, that eventually some
one group fails to meet the limits imposed by the environ-
ment, and these together with the remainder of the colony

4 Walton, Science, p. 216, 1909.

680 THE AMERICAN NATURALIST [ Von. XLIX

—the individual—perish. The experiments of Wood-
ruff (711, ete.) who in extending the work of Maupas and
of Calkins was able to rear several thousand generations
of Paramecium without conjugation, as well as the in-
vestigations of Harrison subsequently elaborated by
Carrel, where human and other animal tissues main-
tained cell division for a prolonged time in an artificial
medium, are here of much interest. In each case the
result is brought about by the favorable artificial environ-
ment, and it is made more clear that death itself is wholly
or in part due to the unfavorable conditions surrounding
an organism.

IV. A Worxkine HYPOTHESIS oF EVOLUTION

Investigations during the last fifteen years, instead of
establishing evolution as the simple process of natural
selection conjectured by Darwin and others, have made it
evident that the results are due to many factors of much
complexity. While the diversity of organisms depends
on variation—their inheritance and non-inheritance—it
is becoming more and more apparent that the term is too
comprehensive and covers variations arising in organ-
isms from causes quite different from one another.

The results reached in the preceding pages indicate the
need of extending the older terminology as used by Plate,
13, and others where variations are separated into
‘*somations’’ or fluctuations induced by the environment
and not inherited, and ‘‘mutations’’ or blastovariations
arising in the germ plasm and inherited, if a clearer —
understanding is to be obtained of evolution and its ap-
plication. Therefore the following scheme is proposed.’

5 Several interesting groupings of variations have been suggested by Spill-

man, Baur and others, none of which, however, appear to meet present con-
ditions,

No. 587] VARIABILITY AND AMPHIMIXIS 681

VARIATIONS

Al, adage originating in accordance with definite
law

A.
Bi, Talend by general environmental stimuli,
(food supply, use and disuse, ete.), but not in-

r s
B2. Notinduced by environmental stimuli; inherited.
C1. Arising through the transference of factors
by the combination of two ancestral lines in
accordance with Mendelian principles, but ex-
hibiting ‘‘per se’’ no definite progress. ..... 2. Amphimutations
‘< mutations’? in part).
C2. Arising through causes at present unknown,
but which, from the progressive results ob-
tained, may be assumed to originate in accord- :
DUNE with dofimte I8WS ea is Oo eiren 3. Cumulations.6
A2, Apparently not originating in accordance with
RE ANER oo dada sci ce oe B. Abnormations.
Bı. Induced during early developmental stages of
the embryo from intracellular (?) stimuli, and
inherited.
C1. Arising through the abnormal segregation
of the hereditary material Sarian .. 4. Malsegregations
‘*mutations’’ in part).
C2, Arising by the loss of hereditary qualities,
Di, Resulting from the functional loss of
factor controlling a character ........... Defactorations
(3 aiid in part).
D2, Resulting from the partial functional
loss of a factor controlling a character. ... 6. Fractionations
is amiata) in part).
B2. Induced during the early developmental stages
of the embryo from extracellular (?) stimuli and

Hot Mhona = oe Sas Se ee a 7. Malformations.7

While any scheme presented must change as new facts
are obtained, a terminology is of value in proportion as it
gives a basis for future progress. The objection that it
is not possible to point out a specific cumulation by no -
means indicates the absence of such progressive varia-
tions taking long intervals of time, by the haphazard

6 Cumulations—from cumulo, to increase—including the names of the
following groups, with the exception of fractionations proposed by Bateson,
are based on the apparent origin of the variations.

7 Many so-called malformations originate as hi siscitiens: ete.

682 THE AMERICAN NATURALIST [ Vou. XLIX

method of nature, in which to bring about a change evi-
dent to mankind. That the weight of evidence, so far as
investigations have gone, is against evolution by means
of the other variations noted, makes the explanation the
more plausible. While it is true that Bateson (714) has
urged the consideration of the proposition that organic
changes occur through the loss of inhibiting factors—de-
factorations—such a double negative theory assumes a
decreasing complexity instead of an increasing com-
plexity of protoplasm, as already pointed out by Castle,
(715) and seems impossible to maintain.

On the interpretation here presented, the diversity of
organic forms is more complex than earlier imagined,
and the problem of positive racial improvement is still
far from solution. Loss as well as segregation factors
may add new forms which really contain nothing new.
To build up and not to break down is the desideratum,
and the data obtained would seem to suggest that pure
line breeding with the employment of statistical methods
to show any progress would be the path leading most di-
rectly to the goal.

VI. Conciusions
1. Direct Conclusions

The following conclusions drawn from the investiga-
tion are primarily statements of fact.

1. Zygospores of Spirogyra inflata (Vauch.) produced
by lateral conjugation or close breeding (quasi-partheno-
genesis) are relatively 26 per cent. more variable in
length and 31 per cent. more variable in diameter as
measured by the coefficient of variation, than those pro-
duced by scalariform conjugation or cross breeding
(sexual reproduction).

2. The size (volume) is greater in the average (mean)
zygospore close bred by lateral conjugation, where the
mean length is 62.38 ». + .178, than in the average zygo-

No. 587] VARIABILITY AND AMPHIMIXIS 683

spore cross bred by scalariform conjugation, where the
mean length is 60.44 u. + .135, The diameter is approxi-
mately the same in both types.

3. In zygospores produced by lateral conjugation there
exists a positive correlation between length and diameter
of .1894=.0460, while in sealariform conjugation the
value is .0934 = .0473. This is in general agreement with
results obtained by others although here the difference is
not significant when the probable error is considered.

4. In the material studied approximately 45 per cent.
of the zygospores were formed by lateral conjugation, the
remaining 55 per cent. by scalariform conjugation.

5. The material studied was strictly homogeneous, and
evidently arose from the same parental stock, both types
of filaments being intermingled with no structural dif-
ferences except those of conjugation. Consequently the
differences in variability are not the result of fluctuability.

2. Indirect Conclusions

The conclusions here presented are generalizations
based on the present investigation as well as the work of
others, and represent propositions concerning which dif-
ferences of opinion may exist.

1. Amphimixis, cross-breeding, etc., decreases and does
not augment variability (cumulability) although amphi-
mutability may temporarily be increased.

2. Close bred forms are more highly correlated in re-
spect to related characters than cross-bred forms.

. Variations, so far as their origin is concerned, may
be separated into (A) Normations consisting of (1)
fluctuations, (2) amphimutations, and (3) cumulations,
and into (B) Abnormations consisting of (1) malsegrega-
tions, (2) defactorations, (3) fractionations, and (4) mal-
formations.

4, Cumulations may best be investigated among organ-
isms produced asexually, by pure lines, or by close breed-
ing than by cross breeding, ete.

684 THE AMERICAN NATURALIST [ Von. XLIX

5. Sexual reproduction and cross fertilization have
been advantageous in the evolution of organisms by limit-
ing ecumulability and thus confining the progress of the
group to a path bounded by the more permanent en-
vironment.

6. Death occurs as a result of the continually forming
body cells becoming so variable through the absence of
control by amphimixis, that eventually some one group
fails to meet the limits imposed by the environment, and
these together with the remainder of the colony—the
individual—perish.

3. Hypotheses

The following opinions in the nature of hypotheses
based to a large extent on the preceding work may be
confirmed or invalidated by future investigations.

1. Variability (cumulability) will be greater in a small
and isolated population than in a large and less isolated
population.

2. Progressive evolution has resulted from factors aris-
ing through cumulations without reference to amphimuta-
tions (Mendelian combinations).

3. Characters once established by cumulations produce
by fluctuations, amphimutations, etc., the diversity of
organic life. Such secondary variations are only in-
directly the products of evolution.

BIBLIOGRAPHY

Baitsell, G. A.
"11, Conjugation of Closely Related Individuals of Stylonychia. Proc.
Soc. Exper. Biol. and Medicine, pp. 122-123.
Baitsell, G. A.
12, Experiments on the Reproduction of Hypotrichous Infusoria,
Journ. Exper. Zool., Vol. 13, pp. 47—76.
Bumpus, H. C.
’99. The Elimination of the Unfit as illustrated by the Introduced
parrow, Passer domesticus. Biol. Lect., Woods Ho
Calkins and Gregory.
713. Vatiaiions. in the Pages of a Single Exconjugant of Parame
cium caudatum. Journ. Exper. Zool., Vol. 15, pp. 429-525.

No. 587] VARIABILITY AND AMPHIMIXIS 685

eater and Phillips.
Comparative sprig ners of Drones and Workers of the Honey
Bee. Biol. Bull., Vol. 6, pp. 18-37
Castle, Carpenter, Clark, ae and Barrows.
The Effects of Inbreeding, Cross-Breeding and Selection upon
the ae and Variability of eg Proc. Am. Acad.
Arts and Sciences, Vol. 41, pp.
East, E. M.
710. The Role of Hybridization in Plant Breeding. Pop. Sci. Monthly,
Vol. 77, pp. 342-355
East and Hayes.
712. Heterozygosis in Evolution and Plant Breeding. Bull. 248,
Bureau Plant Industry, U. 8. Dept. Agric., pp. 1-58.
icscagieln M P
yore Curves and Correlation. London, MacMillan, pp.

Emerson, - i
’10. The Inheritance of Sizes and Shapes in Plants. Am. Nart., Vol.
44, zp af
Emerson and E
3. The ak of Quantitative Characters in Maize. Bull, Agr.
Exp. Station Nebraska
ewenect, B.
’87. Sexual Propagation. Prager medic. Wochenschr., p. 247.
Hayes,
712. Pecans and Inheritance in Nicotina tabacum. Conn. Agr.
Exper. Station Bull. 171, pp. 1—45. ;
Jennings, H. S.
’11. Pure Lines in the Study of Genetics in Lower Organisms. AM.
a , Vol. 45, pp. 79-89.
J oe: S.
ane Mating, Variability and Inheritance of Size in the
onjugation of Paramecium. Journ. Exp. Zool., Vol. 11, No.
4.

s.
713. The Effect of Conjugation in Paramecium. Journ. Exp. Zool.,
Vol. 14, No. 3, pp. 279-391.
Kellogg, V. L.
706. Variation in Parthenogenetic Insects. Science, Vol. 24, pp.

708. Variation in Bees. Biol. Bull., Vol. 6, pp. 217-219.
MacDougal, Vail and Shull.
707. Mutations, Variations =e oo of the @notheras. Car-
negie Inst. Pub. No.
MacCurdy and Castle.
707. Selection and Cross Breeding in Relation to the Inheritance of
Coat-Pigment and Coat Patterns in Rats and Guinea Pigs.
Carnegie Inst. Pub. No. 70.

686 THE AMERICAN NATURALIST [ Von. XLIX

Minot, C. S.
708. The Problems of Age, Growth, and Death. Putnam and Sons,
New York.

Moenkhaus, W. J.
711. The Effects of Inbreeding and Selection on the Fertility Vigor
and Sex Ratio of Drosophila ampelophila. Journ. Morph., pp.
122-154,
Nägeli, C,
’84. Mechanisch-physiologische Theorie der Abstammungslehre.
München.

Pearl, R.

07 A Biometrical te of Conjugation in Paramecium. Biomet-
a, Vol. 5, pp. 97.

Pearl and pane
0 Varlatien and Correlation in the Crayfish. Carnegie Pub. No, 64.

Pearl and Surface.

709. Is there a Cumulative Effect of Selection? Zeit. f. indukt. Ab-
stamm. u. Vererb., Bd. 11,

Pearson, Wright and Lee.

707. A Cooperative Study of Queens, Drones and Workers in Vespa
vulgaris. Biometrika, Vol. 5, pp. 407—422.

Strasburger, E.

’84. Neue Untersuchungen über die Befruchtungsvorgange bei den
Phanerogamen als Grundlage fiir eine Theorie der Zeugung.
Jena.

Spillman, W. J.

710. Notes on Heredity and Evolution. Am. Nart., pp. 750-762.

Walton, L. B.

708. The Variability of the Zygospores of Spirogyra and its bearing
on the Theory of Amphimixis. Science, p. 907. (Notes based
on the first investigations in connection with the present paper.)

Walton, L. B.

712, Amphimixis, Variability and Death; Some Facts and a Theory.
Science, Vol. 35, pp. 935-940.

Walton, L. B.
g The EEAS Control of Organisms and its Significance.
nce, Vol. 39, pp. 479—488.

Warren, E.

99. An Observation on Inheritance in Parthenogenesis. Proc. Roy.
Soc., Vol. 65, p. 154.

Warren, E.

02. Variation and Inheritance in the Parthenogenetie Generations of
the Aphis Hyalipterus triphodus (Walker). Biometrika, Vol.
1, p. 129

Weismann, A.

76. Studien zur Descendenz-Theorie. II. ose die Mechanische
Auffassung der Natur. Leipzig, W. Engelm

No. 587] VARIABILITY AND AMPHIMIXIS 687

Weismann, A.

84. Ueber Leben und Tod. Jena.

Weismann, A,
86. Die Bedeutung der sexuellen oo R fiir die Selektions-
theorie. Tagbl. Naturforsch. Jen
Whitney, A.

12. Reinvigoration produced by Cross Fertilization in Hydatina
senta. Journ, Exper. Zool., Vol. 12, No. 3, pp. 337-362.
Woodruff, L. L

11, hee. Thousand Generations of Paramecium. Arch. Protist., Bd.

21, pp. 263-266.

GENETIC STUDIES OF SEVERAL GEOGRAPHIC
RACES OF CALIFORNIA DEER—MICE?

DR. FRANCIS B. SUMNER

SCRIPPS INSTITUTE, LA JOLLA, CAL.

Some of those present may recall a resolution which
was adopted at a meeting of the Biological Society of the
Pacific, held in Berkeley, in April, 1913, endorsing a
project for the study of certain problems, related both to
genetics and to geographical distribution. During the
same year, the Scripps Institution for Biological Re-
search found it possible to undertake the execution of this
project, and the author of the present paper was chosen
to carry it out. It is my object to-day to offer a pre-
liminary report upon the results of these studies.

To those who have been so fortunate as to work in
fields which yield quicker returns than does that of ex-
perimental breeding, it may seem that something more
than a ‘‘preliminary report”? might reasonably be ex-
pected after the lapse of a year anda half. If any justi-
fication is needed for such seeming slothfulness, I need
only remark that my studies have already necessitated
the trapping of about 600 living mice, of my chosen
species, in four widely distant parts of the state, together
with the rearing of several hundred others which were
born in captivity; and that I have made measurements
of some 500 of these animals, including skeletal measure-
ments of over 400. Care of this rather large family of
pets, statistical treatment of the measurements, continu-
ous meteorological observations at several points, and the
preparation of a certain number of skins and color photo-
graphs, are also to be included in the technique of this

1 Read before a joint meeting of the American Association for the Ad-
vancement of Science (Section F), the American Society of Naturalists,
the American Society of Zoologists, the American Genetic Association, and
_ the Eugenie Research Association, at Stanford University, August 4, 1915.

: 688

No. 587] CALIFORNIA DEER-MICE 689

project. Without’the generous opportunities afforded
me by the Scripps Institution, the work could never have
been undertaken. And of an importance only second
in order I must’ mention the assistance rendered me
throughout these studies by the Museum of Vertebrate
Zoology at Berkeley.

The resolution to which I have referrsd above formu-
lated four questions which were regarded as especially
worthy of consideration in the investigations contem-
plated. These were:

1. To what extent do influences such as external condi-
tions, the exercise of organs or faculties, ete., which pro-
duce modifications of structure or function in the parent,
result in bringing about parallel changes in the offspring?

2. If such changes are, in reality, found to reappear
in the offspring, do they constitute true examples of
heredity?

3. Are the subspecies or geographical races of the sys-
tematic zoologists fixed, in the sense of being hereditary,
or do the differences by which they are distinguished
depend upon conditions which must act anew during the
lifetime of each individual?

4. If these subspecific characteristics are actually found
to ‘‘breed true,’’ do they owe their existence at the outset
to ‘‘mutations’’ or to the cumulative effect of environ-
mental influences, or to the mere fact of isolation, acting
in some way independently of those influences?

To a large section of experimental breeders in this
country, to whom ‘‘genetics’’ is synonymous with Men-
delism, such a formulation of problems as this doubtless
seems hopelessly archaic. ‘‘What is the use of raising
all these dead issues,’’ they will ask, ‘‘as if Weismann
and De Vries and Johannsen had never lived?’’ And as
for the question of subspecies, I suspect that some of our
critics would grant them no existence whatever, outside
the overwrought imagination of certain taxonomists.

Those, however, who have read dispassionately such
able compilations of evidence as are offered us, for

650 THE AMERICAN NATURALIST [ Vou. XLIX

example, by Plate? and Semon? are not likely to fall into
the shallow dogmatism which dismisses the whole ‘‘ac-
quired characters’’ question as once for all settled. And
those who have taken the trouble to carefully examine a
few trays of specimens, representing the subspecies of
some widely ranging bird or mammal, will not so readily
resort to a subjective interpretation of the phenomenon
of geographic variation.

I shall give chief attention to-day to the case of a single
species of white-footed mouse, or deer-mouse of the genus
Peromyscus. According to Osgood,‘ the chief monog-
rapher of this genus, the species maniculatus comprises
about 40 distinguishable geographic races, many of which
are so unlike that they would be given full specific rank
but for the fact that they intergrade insensibly with one
another.

My own special studies have had to do chiefly with
those subspecies of Peromyscus maniculatus which fall
within the limits of the state of California. The first
investigations have naturally been directed toward a
careful examination of mice representing each of these
local races, together with a determination, so far as pos-
sible, of the meteorological conditions to which they are
subjected in nature. <A search for correlations of any
sort between structural and environmental differences
was, of course, early undertaken.

Mice were collected at four points within the state:
Eureka, Berkeley, La Jolla, and in the Mojave Desert
near Victorville. At Eureka, Berkeley and Victorville,
self-recording instruments (thermographs and hygro-
graphs) have been left in charge of assistants for nine
to fifteen months, and recording instruments will be in-
stalled at La Jolla this summer. It is planned to con-
tinue these records for at least two years. The instru-

2 ‘<Selektionsprinzip,’’ vierte Auflage, Engelmann, 1913.

3 “í Das Problem der Vererbung ‘erworbener Eigenschaften,’ ’’ Engelmann,
1912.

4‘‘ Revision of the Mice of the American Genus Peromyscus,’’ U. S. De-
partment of Agriculture, Biological Survey, 1909.

No. 587] CALIFORNIA DEER-MICE 691

ments are placed in positions more nearly representing
the natural environmental conditions of the animals than
is customary for regular Weather Bureau stations (e. g.,
in a redwood forest on the outskirts of Eureka).

The Eureka mice are assigned to the subspecies ‘‘rubi-
dus,’’ those from the desert to ‘‘ sonoriensis,’’ while those

EVREKA

DISTRIBUTION MAP
MUSEUM OF VERTEBRATE

fi i j| Ji fi

AND NEVADA, BASED UPON THE pps BUTION MAP OF Osaoop (1909). The ae.
iest shading denotes the range of P. m. rubidus, the intermediate shading that of
gambeli, the lightest that of sonoriensts. Areas of intergradation between two
races are indicated by dotted lines.

692 THE AMERICAN NATURALIST [ Von. XLIX

from Berkeley and La Jolla are assigned by Osgood to
the same subspecies ‘‘ gambeli,’”’ although, as I shall point
out, there are certain slight differences of type between
the two.

Now, as to characters, I have made 14 measurements of
each completely measured mouse. Certain color char-
acters, not capable of quantitative expression, have also
been taken into consideration. I shall first consider the
measurable parts. I must introduce this discussion by
stating that my comparisons are entirely between animals
of the same body length. When I say that P. m. rubidus
has a longer tail than sonoriensis, I mean that this is true
for mice of equal size. Owing to the impracticability of
giving you a mathematical justification of all the steps
which I have taken, I will ask you to credit me with a
knowledge of the more elementary statistical methods.’
I must also explain that I have thus far failed to kill
and measure many animals from which I shall before
long have full data. At present these are being retained
for breeding purposes. Hence my series of measure-
ments, in certain cases, is very small.

To present these subspecific characters briefly, I may
say that, in respect to tail and foot length, rubidus stands
in a class by itself. It does not require the trained eye of
a systematist to detect the fact that this northern race.
has conspicuously longer tail and feet. In the case of
the tail, this difference is due almost wholly to a differ-
ence in the length of the individual vertebra, not to an
increase in the number of these. The other three races
(sonoriensis and the two lots of gambeli) show no statis-
tically certain differences in either of these characters.
P. m. rubidus likewise has a significantly greater skull
length and probably also a greater cranial capacity.”

5 The detailed data must be deferred until a somewhat later stage of the
work. Some of these were presented to the meeting in the form of graphs.

ê The same was found to be true of the soiree induced modifications
in tail length, described by me elsewhere for white

7 Determined by suitably cleaning and then aaie the skulls, and
weighing the volume of mercury which i filled the cranial cavity.

No. 587] CALIFORNIA DEER-MICE 693

The La Jolla race of gambeli seems to fall second in the
list in this regard.

The only significant difference in ear length is that be-
tween the two races of gambeli, the La Jolla stock having
noticeably longer ears than the Berkeley stock, while
rubidus and sonoriensis appear to be intermediate in this
respect.

As regards color differences, these relate chiefly (1) to
the depth of shade, and (2) to the extensity of the pig-
mented areas. A careful comparison of large numbers
of the Berkeley race (gambeli) and the desert race
(sonoriensis) revealed at least ten recognizable differ-
ences of this class, though in many cases these were
merely different expressions of the same fundamental
difference. None of the distinctions between these two
races are absolute ones, holding between any two individ-
uals of the contrasted races. Rather they are distinc-
tions ‘‘on the whole,’’ expressed by differences of mode
or mean. Taken collectively, however, it is likely that
these characters form an ensemble sufficiently distinct to
reveal the identity of practically every specimen.

The most widely separate of the races, in respect to
color, are rubidus and sonoriensis, the former race being
very much darker than the latter. The two lots of
gambeli occupy intermediate positions between the others.

METEOROLOGICAL Data AT Four CALIFORNIA STATIONS.

Temperature (°F.) Humidity Rainfall
Annual Annual Annual Daily Annual
Mean Range Mean range total
SPOR Ae ook os bs ea 51.6 8.9 86 Very 45
small
SED a BODE DG eae aes 56.19 14,49 838 | Small 26°
Seg ried (San ey hda 60.6 14.7 75 Small 10(—)
Mojave Desert. .......... (Mo- (Mo- (Victor- | High (Victor-
jave)! jave)! ville)!° ville)
63.6 40.4 45+ 6

8 U. S. Weather Bureau.

9 University Observator

10 Computed from observations during present experiments (only one
half year at this station).

11 Rain-gauge records of Mr. Reginald Frost.

694 THE AMERICAN NATURALIST [Vou. XLIX

Now do we find any instances of correlation between
these differences of structure or color and differences in
environmental conditions? The most conspicuous of the
structural differences relate to the greater length of the
tail and foot of the Eureka race (rubidus) as compared
with any of the other three races here considered. It is
of considerable interest to note that there is here an in-
crease in the length of these appendages as we pass to the
northward, a circumstance which is still further empha-
sized by the condition of certain Alaskan subspecies. In
fact, so far as these coastal subspecies of Peromyscus
are concerned, there seems to be, within certain limits, a
reversal of Allen’s principle of the shortening of ‘‘ periph-
eral parts’’ as we pass from south to north. The facts
here revealed are likewise out of harmony with my own
experimental results from white mice, which showed con-
clusively that low temperature and high humidity led to
a decrease, rather than an increase in the length of the
tail and foot.12 A little later I shall point out the pro-
nounced effect of certain other artificial influences upon
the length of these appendages in Peromyscus, though I
must admit that these later experimental results furnish
no more satisfactory clue to the origin of these differ-

ences in nature.
= On the whole, then, these preliminary researches do
not offer much ground for believing that the differences
found in the tails and feet of these wild races of Pero-
myscus result directly from any differences in environ-
mental stimuli, or for expecting that they will respond
appreciably to artificial climatic changes.

Passing to color differences, we seem to have here a
good illustration of that correlation between atmospheric
humidity and depth of pigmentation which has long been
recognized to hold for mammals, birds and some other
animals. If we arrange our four environments in ascend-
ing order with respect to their atmospheri¢ humidity

12 Cf. Journal of Experimental Zoology, April, 1915, and earlier papers
therein cited.

No. 587] CALIFORNIA DEER-MICE 695

(the same order holds with respect to their rainfall), we
have the series: (1) Victorville, (2) La Jolla, (3)
Berkeley and (4) Eureka. Correspondingly, the desert
mouse (sonoriensis) is the palest of the lot, while the La
Jolla mouse, the Berkeley mouse and the Eureka mouse
follow inthe order of increasing pigmentation. This
relation, when viewed in connection with a wide range of
known facts, and with certain experimental data to be
noted later, can hardly be regarded as accidental. Any
exact quantitative determination of the density of pig-
mentation would of course be difficult, and it has not yet
been attempted. But the width of the dorsal median
stripe of the tail is found to serve in some measure as an
index of the extension of the darkly pigmented areas. It
is interesting to note, in order of increasing width:
sonoriensis (28 per cent.), gambeli (32 per cent.) and
rubidus (43 per cent.). (The gambeli considered are
from La Jolla.)

Let us grant then, provisionally, some sort of causal
relationship between atmospheric humidity and the quan-
tity of pigment in the hair or feathers. Now, aside from
our ignorance of the physics and chemistry of the proc-
esses here involved, there is still a most important bio-
logical question left unsolved: Are these differences in
pigmentation between the various geographical races
germinal in their origin or are they purely somatic and
individually acquired?

It was but a few years ago that Mr. J. A. Allen’? was
shocked by the very moderate suggestion of President
Jordan’s'* that perhaps some of our subspecific differ-
ences were ‘‘ontogenetic,’’ and not racially fixed. Mr.
Allen was shocked, though unable to offer any really sub-
stantial evidence in reply. In this uncertainty over so
elementary a matter of fact, everybody suggested decisive
experiments for some one else to perform, but somehow
` no one seemed disposed to perform them. At least the

13 Science, January 26, 1906.
14 Science, December 29, 1905.

696 THE AMERICAN NATURALIST [ Von. XLIX

question of the fixity of the subspecies of mammals and
birds has, to my knowedge, never before been put to ex-
perimental test.*®

Now, in view of the subject matter of the present paper,
it would ill become me to underrate the value of such
tests. But I think that I have had enough to do with the
experimental method in zoology to make me realize its
rigid limitations. It is seldom indeed that we are able
to perform a really crucial experiment and to obtain un-
equivocal results. Moreover, the mills of the gods grind
slowly, while the single human life is short.

I am therefore disposed to attach considerable impor-
tance to what have been called ‘‘Nature’s experiments.’’
Certain of these have been cited by Grinnell and Swarth’®
in their bearing on the subspecies question. For ex-
ample, two well-marked local races or subspecies of song-
sparrow occur in southern California, on opposite sides
of the high mountain range which divides the coastal
plain from the desert interior. These races are sepa-
rated nearly everywhere by the mountain barrier, but at
certain points passes through the latter occur, permitting
of migration from one side to the other.

Now these authors find localities in which the coast
form of song-sparrow has penetrated, for considerable
distances, into the desert and has become established
there. Nevertheless, these invaders, which have perhaps
been exposed for many generations to the desert at-
mosphere, have retained the darker pigmentation and
other characteristics proper to the coastal plain.

- Again, both Grinnell and Taylor’? have taken speci-
mens of Peromyscus maniculatus, which they believe to be
typical representatives of the desert race sonoriensis,

15 It is interesting to note that the often cited case of the Porto Santo
rabbit has been recently put in quite a new light by the investigations of
G. S. Miller, who finds that this transplanted race is merely the unmodified
rabbit of southern Europe. (Catalogue of the mammals of western Europe,
in the British Museum, London, 1912.) |

16 University of California Publications in Zoology, Vol. 10, No. 9, 1913.

17 University of California Publications in Zoology, Vol. 7, No. 7, 1911.

No. 587] CALIFORNIA DEER-MICE 697

high up in the mountains of California and Nevada.
These points are continuous with the main habitat of the
subspecies in the desert lowlands and plateaus, in the
sense that no abrupt barriers intervene, but they present
very great differences in climate and vegetation.

Facts of this sort—‘‘natural experiments,’’ as we may
call them—seem to show that these subspecific differ-
ences manifest themselves in a large degree independ-
ently of climatic conditions, in other words, that they are
of germinal rather than of somatic origin.

But these ‘‘natural experiments’’ are not entirely con-
clusive, for we can never be quite certain what the actual
condition are which Nature has imposed in a given case.
Granting that these darker song-sparrows in the desert
are actually invaders from the coastal plain, we have no
means of knowing how long they have been exposed to
the desert conditions. Also is it definitely known that
their restricted habitat in certain portions of the desert
does not agree with their original habitat in respect to
those factors which are really essential in determining
their characteristic coloration?

My own first attempt at transplantation consisted in
bringing a considerable number of specimens of P. m.
sonoriensis from the vicinity of Victorville to Berkeley.!8
At the latter point, the mice were kept in cages, freely
exposed to the air, and under atmospheric conditions as
nearly natural as possible. For control, numbers of the
Berkeley race were reared in neighboring cages. The
result of this experiment I can state briefly: Neither
the originally introduced animals nor their offspring,
nor their grandchildren, have thus far shown any per-
ceptible approach to the local type. They are still
obviously of the sonoriensis race. If there is any tend-

18 The contrast between the climatic conditions at the two points is indi-
cated, to a certain extent, in the foregoing table. But this does not show
one of the most characteristic differences, namely, the relatively enormous
diurnal fluctuations of temperature and humidity in the desert, as com-
pared with those in the coastal region.

698 THE AMERICAN NATURALIST [ Vou. XLIX

ency for these mice to be modified in the direction of
gambeli, this tendency is not sufficiently great to be de-
tected through any mere gross comparison, based upon
qualitative characters. Unfortunately, these two races
differ significantly in color characters only, so that an
exact quantitative test of this question is not here possible.
But I am now rearing P. m. rubidus at La Jolla, an ex-
periment which would seem to be much more promising
of decisive results than the one just described.

Enough has been done, therefore, to prove that the color
differences between sonoriensis and gambeli are at least
in a large degree germinal and independent of environ-
mental influences acting during a single lifetime.
Whether or not they are wholly germinal, and if so,
whether they can resist change for an indefinite number
of generations, remains to be learned. Suppose that we
had a shifting of the mean to the extent of one per cent.,
or even ten per cent., in the direction of gambeli. Amid
all the natural variability, we certainly should not be able
to detect such a change with the unaided eye.

In interpreting the facts that I have offered as to the
relative fixity of these color differences of Peromyscus,
we should have due regard for various other experiments
which show that environmental influences may produce
notable changes of coloration in the lifetime of an indi-
vidual. As especially comparable with these tests of my
own, though differing completely in the outcome, may be
cited the experiments of Beebe.1® This author produced
a marked increase of pigmentation in the feathers of three
species of birds by rearing them in an atmosphere of
abnormally high humidity.

Beebe’s experiments seem to show that pigmentation
may in some cases be altered during a single lifetime by
changes of humidity. The generally known facts of geo-
graphic distribution show that there is in nature a dis-
tinct correlation between pigmentation and humidity.
My own experiments show that these geographic differ-

19 Zoologica, N. Y. Zool. Soc., Sept. 25, 1907.

No. 587] CALIFORNIA DEER-MICE 699

ences of pigmentation are capable of being relatively fixed
germinally. Bringing these three sets of facts together,
they are, in my opinion, most readily harmonized on the
assumption that the direct effects of humidity upon the
organism may finally become fixed through heredity.”
But I grant that this assumption is as yet far from
proved.

A study of the effects of captivity forms a new line of
inquiry which was hardly considered when the present
researches were undertaken. For some reasons, how-
ever, this line of investigation now seems quite as prom-
ising as the search for climatic effects. In the first place,
my comparatively meager results in this field already re-
veal striking differences between the wild stock and the
individuals of the first generation reared in captivity.
These differences relate to absolute size (the domesti-
cated generation being smaller), to length of tail and foot
(both shorter in the domesticated) and to the length of
the femur and the pelvis, which differ in the same direc-
tion. In the case of the femur, in particular, the differ-
ences are striking, even upon the most casual inspection,
and they are absolutely certain statistically. They hold
for both sexes and for both of the subspecies which have
thus far been tested in this regard. Thus far, no certain
effects upon the cranial capacity of the captive lots have
been detected.

Such differences as those found probably have nothing
to do with the direct effects of external conditions, but
are due to the activities of the animal. For this reason,
they are of much higher value as a test of the Lamarckian
principle, which could help us most in explaining the per-
fecting of active parts through use or of their degenera-
tion through disuse.

The experimental results just referred to suggested
the possibility that some of the differences found to occur
in nature might have had a functional basis. The larger

20 Whether this results from ‘‘somatic’’ or ‘‘ parallel induction’’ need
not concern us here.

700 THE AMERICAN NATURALIST [Vou. XLIX

feet and tail of P. m. rubidus, for example, might be at-
tributed to its (assumed) greater activity. Unfortu-
nately for this theory, the pelvis and femur of rubidus is
no greater (slightly smaller, it would seem) than in the
other races. Had the enlargement of the foot been
due to greater functional activity the skeletal parts named
would probably have also undergone an increase in size.
From a series of measurements upon the skeletons of
wild and domestic fowls and rabbits, Darwin inferred a
relative decrease in the size of the wings of the former
and of the cranial capacity of the latter, under the in-
fluence of domestication. These conclusions may well be
true, but the evidence offered seems to be open to several
objections. (1) They are based on too small numbers
and lack the precision demanded by modern statistical
work. (2) We can only infer the exact nature of the wild
stock from which the domesticated races are descended.
(3) We can not judge of the extent to which artificial se-
lection may have played a part in bringing about these
differences. It was perhaps considerable. (4) We do not
know how much of this modification results from use or
disuse during the animal’s own lifetime. It might be
contended that the change was not congenital at all.
Lapicque and Girard have, in the main, demonstrated
a smaller cranial capacity in domesticated animals as
compared with wild ones, but aside from the more rig-
orous statistical methods employed by these authors, the
biological significance of their results is open to the same
objections as I have stated in the case of Darwin’s.
Hatai™ has shown that the albino rat has a smaller
brain than the wild Norway rat, when individuals of the
same size are compared. This proves nothing definite,
however, as to the effects of domestication, a fact which
the author recognizes. The deficient brain of the albino
may have been part of the same. mutation which brought
about the albinic condition. Or it may have resulted
in part from the selection of tamer individuals. Or the
21 Anatomical Record, Vol. 3, 1909, p. 245.

No. 587] CALIFORNIA DEER-MICE 701

difference between the two races may be purely onto-
genetic. Indeed, Donaldson?? seems to have demon-
strated the effects of exercise upon the weight of the
nervous system of the rat.

By taking a known stock of wild mice and measuring
each successive generation reared in captivity, and by
being careful to avoid selection, it would seem that the
foregoing ambiguities of interpretation might in a large
degree be obviated. It is possible, therefore, that this
phase of the subject deserves quite as much attention as
the problems relating to the physical environment and
the distribution of subspecies.

Hybridization has thus far failed completely between
rubidus and gambeli (48 matings). The Berkeley gam-
beli has, however, been successfully crossed with sonori-
ensis and some young of an F, generation have already
been obtained. Owing to the intergrading and widely
overlapping character of the differences between these
two races, it does not seem likely that they will lend them-
selves well to Mendelian analysis. But it would be idle
for me to discuss the results of these crosses at the present
stage of the experiments. Further attempts will, of
course, be made to obtain hybrids between the more widely
separated races.

(Since writing the foregoing, several successful mat-
_ ings between rubidus and sonoriensis have been effected.)
22 Journal of Comparative Neurology, Vol. 21, No. 2, 1911.

SHORTER ARTICLES AND DISCUSSION

ADDITIONAL EVIDENCE OF MUTATION IN GNOTHERA

IN a group of recent papers Bartlett reports on the remarkable
behavior of certain wild species of Ginothera grown in large cul-
tures, which behavior he regards as strong evidence for the muta-
tion theory of De Vries. The facts are presented very clearly,
but there is, however, a point of view which has not been consid-
ered in the interpretation of the conditions in his material, certain
possibilities that must be reckoned with in the critical examina-
tion of such evidence. The suggestions that I shall offer will
eoncern chiefly the genetic purity of the forms studied, a condi-
tion which is of course basic to studies on mutation as well as to
Mendelian experimentation.

nothera pratincola Bartlett! is a small-flowered, close-polli-
nated species apparently common in the North Central States.
Seven strains derived from wild mother plants at Lexington,
Kentucky, gave rise to a variant, nummularia, which differs from
the parent type in the form of the seedling leaves, foliage, pubes-
cence of the ovary and calyx, and in the manner in which the
calyx is ruptured in the opening of the flower. Nummularia
appeared with a frequency of about 1 plant to every 400 seeds
sown and 1 plant to every 250 seedlings since the germination of
the seeds in the earth was 66 per cent. Two of the strains pro-
duced nummularia in both the F, and F, generations. Further
studies will be undertaken to determine whether pratincola will
continue to give nummularia or whether it may perhaps in later
generations produce stable individuals. Nuwmmularia develops a
low percentage of good pollen (less than 50 per cent.) while
pratincola has a high proportion (90 per cent.) ; nummularia also
forms very few good seeds to a capsule, and of these only 34 per
cent. are viable. Small cultures grown from nummularia seeds
gave no plants of pratincola, but certain new forms appeared.
The high degree of sterility both gametie and zygotic shown by
nummularia is striking and demands study, for it will make a
difference in the interpretation of the behavior of this plant
whether the sterility is physiological or genetic in character.

1 Bartlett, H. H., ‘‘ Additional Evidence of Mutation in Mnothera,’’

Bot. Gaz., Vol. LIX, p. 81, 1915.
702

No. 587] SHORTER ARTICLES AND DISCUSSION 7038

Cultures of pratincola and nummularia should be grown from
germinations established experimentally to be complete, with
records of the residue of sterile seed-like structures, and the two
forms should be crossed with the purest Gnothera known to
determine whether or not the F, hybrid generations are uniform.
Should F, hybrid generations consist of distinct classes we would
be justly suspicious of the purity of the stock.

A second paper of Bartlett? describes a series of cultures of
Ginothera stenomeres, a cruciate-flowered species from Mont-
gomery County, Maryland. Two sharply marked new types were
produced by the typical form of the species, gigas represented by
one specimen with 28 chromosomes, and lasiopetala with hairy
petals. The gigas plant appeared in the F, generation of a line of
stenomeres. The peculiarities separating it from stenomeres are
similar to the distinctions between Lamarckiana and its deriva-
tive gigas. Thus both gigas forms are more persistently biennial
in habit than their parents, both have thicker, broader leaves,
stouter stems, larger buds, thicker fruits, 4-lobed pollen grains,
and twice as many chromosomes. A progeny of 63 individuals
from the original gigas plant consisted of 54 typical gigas, 6
narrow-leaved variants, and 3 ‘‘secondary mutations’’; the form
thus, as with the gigas from Lamarckiana, produces a varied off-
spring. Two of the ‘‘secondary mutations’’ were dwarfs and one
had the characters of lasiopetala, hairy petals, and in addition
certain of the stamens were also hairy. :

Lasiopetala was noted in an F, generation and also in two cul-
tures of an F,; it is of infrequent occurrence, only 5 plants in all
being observed. The plants formed persistent rosettes (steno-
meres being annual) and only one branch produced flowers, these
with hairy petals. The pollen of lasiopetala is 40-50 per cent.
perfect; that of stenomeres 60-80 per cent. An F, progeny of
116 plants from selfed lasiopetala gave 60 per cent. typical
stenomeres and 40 per cent. lasiopetala, thus behaving like
@nothera lata and O. scintillans in throwing their parent form
Lamarckiana.

Of these two new types derived from O. stenomeres the gigas
plant is remarkable as being another of the very few cenotheras
discovered with the quadruploid number of chromosomes (28) ;
triploid forms usually named semigigas have been described from
a number of lines. The hairiness of the petals in lasiopetala is

2 Bartlett, H. H., ‘‘The Mutations of Mnothera stenomeres,’’ Amer.
Jour. Bot., Vol. II, p. 100, 1915.

704 THE AMERICAN NATURALIST [ Vou. XLIX

regarded by Bartlett as a character new to the genus. Bartlett
emphasizes the fact that the characters of neither type could be
interpreted as the result of segregation following hybridization,
which may be true, but I do not think from this that it follows
that neither type can be the result of hybridization. I am not
willing to admit that hybrids present only combinations of char-
acters derived from their parent lines. It seems to me reasonable
to believe that in hybrids at times the interaction of elements
modifies the old or produces new factors. The species stenomeres
and the derivatives gigas and lasiopetala have not been tested for
genetic purity by cross-breeding with relatively stable types and
the problems of gametie and zygotic sterility have not yet been
attacked.

The final paper of this group of Bartlett’s* deals with an ex-
tremely interesting situation developed in cultures of Gnothera
Reynoldsii from Knoxville, Tennessee. This is also a small-
flowered, close-pollinated species, and its peculiarity lies in an
ability to throw extraordinarily large classes of dwarfs. There
are two types of dwarfs: (1) semialta somewhat smaller than the
typical Reynoldsit and intermediate between it and the smaller
dwarf; (2) debilis. A plant of Reynoldsii in the F, produced 29
individuals like itself, 32 plants of semialta, and 18 of debilis,
i. e., 60 per cent. of its offspring were dwarfs. Semialta throws
debilis, but no Reynoldsii. Debilis apparently can produce no
Reynoldsii or semialta and breeds true except for an occasional
variant bilonga which was also found in one culture from typical
Reynoldsii.

With respect to the dwarfs we have here presented a beautiful
series leading from the unstable parent type Reynoldsii through
the more stable semialta to the most stable and most extreme
dwarf debilis. Bartlett calls the behavior mutation en masse, but
confesses that it bears a certain degree of resemblance to Men-
delian segregation. We should very much like to see this study
repeated on a larger scale and with experimental germination of
the seed so that we may be sure of the ratios and also certain that
the cultures have given us all of their possible progeny. Segre-
gation en masse seems to the writer likely to be a more probable
explanation of the phenomena than mutation.

The form bilonga derived from the dwarf debilis offers a partic-
ularly interesting problem. It is similar to semialta except that

3 Bartlett, H. H., ‘‘Mutation en masse,’? AMER. Nat., Vol. XLIX,
p. 129, 1915.

No.587] SHORTER ARTICLES AND DISCUSSION 705

the fruits are twice as long. The capsules sometimes reach the
length of 70 mm. and average above 60 mm.; they are very much
the longest fruits reported in the subgenus Onagra. Bartlett re-
gards this large size of capsule as the origin of a new character.
Now capsule size obviously depends upon the number of ovules
produced which develop into seeds. It thus becomes an impor-
tant matter to obtain the data on ovule sterility in the species
Reynoldsii and the derived forms semialta, debilis and bilonga.
Ovule sterility is widespread among the Œnotheras as all stu-
dents of the genus know. Should it be found that bilonga pro-
duces a very much greater number of ovules than debilis and the
other types this fact would indicate a true progressive advance.
It may, however, be found that the smaller size of the capsules of
Reynoldsti semialta, and debilis is due to ovule sterility, i. e., to
the inability of a large proportion of the ovules to set seed. This
would.point to a very different interpretation of the conditions in
bilonga, and might indicate that bilonga is an example of rever-
sion towards an ancestral type in which a large capsule was corre-
lated with a high degree of ovule fertility.

In the comments which I have presented on the extremely
interesting facts discovered by Bartlett no attempts have been
made to offer exact explanations in line with Mendelian analysis.
It is not difficult to spin hypotheses on assumptions which have
been neither established nor disproved, but such creations are
hardly worth the effort when the facts are within grasp. My
main point is a constant questioning of the genetic purity of the
material with which Bartlett has worked from the standpoint
developed in my forthcoming paper ‘‘The Test of a Pure Species
of Gnothera.’’* It is impossible to discuss this subject in the
short space of a review. The most important test is that of cross-
breeding with the purest species known, to judge from the uni-
formity of the F, hybrid generation whether or not the parent
types are pure. I also firmly believe that all exact genetical work
on cenotheras must make use of methods of experimental germi-
nation to ensure complete progenies from the viable seeds and tc
permit the preservation of a residue of ungerminated structures
that may be examined.® There is in addition the determination

4To “apo in the Proceedings of the American Philosophical Society,
Vol. LIV, 1915.

5 Davis, x M., ‘‘A Method of Obtaining Complete Germination of Seeds
in Enothera and of Recording the Residue of Sterile Seed-like Structures,’’
Proc. Nat. Acad. Sci., Vol. I, p. 360, 1915.

706 THE AMERICAN NATURALIST [Von. XLIX

of degrees of sterility both gametic and zygotic, and the consid-
eration of whether such sterility is genetic or physiological.
From such tests it is possible to reach much clearer conclusions on
the genetic purity of @nothera material than has been possible
in the past.

Finally reference should be made to the important confirmation
by De Vries® of the studies of Stomps on @nothera biennis L.
In large cultures totaling 8,500 plants from Stomps’s selfed line
De Vries obtained 8 plants of a dwarf biennis nanella about 0.1
per cent., 4 plants of biennis semigigas (21 chromosomes) about
0.05 per cent., and 27 plants of the color variety biennis sulfurea
about 0.3 per cent. Since the percentages from Lamarckiana
are for nanella 1-2 per cent. and for semigigas 3 per cent., it
would appear that biennis is the more stable of the two species,
although the color variety biennis sulfurea is a new type in
experimental studies in @inothera. A culture of over 1,000
plants from selfed seed of biennis sulfwrea, all with pale yellow
flowers, produced 2 dwarfs, thus giving what De Vries calls a
‘‘double mutant,’’ O. biennis mut. sulfurea mut. nanella.

This behavior of Gnothera biennis is to the writer much more
trustworthy evidence for mutation than that presented from the
studies on Lamarckiana since biennis has a record of a long
history as a species on the sand hills of Holland, where there ap-
pears to have been little probability of recent contamination.
However, the showing of ‘‘mutants’’ from biennis does not ap-
pear very encouraging for the mutation theory of organic evolu-
tion when it is remembered that biennis nanella is frequently
weakly or diseased, that biennis sémigigas is self sterile, and that
biennis sulfurea appears to be a retrogressive form having lost the
power of producing normal yellow flowers. Although the Dutch
biennis of all the cenotheras so far brought into the experimental
garden still seems to me the form most free from suspicion of
genetic impurity, nevertheless, the line of Stomps’s has not, so
far as we know, been subjected to all of the tests of a pure species.
Until these tests are made it is not safe to assume that this mate-
rial is wholly pure. It seems to me not improbable that other
species of @nothera will eventually be isolated more stable than
the Dutch biennis. BRADLEY Moore Davis

UNIVERSITY OF PENNSYLVANIA,

June, 1915

6 De Vries, Hugo, ‘‘The Coefficient of Mutation in @nothera biennis L.,’?

Bot. Gaz., Vol. LIX, p. 169, 1915.

No.587] SHORTER ARTICLES AND DISCUSSION 707

THE VALUE OF INTER-ANNUAL CORRELATIONS

IF £1, £a, Lz +++ Zn be measures taken on the n individuals of a
series in a given year and 2,', £3, £, +++ Zn’ be similar measures
taken in a subsequent year, the correlation between the first and
second measures on the same individual rss’, may be designated
as a direct inter-annual correlation.1 The purpose of this review
is to illustrate the usefulness of such constants, with a view to
extending their application, by bringing together examples of
inter-annual correlations from various fields.

The immediate value of such coefficients may be purely scien-
tific, economic, or both theoretical and practical.

Practically such means of prediction as correlation and regres-
sion formule should find wide application in breeding operations
where it is desirable to weed out or send to the butcher at the
earliest possible moment those individuals which can not be kept
with the maximum profit. If the correlation between the egg
production of a fowl in her pullet year and her laying capacity
in any subsequent year be high, it is clear that those which on the
average are to prove unprofitable may be sent to the pot when
most desirable for that purpose, and before they have consumed
two or more years’ feed without yielding the maximum return in
eggs. If, on the contrary, there be no correlation, the labor of
selection in the pullet year is an unnecessary expense. Ifa cow’s
milking capacity be closely correlated with her milking record in
her heifer year, the culling of dairy herds may be profitably
carried out in the first year. In plant breeding experiments,
involving either sexual or vegetative reproduction, selection of
individuals for future propagation must be made, and at as early
a date as possible. If the future yield per plant of hay can be
estimated with considerable accuracy from a first year’s culture
the process of selecting clonal strains can be carried out with far
greater rapidity than if one must wait for the results of subse-
quent years’ tests. In all such cases the finality of a first judg-
ment must depend in large degree upon the closeness of correla-
tion between the results of successive experiments—in short upon
the value of the inter-annual correlation coefficient.

1 Cross inter-annual correlations in which the measures taken are of a
different sort are sometimes useful, but examples of such are not consid-

view,

708 THE AMERICAN NATURALIST [ Von. XLIX

In dealing with egg production Pearl and Surface? give as the
correlation in number of eggs for first and second year

r= .032 + .083,

a value which, though positive, is clearly insignificant with re-
gard to its probable error.

Thus in this particular case the performance of the first year
furnishes no clue to that of the second. With respect to egg-
laying capacity, the record of the pullet year furnishes no cri-
terion for elimination from the flock.

For milk yield in cattle the case seems to be quite different.
Gavin? has found that there is a medium correlation between (a)
the ‘‘ revised maximum ’” yield in quarts of successive lactations,
and (b) between the ‘‘ revised maximum ’’ of the individual
lactation periods and the highest revised maximum reached by
the animal.

So far as I am aware the only worker who has published
correlations between the characters of the same plant individuals
in different years is Clark® whose results have been noted in these
pages by Pearl.®

The correlation tables and constants show that plants of a
given class in any year (height or weight of hay produced) will
be highly variable in a subsequent year, but will on the average
deviate from the mean of the whole culture of the year in the
same direction and to about half the extent of the type selected
in the preceding year. Thus if selection were made on the basis
of a single year’s test only, many individual plants of low yield
would be discarded which in a subsequent year would have taken
higher rank, while high-yielding plants would be retained which
subsequently would give disappointing results. On the whole,
however, the yield of a hay plant one year does furnish a valu-
able index to its yield in a subsequent year.

2 Pearl, R. and F. M. Surface, ‘‘A Biometrical Study of Egg Produc-
tion in the Domestic Fowl,’’ I, Bull. Bu. Anim. Ind., 110, 66, 1909.

3 Gavin, Wm., ‘‘Studies in Milk Records: On the Accuracy of Estimat-
ing a Cow’s Milking Capacity by Her First Lactation Period.’’ Jour. Agr.
Sci., 5, 377-390, 1913.

4‘‘ Revised Maximum’’ milk yield is «a maximum day yield which is
three times reached or exceeded in a lactat

5 Clark, C. F., ‘‘ Variation and paca i in aS ? Bull. Cornell Agr.
Exp, Sta, 279, 1910.

€ Posti, R., Aune. Nar., 45, 418-419, 1911.

No.587] SHORTER ARTICLES AND DISCUSSION 709

That we are dealing with a real measure of the relatively per-
manent differentiation of individuals, and not with merely tem-
porary differences due to growth, is indicated by the fact that
the correlations between a first and a third year are about the
same as those between a first and a second or a second and a
third.

In other fields of plant industry such methods may be profit-
ably applied. For example Sievers’ after discussing at some
length the question of the differentiation of belladonna plants
with respect to alkaloidal content, warns the reader that ‘‘ the
investigation has hardly progressed far enough to yield any defi-
nite conclusions ’’ but says in summarizing his data:

A considerable number of plants with leaves rich in alkaloids in one
season are found to have equally rich leaves in the following season.
Furthermore, they frequently manifest the same characteristics at the
various stages of growth during the season in eomparison with other
plants. The same facts are true with regard to plants which bear
leaves with a low percentage of alkaloids.

How much more definite is the information conveyed by the
simple statement that the inter-annual correlation! between the
alkaloidal content for 1911 and 1912 is

r==.513 + .066!

Such studies as those by Stockberger on individual perform-
_ ance in hops? may be facilitated by the use of inter-annual corre-
lation coefficients. He gives only the extremes of his series of
individuals, but from these the correlations between yield per
hill for different years are:

Lowest Hills Highest Hills
1909 and 1910 ..:...... 29 + 17 59 + .18 =
1910 and 390)... a. 55 + 13 52 + 14
1009 tnd WIL... 43 + 15 30 + .18

Such constants, deduced from materials which almost certainly

T Sievers, A. F., ‘‘ Individual Variation in the Alkaloidal Content of Bel-
ladonna dae! Jour. Agr. Res., 1, 129-146, 1913.

8 In computing this coefficient a number of inconsistencies in the data
table were issand. The constant as given is probably as nearly correct
as can be found from the available da

9 Stockberger, W. W., ‘‘A Study of Individual Performance in Hops,’’
Prac. Amer. Breed. Ass., 7, 452-457, 1912.

710 THE AMERICAN NATURALIST [Vou. XLIX

do not show the full strength of the correlation, remove at once
all question concerning the relatively permanent differences in
productiveness of the individual hills.

Consider next an illustration from hybridization of measur-
able characters.

Goodspeed and Clauson’ have given the mean values of meas-
urements of the flowers of individual plants of Nicotiana hybrids
cultivated in 1912 and of corollas of the same plants cut back
and flowered in 1913. The correlations between the mean di-
mensions for the two years I find to be:

N. Tabacum var. macrophylia 2 X N. sylvestris g
F, plants, N = 21.
For spread of corolla, r = .044 + .147.
For length of corolla, r==.169 + .143.
Hybrid produced by crossing F, of the hybrid N. Tabacum
** Maryland ” 2 by N. Tabacum 4, with N. sylvestris, N=19.
For spread of corolla, r==.560 + .106
For length of corolla, r==.788 + .059.

These correlations show at once the high degree of uniformity
of the F, of the first as compared with that of the second series.
In all four cases the signs of the coefficients are positive, but
those of the first class are insignificant in comparison with their
probable errors. In both cases length of corolla is more closely
correlated than breadth. Possibly this is due to errors of sam-
pling only, or to greater difficulty in obtaining an exact measure .
of the spread of the limb. It may, however, indicate that some
characters are more sharply and permanently differentiated
from individual to individual than others.

That the latter may sometimes be the case is clearly shown by
unpublished data of my own for the ligneous perennials Staph-
ylea trifolia and Hibiscus Syriacus.*?

10 Goodspeed, J. H., and R. E. Clauson, ‘* Factors Influencing Flower Size
in Nicotiana with Spøeial Reference to Qisim of Inheritance,’’ Amer.
Jour. Bot., 2, 232-274, 1915

11 The AREA are based in all cases on mean values of the characters
of ovaries of shrubs well established in the Missouri Botanical Garden. In
such work the number of individuals can never for practical reasons be very
large, if a fairly large number of countings be made for each shrub. Fur-
thermore much of the work which one does may be lost by some accident
which precludes the securing of countings from each individual every year.
. If an individual is not represented in both of a pair of years it must be
omitted entirely.

No. 587]

SHORTER ARTICLES AND DISCUSSION

711

The accompanying tables show the correlations deduced for

the characters indicated.'”

INTER-ANNUAL CORRELATIONS FOR FRUITS OF STAPHYLEA

Correlation for | Correlation for pa ergo for

Relationship 1906 an sa pen Be 1007's ae 1908, 1908 1909,

tes ong srg ovules .445 = .069 | .816 = .058 | 872 = 036

Seeds seeds .063 = .154 | .064+.173 | .056 + .150

Ravan d asymm 748 + .068 | .102.+ .172 | .205 + .145
Libsilacs composition ae y lens A com-

TOON on ee a a ahs ss .601 + .099 | -294 = .159 | 335 = .134

INTER-ANNUAL CORRELATION FoR Fruits or HIBiscus, n= 23.

Correlation for

Relationship 1907 and 1908
PUPS BU OIE os. co Carra aa ie is eke es 451 + .112
Brors and PINE oo ee a e a .836 + .042
yale And OO aoc e ea et eee aa e 941 + .016
Goods and PoE ee cia 630 + .085
Asymmetry ANG IMMO -e ok ic Fi ayes 747 + .062
Locular composition and locular composition ........ 725 + .067
POP BOG LOr oe 6 seas beet bsuctveteas yey 610 + .088
COTTE EON BNO SONORON >.: iruek koraan .035 2 141

The constants are very irregular in magnitude, but are with-
out exception positive in sign. In many instances they are large.
Thus in these individual shrubs which taxonomically show no
differences® there is nevertheless a distinct differentiation in
respect of the great majority of the characters examined.

While the probable errors are large the evidence warrants the
conclusion that some are decidedly more highly correlated than
others.

12 Sepals = mean number of sepals in calyx.

Bracts = mean number of bracts in involucre.

Ovules = mean number of ovules formed per fruit.

Seeds — mean number of seeds matured per fruit.

Asymmetry = average radia] asymmetry in the distribution of the num-
ber of ovules per locule. For method of computation see Biometrika, Vol.
VII, pp. 477-478, 1910, and AMER. Nat., Vol. XLVI, p. 480,

Locular composition = average number of locules per frait with an odd
number of ovules. See citations above.

Fertility — coefficient of fertility (mean seeds per fruit) (mean ovules
per fruit).

Correlation = sigma of correlation between number of ovules and
number of seeds per loc

13 I believe one of the Hibiscus shrubs had lighter flowers than the rest.

712 THE AMERICAN NATURALIST [ Von. XLIX

In Hibiscus the differentiation of the individuals with respect
to number of bracts seems to be greater than that for number of
sepals. For both Staphylea and Hibiscus the correlation for
ovules is generally high. It is in every instance higher than
that for mean number of seeds matured per fruit. Correlation
for both mean number’ of seeds per fruit and relative number
of seeds matured has a moderately large value in Hibiscus, but
in Staphylea it is sensibly 0. In both species such character-
istics of the ovary as radial asymmetry and locular composition
seem to be rather sharply differentiated from individual to indi-
vidual. This is probably due in part to differentiation with
respect of number of ovules per fruit, but further discussion of
the problem would be out of place in a note, the only purpose
of which is to call attention to the usefulness, in both applied
and pure science, of a quantitative means of detecting and ex-
pressing permanent differentiation.

In this brief review I have made no attempt to discuss fully
all the biological phases of the problems suggested. The analysis
of the data may in several instances be carried much further by
the use of the statistical tools. Perhaps enough has been said to
indicate that inter-annual coefficients may be of real service in
practical animal husbandry, in plant breeding and in mor-
phology and physiology. More than usefulness is not to be ex-
pected of any method.

J. ARTHUR Harris

THE PHENOMENON OF SELF STERILITY

In my paper which appeared in THE AMERICAN NATURALIST,
Vol. XLIX, p. 79, the last seven lines on page seventy-nine
should read as follows:

Self-sterile plants crossed with self-sterile plants gave only self-sterile
offspring. Certain self-fertile plants, however, gave only self-fertile off-
spring either when self-pollinated or when crossed with self-sterile plants.
Other self-fertile plants gave ratios of 3 self-fertile to 1 self-sterile off-
spring when self-pollinated, and ratios of 1:1 when crossed with pollen
from self-sterile, ete.

E. M. East.

VOL. XLIX, NO. 588 “ DECEMBER, 1915

SHe.

As

. The F, Blend accompanied by Genic Purity. Dr. H. H. LAUGHLIN

THE
AMERICAN
NATURALIST

A MONTHLY JOURNAL
Devoted to the Advancement of the Biological Sciences with
Special Reference to the Factors of Evolution

CONTENTS
Page
Some Experiments in Mass Selection. Professor W. E. CASTLE - - -713
The Inheritance of Black-eyed White Spotting in Mice. Dr. C. C. LITTLE - 727
- ~ 741

am

The ne TREC of the “Blanket Te of Freshwater Pools. EX perem
PLA 752

Shorter pee and Discussion: Practical Vitalism. Dr. A. GURWITSCH - 763
Index to Volume XLIX = - - = ~ - = 771

THE SOIENCE PRESS
LANCASTER, PA. GARRISON, N. Y.
NEW YORK: SUB-STATION 84

The American

SS intended for erie and books, ete.

M
sent to the Editor of THE AME
Short articles containing summ

Naturalist

., intended for review should be

RICAN rata RALIST, Garrison-on- Hudson, n 1

es of research work be the

problems of organic evolution are sapotialty welcome, oud will be given prefereens

in publicatio
Gn

e
Further thorn will be supplie
Subscri

anadian postage twenty-five cents additional.

ptions and advartiscun nie oe her sent to the A say yie

a reprints of gor on a are supplied to authors free of charge.

The
postage is fifty cents and

eign
The b arge for single copies is

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

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered as second-class matter, April 2, 1908, at the P
c

Garrison, N, Y.

ost Office at Lancaster, Pa., under the Act ot

ongress of March 3, 1879.

FOR SAL
ARCTIC, ICELAND and GREENLAND
RDS’ SKINS
Well Farr Low Prices
articulars of
G. DIN NESEN. Bird Collecto
Husavik, North Iceland, Via Leidle, Eneli

JAPAN NATURAL gea RAITAR

Perfect Condition and Lowest Pri
Specialty: Bird Skins, Oology, NEAR Marine
nimals ee others. Catalogue free. Correspond-
ence solici
T. FUKAI, Naturalis
Konosu, pe Japan

BOOKS FOR SALE

relating to all branches of

NATURAL HISTORY

Catalogue on application

JOHN D. SHERMAN, Jr.

Avenue MOUNT VERNON, N. Y.
The University of Chicago]
S instruction during the ~
mer Quarter on the same basis a

during the other quarters of ie |
emic

The undergraduate coll leges, the
5 a

|
|

Marine Biological Laboratory
Woods Hole, Mass.

widen Laem se —— bsg Paystiocy” and Bor

r th
of 1nembers of the staff, The
for such a table is $50.00.

Courses of laboratory instruction
with lectures are offered in Inverte-

INSTRUCTION
July—August

Eco!
ago eam Each course req
ef e of the student. Fee.
$50.00. se lecture poage 7 Bio the
o
Agr ama Am off “tem §

SUP
DEPARTMENT
Open the Entire Year

otanical
gent on application. State ate
is desi =- For price ne and all
inform regarding material,
bie og
GEO. M. GRAY, Curator, Woods Hole, Mass

The annual announcement will be sent on -a to
Director, Laboratory, Woods Hole

THE
AMERICAN NATURALIST

VoL. XLIX. December, 1915 No. 588

SOME EXPERIMENTS IN MASS SELECTION

PROFESSOR W. E. CASTLE
Bussey Institution, HARVARD UNIVERSITY

Ar the close of an interesting review of ‘‘seventeen
years selection’’ of the character winter egg production
in Barred Plymouth Rock fowls, made at the Maine Agri-
cultural Experiment Station,’ Dr. Pearl compares his re-
sults with those of Phillips and myself? in selecting for a
like number of generations the hooded pattern of rats and
concludes that the same interpretation should be given to
both series of experiments, viz., that selection can change
a population but not a character.

Without discussing for the moment the validity of the
now world-famous generalization of Johannsen, which
Pearl here accepts for his fowls and seeks to extend to
our rats, I wish to point out some differences between the
two cases which make a direct comparison between them
difficult and conclusions based upon them of unequal
validity.

The character winter egg production in fowls is on
Pearl’s showing extremely difficult to determine. It is
necessarily an unknown quantity in all male birds, which
themselves produce no eggs, and any influence which

1‘‘ Seventeen Years Selection of a Character Showing Sex-linked Mendel-
ian Inheritance,’’ AMERICAN NATURALIST, Vol. 49, pp. 595-608, 1915.

2**Piebald Rats and Selection,’’ Publ. No. 195, Carnegie Institution of
Washington, 1914,

713

714 THE AMERICAN NATURALIST [Von XLIX

males may exert on the egg-production of their daughters
can be tested only by an indirect and rather uncertain
process. Only in the case of females is the character di-
rectly measurable and then only for such females as (1)
are hatched ‘‘after April 1 and before June 1,’’ (2) sur-
vive all the accidents of chickhood and adolescence, (3)
escape all attacks of disease and are kept continuously
free from parasites, and (4) are properly fed and housed.
For any bird which dies, is disabled or becomes seriously
ill under ten months old, the character is an unknown
quantity. These limitations make the proportion of
birds which can be accurately rated as regards the char-
acter extremely small, and reduce correspondingly the
material on which selection can be practised.

Contrast with this situation that regarding the hooded
pattern of rats. This character is possessed by every in-
dividual of both sexes and is inherited equally through
either sex. The character is fully developed in its final
form within a week after birth, months before sexual
maturity is attained. This makes it possible to grade the
animals accurately while they are still very young and to
discard at once all individuals which fall below the
adopted standard. Selection thus has a vastly greater
amount of material to work with, and the variation in
each generation can be ascertained with a completeness
and accuracy quite impossible in the case of winter egg
production in fowls.

It is scarcely necessary to point out that upon the com-
pleteness of one’s knowledge of the character and extent
of variation depends his ability to take advantage of that
variation by systematic selection. By this criterion win-
ter egg production is very poor material on which to base
an experimental test of ‘‘mass selection,” whereas the
hooded pattern of rats is material admirably adapted for
the purpose. Many times has the fact been commented
upon that Mendel’s fortunate choice of peas as material
for his studies of hybridization was largely responsible
for his success where others failed. If one wishes to test

No. 588] EXPERIMENTS IN MASS SELECTION 715

a theory he must choose material suited to the purpose.
No adequate test of the efficacy of mass selection can be
obtained from material which can not be accurately
judged in the mass.

Pearl points out further limitations of his material in
the statement ‘‘that phenotypic variation of the character
fecundity, in fowls, markedly transcends, in extent and
degree, genotypic variation.” That is, non-heritable
causes of fecundity are in excess of heritable causes and
serve to obscure the occurrence of the latter. Further,
Pearl says:

It is quite impossible in the great majority of cases to determine with

precision what is a hen’s genetic constitution with respect to fecundity
from an examination of her egg record alone.
If then one has reared his pullets to the age of one year,
has kept them free from disease and parasites, has fed
and housed them properly and has even trap-nested them
and recorded their eggs all winter, still he has no suff-
cient basis on which to base a selection. He must first
rear and test their progeny in the same way. Pearl’s
statements on this point, the accuracy of which I do not
question, are sufficient to show the entire unsuitability of
his material for testing the efficacy of mass selection.

One might with propriety even question whether such
a thing as inherited capacity for winter egg production
exists in fowls, but on this point, I think, another inves-
tigation? made by Pear] is conclusive, in which he crossed
Cornish Indian game fowls, which are poor winter layers,
with Barred Plymouth Rocks which are fairly good winter
layers. Reciprocal crosses were made in both of which
the daughters showed resemblance to the racial winter
egg productiveness of the sire’s race. This result indicates
that a sex-linked genetic factor of some sort exists which
affects winter egg production in fowls. But since the fe-
cundity of the offspring was obviously influenced by the
mothers’ race as well as by the father’s race, Pearl was

3‘‘The Mode of Inheritance of Fecundity in the Domestic Fowl,’’ Jour.
Exp. Zool., Vol. 13, p. 153, 1912.

716 | THE AMERICAN NATURALIST [ Vou. XLIX

led to suggest the existence of a second fecundity factor
which was not sex-linked. He assumes that this second
factor, like the first, is a Mendelizing factor, but without
any sufficient published evidence for either conclusion.
To this I called Dr. Pearl’s attention soon after the pub-
lication of his paper and suggested that if possible the
data be put on record in such form as to allow of testing
this and other hypotheses concerning the genetic factors
concerned. For one-factor, two-factor, ten-factor and
infinity-factor Mendelian hypotheses would call for very
different ratios and distributions of fecundity among the
offspring. He replied that the data could not be so given
without an amount of work which he considered unprofit-
able. We are left, therefore, with only this information
concerning Pearl’s pullets, whether each one laid more or
less than 30 eggs in its first winter. If we knew what
number each one laid, we might form an intelligent
opinion as to whether Mendelian factors are involved,
and if so how many, in the same way that we can test
Mendel’s conclusions concerning the independent inheri-
tance of yellow cotyledon color and round seed form in
peas because he tells us the actual proportions of the
various sorts of peas reported for each plant. Being
denied such information by Pearl, it is useless to dis-
cuss his two-factor hypothesis, for its correctness can be
neither proved nor disproved.

Leaving aside the question whether any inherited factor
has changed as a result of selection in Pearl’s experi-
‘ments, which we have no means of investigating, we can
consider only the question whether the gross winter egg
production has changed. As a basis for judgment he
gives us the averages of winter egg production year by
year for sixteen years. Pearl’s graphic presentation of
the data (assuming that the considerable fluctuation re-
corded is not significant) indicates a steady decline of
the general flock average during the first nine years of
the experiment and a steady recovery and further in-
crease during the next seven years, which he ascribes to

No. 588] EXPERIMENTS IN MASS SELECTION 717

the different basis of selection in the two periods. But
it is hard to believe that this entirely explains the dif-
ference in result. One notices for example that during
the period of ostensible decline the highest average fecun-
dity (45.23) is recorded when the number of birds under
observation is smallest (48) and the lowest average
(19.93) is recorded when the flock is largest (780). Fur-
ther, in the later seven-year period of ‘‘improvement,’’
the number of birds tested declines as their average
fecundity rises. Has not the better environment and
lessened competition of small numbers possibly some-
thing to do with the changes noted? Is it certain that
genetic agencies are responsible for the differences ob-
served? Pearl himself nowhere states that the selection
practised during the earlier period had produced posi-
tive deterioration; he merely states that ‘‘there was no
change of the mean in the directon of the selection’’ dur-
ing this period when selection was based on high produc-
tion without progeny tests. But as soon as progeny tests
were made an additional feature of the basis for selection
Pearl notes immediate results, viz., the immediate isola-
tion of a strain which in its first year made a record for
high productiveness only once equalled in the six subse-
quent years. How many successive selections were made
in this period, we are not informed, but since it would
require at least two years to make a combined perform-
ance and progeny test, it would seem that not more than
three successive selections can have been*carried out on
this basis in the seven year period from 1908 to 1915. It
may fairly be questioned whether this is an adequate test
of the effectiveness of mass selection. The total number
of individuals tested during this period is, according to
Pearl’s table, 1,655. For the entire seventeen years of
selection it is 4,842.

The total number of animals graded in our selection
experiments with rats heretofore published is 20,645, and
the number of generations involved 13. Since those
figures were compiled, four additional generations of

718 THE AMERICAN NATURALIST [ Vou. XLIX

rats have been reared in the straight selection series,
bringing the total number of animals observed in this ex-
periment up to 33,249, and the total number of genera-
tions of selections up to 17, numbers certainly more nearly
justifying the term ‘‘mass selection’’ than those studied
by Pearl. As no previous account of this experiment has
been given to readers of the Naruratist, a brief review
of its salient features may be appropriate here.

Experiments made by MacCurdy and by Doncaster
had shown that the hooded pattern of rats is a Mendelian
recessive character dominated in crosses by the ‘‘self’’
or entirely pigmented condition of wild rats and of cer-
tain tame races. The F, ratio obtained in crosses be-
tween hooded and self rats is an unmistakable mono-
hybrid ratio, viz., 493 hooded: 1,483 self, or 24.9 per cent.
hooded. The hooded pattern is subject to slight fluctua-
tions in the relative amounts of pigmented and unpig-
mented surfaces, and though these slight plus and minus
variations are such as are usually disregarded in Men-
delian analyses, MacCurdy’s investigations had indicated
that they are to some extent inherited. It was our pur-
pose in starting the selection experiments to ascertain
whether the observed fluctuations were capable of in-
crease and summation through the action of repeated
selection, a possibility denied for all such cases by de
Vries and Johannsen on theoretical grounds and quite
incompatible with notions prevailing then as to the
‘‘oametic purity’’ of recessives. This ‘‘pure line” idea
Pearl still maintains on the basis of his observations of
the winter productiveness of his pullets. But, as I have
tried to show, his material is no more adequate than that
of Johannsen, which involved no demonstrated Mendelian
character whatever. For, though Pearl asswmes that
winter egg productiveness of fowls involves a ‘‘sex-
linked Mendelian character’’ he has withheld from pub-
lication the only facts on which such an assumption may
legitimately be based.

Our selection experiments with hooded rats began in

No. 588] EXPERIMENTS IN MASS SELECTION 719

1907. The initial stock consisted of less than a dozen indi-
viduals all ‘‘pure recessives,’’ which produced only ‘‘

cessive’? hooded young, in accordance with Mendelian
expectation. But though all the young were recessive
(hooded), all were not exactly alike, and to assist in their
classification we devised arbitrary ‘‘grades’’ of increased
(plus) or decreased (minus) pigmentation as compared
with the modal (zero) condition in our hooded race. The
scale of ‘‘grades’’ is shown in part in Fig. 1. It has

$

+
AEMT set of grades used in classifying the fluctuating variations

uke
of oa

been found necessary to extend it in both directions,
beyond the range shown in the figure, in order to
admit the new grades of rats which have made their ap-
pearance as the experiment progressed. The first plus-
selected parents produced 150 offspring ranging in grade
from + 1 to +3, mean + 2.51. The first minus-selected
parents produced 55 offspring ranging in grade from — 2
to + 4, mean —1.46. It will be observed that the ranges
of the young produced in the two selections were prac-
tically continuous with each other, though they did not
actually overlap. But actual overlapping did occur in the
following generation, in which no advance was made in
the mean grade of the parents, practically all the available
females being used as parents in an effort to increase the
stock. The grade of the offspring also remained prac-
tically stationary in this second generation (see Tables I

720

THE AMERICAN NATURALIST

TABLE I

[Vou. XLIX

RESULTS OF THE PLUS SELECTION OF HOooDED RATS CONTINUED THROUGH

SIXTEEN SUCCESSIVE GENERATIONS

Lowest Highest Standard Number of
Mean Grade | Mean Grade
vor ranai | of Oxbetns | Gorn | Gene | oan St | Gieli
1 2°51 2.05 +1.00 +3.00 .54 150
2 2.52 1.92 — 1.00 +3.75 to 471
3 9.78 2.51 + .75 +4.00 Oo 341
4 3.09 2.78 + .75 +3.75 47 444
5 3.33 2.90 + .75 +4.25 .50 610
6 3.52 a1 +1.50 +4.50 49 861
T 3.56 3.20 +1.50 +4.75 .55 1,077
8 3.75 3.48 +1.75 +4.50 44 1,408
9 3.78 3.54 +1.75 -4.50 "30 1,322
10 3.88 3.73 +2.25 +5.00 .36 776
11 3.98 3.78 +2.75 +5.00 -29 697
12 4.10 3.92 +2.25 +5.25 soL 682
13 4.13 3.94 +2.75 +5.25 -34 529
14 4.14 4.01 +2.75 +5.50 -34 1,359
15 4.38 4.07 +2.50 +5.50 .29 3,690
16 4,45 4.13 +3.25 +5.87 29 1,690
16,107
TABLE II
RESULTS OF THE MINUS SELECTION OF HoopED RATS CONTINUED THROUGH
VENTEEN SUCCESSIVE GENERATIONS
Mean Grade | Mean Grade — Biphost ere Number of
See | ot Parents | ct Oftepeton Oa pst ge Posed Offspring
1 — 1.46 — 1.00 + .25 — 2.00 51 55
2 — 1.41 — 1.07 + .50 — 2.00 AQ 132
os —1. —1.18 0 —2.00 48 195
4 —1.69 —1.28 + .50 22o 46 329
5 —1.73 —1.41 0 — 2.50 50 701
6 — 1.86 — 1.56 0 — 2.50 44 1,252
rd —2.01 — 1.73 0 — 2.75 .35 1,680
8 — 2.05 — 1.80 0 Zio .28 1,726
9 —2.11 —1.92 — .50 —2.75 28 1,591
10 — 2.18 — 2.01 — 1.00 3:25 -24 1,451
11 — 2.30 — 2.15 — 1.00 — 3.50 .35 984
12 — 2.44 2.23 — 1.00 —3.50 ot 1,037
13 —2.48 — 2.39 —1.75 —3.50 .34 1,006
14 — 2.64 —2.48 — 1.00 3.50 .30 Viz
15 — 2.65 — 2.54 —1.75 —3.50 .29 1,438
16 — 2,79 — 2.63 —1.00 — 4.00 ae 1,980
17 —2.86 — 2.70 —1.75 —4,25 -28 868

No. 588] EXPERIMENTS IN MASS SELECTION 721

and IT). In the third and all subsequent generations selec-
tion was made as rigorous as possible consistent with the
maintenance of a strong colony from which to make fur-
ther selections. Following each selection an advance in
the average grade of the offspring took place attended by
a steady movement in the direction of the selection on the
part of both the upper and the lower limits of variation.
The sixteenth plus selection produced 1,690 offspring (a
larger number of individuals than is contained in Pearl’s
entire seven-year series) every one of which fell beyond
the original range of variation, which was from + 1 to + 3
in the first plus selected generation and from + 34 to +5
in the sixteenth generation. What this change signifies
will be better appreciated when I state that +6 in our
grades is a wholly pigmented or ‘‘self’’ rat, and that the
extreme variation noted, + 53, signifies a rat wholly pig-
mented except for a few white hairs between the front
legs. The whole race has accordingly been changed so
that no individual is longer produced which falls within
the original range of variation. Not a dozen rats in this
entire generation would be allowed by a fancier in the
category of ‘‘hooded”’’ rats.

In the minus selection series the results secured are
scarcely less striking. Only a very few individuals of the
1,980 sixteenth generation rats, or the 868 seventeenth gen-
eration rats fell within the original range of variation,
which in generations 1-3 went no farther than grade — 2.
In all other individuals of the sixteenth and seventeenth
generations the ‘‘hood’’ was reduced to an extent never
seen in the hooded rats of the fancier, the white areas
having covered the neck and in extreme cases the fore-
head also, leaving only the nose and a patch round the
eyes and ears still pigmented.

Pearl (p. 607) commenting on the results of his selec-
tions states that he had no reason to think that at the close
of the series any individual had been produced superior
in productiveness to those which occurred at the outset,
but that he had merely secured more of them, thus raising

T2: THE AMERICAN NATURALIST [ Vou. XLIX

the average. With the rats, however, a very different
condition exists. The average is not changed by increase
of high-grade individuals merely or chiefly. At the pres-
ent time every individual in the plus selection series and
nearly every individual in the minus selection series is of
higher grade (plus or minus respectively) than any indi-
vidual in the race at the outset. It is not a fallacious
change of averages which has taken place; a genuine and
permanent racial change has occurred, following step by
step upon repeated selection. Generation by generation
new grades of offspring have come into existence, more
extreme in character than any which existed before, and
simultaneously with the advance of the outer limit of vari-
ation the inner limit has receded. No great change in
variability has attended the selection. The standard devi-
ation has decreased somewhat to about three fifths of its
original amount, but has scarcely altered in the last eight
or ten generations (see Tables I and IT). Rather there has
occurred a change in the modal condition of the character,
about which fluctuation continues very much as before.
When the position of the mode changes, as a result of
selection, the position of the average and of the upper and
lower limits of variation change with it. In a word the
character changes.

In our 1914 publication Phillips and I were conservative
about asserting a change in the single Mendelian unit-
character manifestly involved in the hooded pattern. We
suggested the possibility that other as yet undiscovered
factors might be responsible for the apparent changes
observed and awaited the result of experiments then in
progress to show whether such a possibility was admis-
sible. I have no hesitation now in saying that it is not.
All the evidence we have thus far obtained indicates that
outside modifiers will not account for the changes ob-
served in the hooded pattern, itself a clear Mendelian
unit. We are forced to conclude that this unit itself
changes under repeated selection in the direction of the
selection; sometimes abruptly, as in the case of our ‘‘mu-

No. 588] EXPERIMENTS IN MASS SELECTION 723

tant’’ race, a highly stable plus variation ; but much oftener
gradually, as has occurred continuously in both the plus
and the minus selection series. The permanency of these
cumulative changes we have tested by repeated crossing
of both selected races with the same wild race. The first
cross seems to undo to a slight extent the work of selection,
causing regression in both plus and minus selected races,
but a second back cross with the wild race causes no fur-
ther regression. Thus, plus-selected rats of mean grade
3.45 were crossed with wild rats and the recessive char-
acter was recovered in F, in 75 individuals, 24 per cent.
of the entire generation. These 75 extracted hooded rats
were of mean grade 2.89, a regression of .56 on the mean
grade of their hooded grandparents, which is about double
the regression shown by the plus selected race when not
crossed with wild rats. It seems proper therefore to at-
tribute to the wild cross a part of the regression observed
in this case and this I have expressed by saying that cross-
ing the selected race with wild rats tends to undo the work
of selection. The suggestion was tentatively adopted by
Phillips and myself that this wndoing consisted in the re-
moval of ‘‘modifiers’’ of some sort, possibly independent
Mendelizing factors. If this explanation were correct,
further crossing with wild rats should tend still further
to ‘‘undo’’ the work of selection, so that ultimately the
extracted hooded race should return completely to its orig-
inal modal state, the zero grade. To test this matter,
extracted hooded rats ranging from grade +2 to +4
(mean grade 3.01) were crossed back a second time with
pure wild rats. The theory of independent modifiers
would lead one to expect further regression as a result of
this cross, but no regression was this time observed. In-
stead an advance of .32 took place bringing the mean of
the twice extracted hooded recessives back to about the
grade of the uncrossed race. The mean grade of the once-
extracted grandparents, loaded in proportion to the num-
ber of their twice-extracted hooded grandchildren, was
3.01; the mean of the 263 hooded grandchildren was 3.33.

724 THE AMERICAN NATURALIST [ Vou. XLIX

The number of these grandchildren is large enough to
leave no doubt as to the conclusion that no further regres-
sion attended extraction of the hooded character a second
time from the wild cross. The proportion of hooded in-
dividuals to non-hooded is also an unmistakable mono-
hybrid ratio, viz., 263 hooded to 759 non-hooded, or 25.7
per cent. hooded in a total of 1,022 individuals.

This result indicates clearly the untenable character of
our provisional hypothesis to explain the altered grade of
hooded rats under selection and crossing, by invoking the
action of independent modifying Mendelian factors. No
evidence is forthcoming from further and more extensive
experiments that such modifying factors are concerned in
the result. It seems rather that the hooded character,
which is a mosaic or balanced condition of pigmented and
unpigmented areas, is slightly unstable. It oscillates reg-
ularly about a mean condition or grade, these oscillations
being not phenotypic merely but in part genotypic so that
selection brought to bear upon them is immediately and
continuously effective.

There may exist cases of continuous variation purely
phenotypic, as that of Johannsen’s beans seems on his
showing to be. In other cases phenotypic variations may
so largely exceed genotypic variations that it is difficult to
discover and isolate the latter, as has been Pearl’s ex-
perience. But our experiments with rats show beyond
reasonable doubt that genotypic variation, as well as phe-
notypic, may assume a continuous form, and if it does no
one can question its further modifiability by selection. In
denying effectiveness to selection in the case of continuous
variation, it has been assumed, tacitly by DeVries and
expressly by Johannsen, that continuous variation is
wholly phenotypic. This assumption being disproved,
the pure-line theory which rests upon it lacks adequate
support.

It seems strange looking backward that the idea should
have become so widely accepted that continuous or fluctu-
ating variations are wholly phenotypic. For a continu-

No. 588] EXPERIMENTS IN MASS SELECTION 725

ous variation signifies only the combined result of several
independent agencies. In purely phenotypic variation
(such as possibly Johannsen has observed) these agencies
are obviously environmental and so do not affect the in-
heritance. But in a case of multiple genetic agencies (the
existence of which everyone recognizes) a continuous
series of variations may result which would be amenable
to selection. Pearl and all other pure-line advocates ad-
mit the existence of such cases. But the same thing would
result if, aside from purely phenotypic variations in a
character, its single factorial basis should undergo quanti-
tative variation. It is precisely this last named category
of cases which alone can explain our rat results. And it
is precisely this category of cases which the pure-line ad-
vocates, unable to disprove, boldly deny. Driven from all
other defences they cling to.this as their last line and
solemnly repeat challenges issued years before in mo-
ments of greater confidence. Thus Pearl closes his paper
with a renewal of the opinion expressed by him in 1912.

It has never yet been demonstrated, so far as I know, that the abso-
lute somatic value of a particular hereditary factor or determinant (i. e.,
its power to cause a quantitatively definite degree of somatic develop-
ment of a character) can be changed by selection on a somatic basis,
however long continued.

Our observations on rats are submitted as a sufficient
answer to this challenge.

I do not suppose that Pearl means to be taken seriously
when he says (p. 608):

The extreme selectionist appears to believe that in some mysterious
way the act of continued selection, which means concretely only the
transference of each selected individual from one cage or pen to another
to breed, will in and of itself change the germ-plasm.

I have never heard a selectionist, however extreme, ex-
press such a view; certainly I, whose views are attacked
in the next sentence, have never entertained such an idea.
But Dr. Pearl knows, as well as I do, that while the germ-
plasm of the individual remains unmodified upon its trans-

726 THE AMERICAN NATURALIST [Vou. XLIX

fer from one cage to another, the character of the germ-
plasm of its descendants, and so of the race, depends very
largely upon what mates are transferred to the same cage
with it. This is where the selection comes in and there is
nothing ‘‘mysterious’’ about it either.

The idea that selection can bring about no change in the
germ-plasm of the race ‘‘except by sorting over what is
already there,’’ to which Pearl gives expression, rests on
the assumption that the germ-plasm never changes. What
ground have we for such an assumption? No more than
for the idea of the unchangableness of species, which for-
merly prevailed. Even Johannsen admits that large
germinal changes (‘‘mutations’’) sometimes occur. He
himself records having observed them. Why should we
be so skeptical about the occurrence of minor germinal
changes? Itis easy to overlook them when purely somatic
changes are associated with them and outnumber them as
they possibly do in Johannsen’s beans and Pearl’s fowls
but a single clearly established case should suffice to estab-
lish their existence and their importance in evolution.

THE INHERITANCE OF BLACK-EYED WHITE
SPOTTING IN MICE

C. C. LITTLE

Buack-EYEeD white varieties of rodents have long been
recognized and used as material for genetic investigation.

Cuénot, Morgan and Durham with mice and Castle with
guinea-pigs have utilized this particular color variety in
breeding experiments. For the most part they are agreed
that black-eyed white varieties represent an extreme con-
dition of the ordinary ‘‘spotted’’ or ‘‘piebald’’ series.

Cuénot (1904) in treating the inheritance of spotting
concludes that there exists a continuous series of partially
pigmented forms extending on the one hand from mice
with white on the tail, or with a small white ventral patch,
or with small white forehead spot, through a series of
decreasingly pigmented forms until the black-eyed white
form is reached at the other end of the series. As toa
factorial explanation for the phenomena observed in the
inheritance of spotting, Cuénot feels that there are nu-
merous stages of the spotted condition (P) which he desig-
nates by pt, p°, p?, p* as progressively whiter forms are
considered. He believes, however, that the details of
spotting are not represented in the germ cell. He further
mentions the failure to obtain any particular stage of
spotting in a true breeding condition. Selection of nearly
solid-colored forms has enabled him to obtain animals
with greatly increased white areas.

Durham (1908) has obtained some evidence for two
different types of spotting, one recessive to solid-coated
forms and one dominant to them. She has reported sev-
eral crosses which I have considered in more or less detail
in another paper (Little, 1914). None of the crosses pre-
sented by her can be considered as critical tests of the
presence of two distinct spotting factors. Morgan (1909),

727

728 THE AMERICAN NATURALIST [ Vou. XLIX

who has worked with the same types as Durham, feels
uncertain as to the real significance of black-eyed whites
and as to the occurrence of a distinct factor for dominant
spotting. This uncertainty I also felt and have tried to
show further reasons for not considering Miss Durham’s
work as establishing the existence of a dominant spotting
factor.

Castle (1905) has found that in guinea-pigs black-eyed
whites behave in inheritance in much the same way that
the same type of mouse behaves, namely that black-eyed
whites do not breed true but give, when crossed inter se,
a whole range of spotted forms in addition to some like
themselves.

One can by selection progress in either direction through
a series of spotted forms, decreasing or increasing the
number and extent of pigment patches. Great difficulty,
however, was encountered in trying to fix the color pat-
tern at any particular stage in the series. Up to the
present time this has not been proved possible.

EXPERIMENTAL

In the early winter of 1913 Dr. Castle obtained from a
fancier in England two pairs of black-eyed white mice.
These he kindly handed over to me for investigation.
From the outset the progeny of these mice e to be
extremely healthy and vigorous.

1. Black-eyed White Crossed Inter Se

This cross gave two distinct classes of young, black-
eyed white and ‘‘piebald.’’ The distinction between the
two classes can best be shown by the tabulation of their
progeny on the basis of the amount of dorsal pigmenta-
tion they possess. I have for some time estimated the
per cent. of the dorsal surface pigmented in the case of
all spotted animals recorded. This gives a basis for clas-
sification which, though it may at first glance seem to
inexact, nevertheless has been shown by comparing the

No. 588] INHERITANCE OF SPOTTING IN MICE 129

estimates of two or more investigators on any one animal
to be surprisingly exact and fully as satisfactory as any
other system of grading.

TABLE I.
Per Cent of Dorsal Pigmentation

Type of Cross S| 3) S/8| 8/3| S| 9| 3| 3| 3| 3 £| £| 3| 3| 8| 28| 2
g/g agai g/e gi aigggaigaddigiiiz

Black-eyed white in-
ter se 56 1/0/1/3| 5} 6) 7| 4) 3/11; 0}1/0/0/ 2) 1| 0/1] 0

— ae white X
iebald 105; 0} 0} 1 |1/12) 8)11) 5) 6) 4) 4/1)3)4)] 9 4; 913) 0
Pisbald X piebald ARI 0.0} 0/9/0} 4| 6| 7/13) 6 2| 9/4|6|2| 5 8/15) 2) 0
Total 161; 1/0 2 [4/21 0 /25)28)15) 7 |18 619) 6 /16)13'24) 6| 0

From Chart I it will be seen that 44 of the 75 young
obtained fall in the class between 0 and 5 per cent. of
dorsal pigmentation. These are the black-eyed whites.
é

CHART 1

-

0-5 6-0 WAS WO 31-25 1630 3-3F KNO W-VF USP -SF bo S-S Go THIS N-S -IS FOR 9-95 Tew

730 THE AMERICAN NATURALIST [Von XLIX

The remaining 31 young are more or less scattered along
the range of ‘‘piebald’’ forms. The gap between the two
classes is a considerable one and is certainly significant.

b N

Fig. 1 Fic. 2

Figs. 1—4 are diagrammatic and are intended to show
the two groups of spotted animals. Figs. 1 and 2 show

No.588] INHERITANCE OF SPOTTING IN MICE 731

the common range of variation within the black-eyed
white type and Figs. 3 and 4 the same for the ‘‘piebald’’
type.
2. Black-eyed White X Piebald

This mating brought out two interesting facts. First,
all black-eyed whites behaved in essentially the same way,
approximately an equal number of black-eyed white and
piebald young being produced. Second, the same dis-
tinctness between the two types held good, as will be seen
from the chart given below (solid line).

3. Piebald « Piebald

Piebald animals from black-eyed white parents and
from the cross of piebald X black-eyed white were mated
inter se. They produced only piebald young, 93 in
number.

The distribution of these young according to the degree
of dorsal pigmentation they possessed is shown by Chart
2 (dotted line).

It will be noticed that there is no approach to the black-
eyed white condition (0-5 per cent.). There are also in-
dications of two main modal points, one at 41-50 per cent.
and one at 80-90 per cent. A complete curve formed from
the sum of all piebald animals included in Table I, is
given in Chart 2 (broken line).

This further emphasizes the bi-modal nature of the
curve in the case of piebald mice and makes it seem likely
that there are two genetically distinct grades of this
variety. It is hoped that opportunity will arise in the
future to investigate this point more accurately.

4. Discussion
From the three types of matings given above the fol-
lowing facts may be deduced: (1) The inheritance of the
characters in question is alternative, not blending in
nature; (2) black-eyed white is epistatic to ordinary pie-
bald spotting.

732 _ THE AMERICAN NATURALIST [ Vou. XLIX

e CHART 2
BLACK- EYED WHITE Aang (SOLID LINE) —
PIEBALD X PIEBALD (DOTTED Line ).....
COMBINATION OF ALL means hers LINE)—-—

So

a-r GH TT (M-10 tu-a 426-30 31-95 "36-40" Hi 4s! e-m! prs R-lo GIGS! bé-70 6 I-75 ' T6- fp Bie BS! g6-fo Mirar” 9-ta)
The behavior of black-eyed whites in crosses 1 and 2,
Table I indicates that they are always heterozygous domi-
nants and that they can not, therefore, be obtained in a
condition to ‘‘breed true.”
With this in mind it is interesting to calculate the ex-
pected ratio when black-eyed whites are crossed inter se.
If black-eyed white is due primarily to a dominant factor

No.588] INHERITANCE OF SPOTTING IN MICE 733

which obeys the ordinary laws of mendelian inheritance,
we should expect that black-eyed whites would be obtained
of two genetic types, homozygous and heterozygous. If
now black-eyed whites were mated together at random,
the matings should be either (1) DD x DD, (2) DD x DR
or (3) DRX DR. In the case of (1) and (2) only black-
eyed white young should be produced, while type (3)
should give approximately 3 black-eyed whites to one pie-
bald. Random matings would therefore produce a ratio
of black-eyed whites to piebalds considerably in excess
of 321.

If, on the other hand, the DD form of black-eyed white
mice behaves in a fashion similar to the homozygous yel-
low mice, failing to develop, we should expect a ratio of 2
black-eyed whites to one piebald young, no matter what
the origin of the black-eyed white parents might be, when-
ever two black-eyed whites are bred together.

The results are as follows:

Black-Eyed White X Black-Eyed White

Black-eyed White Piebald
Obsörvod nci aie ORS eS 57 39
Expectod 8:1- ratio «6. cei. eee 64 32
expected 3:1 tso Gis sikiwiecns 6 72 24

When one realizes that the ratio in one case should be
considerably higher than 3:1, it seems that the results in-
dicate a 2:1 ratio and the heterozygous nature of black-
eyed whites.

To further test this hypothesis individual tests of
twenty-one black-eyed whites coming from black-eyed
white parents were made by crossing with piebald
animals. If the DD combination is possible, approx-
imately seven of the twenty-one tested should be of that
constitution. All of them, however, proved to be hetero-
zygous. While the numbers should be supplemented by
further tests, they are certainly sufficient to serve as a
basis for a tentative conclusion that black-eyed white
mice are always heterozygous. :

734 THE AMERICAN NATURALIST [Von XLIX

The numbers from the cross of piebald X black-eyed
white are more extensive and closely approximate a 1:1
ratio. The numbers obtained are 105 black-eyed whites
and 102 piebald, while the 103 of each would have been
exactly an equality ratio.

The behavior of the piebald animals when crossed inter
se is exactly what would be expected if piebald was hypo-
static to black-eyed white and distinct from it in inheri-
tance.

The next question to be considered is the relation of
black-eyed white to ‘‘self’’ or solid coat, in inheritance.

RELATION oF BLACK-EYED WHITE TO SELF

A preliminary investigation of this question has been
made. The ‘‘self’’ race used was really technically not
a ‘‘self” but genetically it carried neither the black-eyed
white nor piebald spotting factors. Somatically the self
race used was a ‘‘blaze’’ race of the type which I have
previously put on record. Further crosses which I have
made between black-eyed whites and true selfs have

‘shown, even in early stages, clear evidence that the be-
havior of the blaze and true self races is directly com-
parable.

1. “Self”? X Black-eyed White

The F, generation produced by crossing self (blaze
F6B) animals with black-eyed whites consists of two very
distinct forms. These have been produced in a ratio of
50 Type ‘‘A’’ to 47 Type ‘‘B.’’ The first of these, Type
**A,’’? is shown in Fig. 5. While the percentage of dorsal
pigmentation of this type is subject to some variation
(see table), it will be noticed that they are ordinarily be-
tween 80 and 90 per cent. colored. The spots of color

$/3/3|3|3|/2|8|8|3|3|2|3

Type “A” t a: Tis

: PLP Desa 2) f/E/22]3
Black-eyed white X self (blaze))0/1/2/0/1}1/2 0/9 |15)10)8/}1

No.588] INHERITANCE OF SPOTTING IN MICE 735

appear to have slightly more irregular and less clearly
defined outlines than do those of the ordinary piebald
mice and many of the spots are distinctly smaller in size
(compare Figs. 3, 4 and 5). Just how much of this ap-

——— N oe : aa
= OO

ey

D V D v
Fic. 5 Fie. 6

pearance is due to true genetic difference between the two
types of spotting is of course problematical and must
remain so until a larger mass of data is available.

Concerning class ‘‘B’’ (Fig. 6) little need be said save
that they appear in every way identical with heterozy-
gotes ordinarily obtained in a cross between ‘‘self’’ and
‘‘niebald’’ animals. They vary from entirely solid colored
animals to those having approximately 20 per cent. of the
ventral surface white. They may be tabulated as follows:

Per Cent of White on Ventral Surface

5 0 | 1-5 | 6-10 | 11-15 | 16-20 | 21-25 | 26-30
fee SB? o n eee a eee ad

2. Type “A” Animals Crossed Inter Se

Type ‘‘A’’ animals obtained in F, are distinctly ‘‘spot-
ted.” They have a clearly discernible amount of white

736 THE AMERICAN NATURALIST [Vou. XLIX

and are not in the least like heterozygous ‘‘selfs’’ of any
recorded type. When crossed together they give three so-
matically distinct classes of young, ‘‘self,’’ ‘‘piebald’’ or
like class ‘‘A,’’ and black-eyed white. The numbers ob-
tained are 15 ‘‘self,’’? 31 spotted (piebald or like class
**A’’) and 11 black-eyed whites.

3. Type ‘‘A’’ X Piebald

To test them further type ‘‘A,’’ animals of this class
were crossed with homozygous piebald mice extracted
from the black-eyed white crosses. Again three general
classes of young were obtained as follows: 45 ‘‘self,’’
54 spotted (piebald or like type ‘‘A’’) and 29 black-eyed
whites.

4. Type ‘‘B’’ X Piebald

To compare the behavior of types ‘‘A’’ and ‘‘B”’ this
cross was made. Only two classes of young resulted as
follows: 82 class ‘‘B’’ and 78 piebald. No black-eyed
whites were obtained.

Discussion

The question now arising is whether the factors for
self, black-eyed white, and piebald are allelomorphic or
independent in inheritance.

From the nature of the F, generation it is certain that
the black-eyed white animals are forming two kinds of
gametes in respect to their spotting factors.

If now the conditions ‘‘self’’ coat, ‘* black-eyed white”?
and ‘‘piebald’’ are all related as members of a system of
triple allelomorphs, we can express the cross as follows:

S=self factor.
W= black-eyed white factor.
sp=piebald factor.

Then

S S=self X Wsp=black-eyed white
gametes § WwW
sp
F, Generation S W= Type A, Fig. 5
S sp = Typo B, Fig. 6

No.588] INHERITANCE OF SPOTTING IN MICE 737

If now animals of Type A are bred inter se we should
expect

S|wWxS W
LSS self
2 SW =like Type “Ar”
1 WW = (not formed because homozygous)

The one WW individual could not be formed since by
experiment it has been shown that W can exist in only
one of the two gametes forming a zygote. When W meets
S, an animal like Class A is produced, when it meets sp
a black-eyed white results.

The expectation therefore is that, if a system of triple
allelomorphs is operative here, we should have no black-
eyed whites formed from mating together class “A”
animals.

The result of this mating quickly settles the above hy-
pothesis for 15 ‘‘self’’ colored, 31 spotted (like or nearly
like Type ‘‘A’’), and 11 black-eyed whites have been
obtained.

It is clear, therefore, that ‘‘ black-eyed white’’ depends
upon a factor which is at least partly independent of that
producing ‘‘piebald’’ spotting. Let us suppose that this
is the case and that ‘‘black-eyed whites” always carry
piebald in all of their gametes and an epistatic inhibiting
or restrictive factor producing increased whiteness in one
half their gametes. If W equals restrictor and w its ab-
sence and sp equals the factor for piebald spotting, all
black-eyed whites will be Wwspsp, in zygotic formula and
will form two sorts of gametes, Wsp and wsp.

This will account for the results in mating black-eyed
whites inter se due to the failure of the WWspsp zygote
to continue its mee Mncose ‘nbd because of the double dose

of W.

If now black-eyed whites Wwspsp are crossed with selfs
wwssS, two classes of F, zygotes will result, WwSsp and
wwSsp. The former will produce a new zygotic combi-

738 THE AMERICAN NATURALIST [ Vou. XLIX

nation really differing from the black-eyed whites in the
substitution of a ‘‘self’’ bearing gamete for a ‘‘piebald’’
one in the zygotic formula. The result is an animal like
Type ‘‘A,’’ Fig. 5; Type ‘‘B,’’ Fig. 6 shows the other F,
type which is entirely free from the W factor and which
is merely a heterozygote between ‘‘self’’ and ‘‘piebald.”’

If class ‘‘A’’ animals are crossed inter se we should on
this new hypothesis expect the following results.

OW WS oora not developed

S EWE Fs cnc very dark spotted

2 WWD siari not developed

© WWO -ree like parents (type ‘‘A’’)

2 WD ices os not developed

2 Wwspsp ....... black-eyed white

I WROSD iia ceu ‘self’?

2 WWD s.is :oe ‘‘self’?’ or ‘‘self’’'with white ventral patch (type ‘‘B’’)
1 wwspsp ........ ‘*piebald’?

Four of the 16 zygotes in F, would have two doses of
W and would not develop. Of the remaining 12, seven
would have some degree of white spotting depending upon
whether they were WwSS, WwSsp or wwspsp in formula;
three would be ‘‘solid’’ colored or like type ‘‘B’’ of F,
and two would be black-eyed whites.

On this hypothesis the F, generation would be as fol-
lows:

Observed Expected
DOR -oree ea ee eG 15 15
OOO -e iirrainn ra 31 35
Biack-oyod. whitòs 66 ei id ce ec cues 11 10
57 60

A further test of the nature of type ‘‘A’’ is possible.
If they are bred to piebald animals, four classes of young
should result as follows.

WOD 0G oa iv nee yk cel ks Cinna kiss like class ‘A
OU ices oh oes ed ce ke hes aaa black-eyed whites
NS i si as oi edn os Wee aaa solid colored
MONI 6 cei nec cs sé vee peensiciveses

Lumping together the WwSsp and the wwspsp animals

No.588] INHERITANCE OF SPOTTING IN MICE 739

we should have 2 spotted, 1 black-eyed white and 1 self.
The results are as follows:

Observed Expected
MIOCENE S ae ee Ges es 64
OE ge i Oro ce oe eee as 45 32
Hlack-eved: White: oss sad fy. Cosix% 29 32
128 128

Whether the excess of ‘‘self’’ animals is significant is,
of course, a question to be borne in mind but it is ex-
tremely doubtful whether it is due to anything more than
a chance deviation. |

Type ‘‘B’’ animals have, upon mating with ‘‘piebald’’
individuals, given very close to the expected ratio of 1
type “B to 1 ‘‘piebald.’’ The exact numbers are
82:78; expected ratio 80: 80.

Is BLACK-EYED WHITE IN Mice an ALLELOMORPH OF
ALBINISM?

The experiments of Castle and Wright have shown that
a dark red-eyed variety of guinea-pig exists which is an
allelomorph of dilute pigmentation and of albinism. This
possibility in the case of mice is eliminated by crossing
black-eyed white with albino, when on the supposition
that the condition found in guinea-pigs holds true in mice
all the young should be either black-eyed white, albino or
dilute pigmented. Actually there were obtained from a
single mating of this sort five young, all intensely pig-
mented, two blacks and three browns; thereby eliminating
the possibility that black-eyed white, in mice, is an
allelomorph in the albino series.

CoNCLUSIONS

The fact that black-eyed white spotting in mice ap-
pears to be due to a factor independent of and supple-
mentary to the factor for ‘‘piebald’’ spotting leads to
interesting speculation as to the nature of spotting and

740 THE AMERICAN NATURALIST [ Von. XLIX

indicates that spotting in mice is dependent upon more
than one pair of clear-cut mendelizing factors. Modify-
ing factors which may be more or less difficult to analyze
but which nevertheless are certainly present, contribute
to the extent of variation in spotted races.

‘*Blaze’’ or forehead spotting is apparently independent
of ordinary ‘‘piebald’’ spotting, as I shall hope to show
in a future paper; ‘‘black-eyed white’’ is primarily due
to an independent genetic factor and ‘‘piebald’’ makes a
third independent type. If now in the ‘‘piebald’’ stock
there exist at least two genetic races as are indicated by
the curve of all piebald animals obtained in the ‘‘black-
eyed white’’ crosses, the condition is still further com-
plicated. At all events one can truthfully say that the
distribution of pigment occurring as it does along a series
from ‘‘self’’ colored to ‘‘ black-eyed white’’ animals, offers
a field for the activity of many mendelizing factors.
There is no a priori reason why this should not be true,
there are many experimental reasons steadily increasing
why it appears to be true.

Spotting in rodents is tempting as genetic material be-
cause of the clear patterns and contrast between colored
and white areas. It is, however, as a character extremely
sensitive to minute quantitative and qualitative changes
and its apparent genetic simplicity is a snare and a de-
lusion.

LITERATURE CITED
Castle, W. E. 1905. Carnegie Inst. of Wash. Publ. No. 23, 78 pp.
Castle, W. E. 1914. Am. Nar., Vol. 48, pp. 65-73.
Cuénot, L. 1904. Arch. Zool. Reo. et Gen., Notes et Revue (4), Vol. 2,

Durham, F. M. 1908. Rept. Evol. Comm., No. 4, p. 41.

Little, C. ©. 1913. Carnegie Inst. of Wash. Publ. No. 179, pp. 11-102.
Little, C. ©. 1914. Am. Nart., Vol. 48, pp. 74-82.

Morgan, T. H. 1909, Am. Nart., Vol. 43, pp. 493-512.

Wright, S. G. 1915. Am. Nart., Vol. 49, pp. 140-148.

THE F, BLEND ACCOMPANIED BY GENIC
PURITY

A Description OF MECHANICAL CHARTS FOR ILLUSTRATING
Menpevian Hereprry 1x Hacw or Tares WELL-
KNOWN Cases oF BLENDING INHERITANCE IN
THE First HYBRID GENERATION

HARRY H. LAUGHLIN

EvGENICcS Recorp Orrice, Coup SPRING HARBOR, N. Y.

THE mechanical charts herewith figured are the first of
a series prepared for the purpose of presenting graph-
ically and schematically the established facts of heredity.
These particular mechanisms, illustrating blending in-
heritance, consist essentially of wooden slabs on which the
gametic formule of the several generations are charted

—those for P, and F, are written on flat surfaces, ©

while that for F, is inscribed on cylinders which turn
freely. A capital letter represents a gene; the corre-
sponding small letter the absence of that gene. The lo-
cation of genes, whether they lie in the same chromosome

i. e., are linked, or in different chromosomes, is shown’

graphically by placing their symbols in the same or in
different squares, or upon the same or different half-
cylinder surfaces. In each of these selected cases the
individuals of the P, generation are homozygous in re-
spect to both of the traits or allelomorphic phases con-
cerned. The genes contributed by the P, generation to
the F, zygote are charted on the starred faces of the
freely turning cylinders. The back of each spool contains
the same inscription as the face of its partner cylinder.
Each face of a cylinder represents a chromosome—the
two faces the two chromosome types in reference to the
741

?

[Von. XLIX

T

fag
N
f
~
N
É i
Z
+

E
O
an
aa

4

THE AME

No. 588] F, BLEND AND GENIC PURITY 743

traits lying in that particular chr , which each F,
individual as a parent is capable of passing on. There-
fore, by turning the spools so that all possible combina-
tions are made, one can read off directly all of the dif-
ferent hereditary potentialities to be had by inbreeding
the F, generation. Consequently the F, line (which is
charted on a flat surface) is simply a record of such
combinations.

For the purpose of this study a case of blended inher-
itance is one in which the development in F, of a given so-
matic trait—regardless of whether it develops from one
or more genes—is about midway between its development
in the two parents, each of which is of pure stock in refer-
ence to the trait concerned. Until about the year 1910
students of heredity were unable to coordinate the general
rule of dominance and segregation on one hand, with the
frequent exception of blending and segregation on the
other. Now the existence of at least three different routes
by each of which nature arrives at the somatic blend in F,
are recognized, and each finds ready interpretation in con-
sonance with the theory of the pure gene. The first of
these is the dilution or true blend route, by which nature
appears to travel in the classical cases of the Blue Anda-
lusian' fowl resulting from the crossing of splashed-
white and black parents, and of the pink four o’clock
(Mirabilis jalapa) resulting from the crossing of red and
white parents.

The ordinary mode of inheritance is strongly duplex—
that is, the zygote normally possesses two genes for each
trait, either one of which genes is usually sufficient—with
possibly a liberal surplus of valence—to give full somatic
expression to its correlated trait. In such cases complete
dominance in F, and clear-cut segregation in F, are the
rule. Occasionally, however, in cases wherein a duplex
parent possesses a strong somatic development of a trait,

1‘*Mendel’s Principles of Heredity’? (3d Impression, 1912), p. 51, by
W. Bateson.

744

THE AMERICAN NATURALIST

=

Plumage Color in Andalusian

tically equalValenc
or o affecting t yh ae ha
Bath of thead ir inthe

eco

P.
Somat Soma:White

Rac.

Ww Wii i.
Fann Eas

sii sdad lak ia dae Cowes which
somatic peg waged womens, Se eel areas from patches

on is aor ots i
HM whe show a black or blue ji
pee i EA
foie wants

by one equally post
blue because of Ad mtrinsic pir aeni a gn ar that hai
the genes N and WX le in the same chromogome.

+ N nigrum) i.e.black

Fowl:-
4s due to lwo inten oe poen for dom-
inant while and me ck- of atang

the

Mechanism for lllust
Inheritance of

[Vou. XLIX

SSS

Nar|

—_

conus ts ting the Party nt ai,

E

Fic. 2. Chart showing the F, Blend Associated with Genic Dilution—the True
Blend

No. 588] F, BLEND AND GENIC PURITY 745

a single gene—from the paternal or the maternal line only
—for such trait, in the zygote, is not sufficient to give a
somatic development of the trait equal to that possessed
by the duplex parent. In such cases, therefore, the unit
trait in question is blended in the F, soma—a case of
imperfection of dominance.? Nevertheless, in such cases
segregation is just as clean-cut in the germ-plasm as it is
in the cases accompanied by strong somatic dominance.

In Andalusian fowl ‘‘W’’—dominant splashed-white—
and ‘‘N’’—(nigrum) black—are two opposing and allelo-
morphic genes of nearly equal valence in ontogenesis.
Their combination and interaction determine plumage-
color in the offspring. The black Andalusian is duplex
for black plumage-pigment, while the splashed-white is
duplex for dominant splashed-white. The F, offspring
are ‘‘blue’’—a shade really intermediate between: the
white and the black. Moreover, the genes ‘‘W’’ and “N”
evidently lie in the same* chromosome. The evidence for
this consists in the fact that in the F, generation, result-
ing from inbreeding two blue Andalusians, neither albinic
white nor jungle‘—pure or modified—patterned fowl re-
sult, which would be the case if ‘‘N’’ and ‘‘W”’ lay in
different chromosomes, permitting, in some F, zygotic
combinations, the elimination of both ‘‘N’’ and “W.”
For further explanation of this particular type of blended
inheritance see the accompanying figure descriptive of the
mechanical chart ‘‘Plumage-Color in Andalusian Fowl.’’

The second type—that of multiple factors—is typified
by the inheritance of black skin-pigment in man. It isa
matter of common knowledge that a mulatto of the first
generation is about intermediate in density of black skin-
pigment between his white and his black parents. In 1913

2‘*Tmperfection of Dominance,’’ American Breeders Magazine, No. 1,
Vol. 1, p. 39, 1910, by C. B, Davenport.

3‘‘ Heredity and Sex,’’ p. 93 et seq. (Columbia University Press, 1913),
by Thomas H. Morgan.

4‘*New Views about Reversion,’’ Proceedings of the American Philo-
sophical Society, Vol. XLIX, No. 196, 1910, by C. B. Davenport.

746

THE AMERICAN NATURALIST

Black Skin-Pigment in
Ma Ti:=

i/sdue to two segregable genes in

each gamete. :
2The potentiality ofe each gene finds

Le ed

PIIGAD

regardless of the presence or ab-

sence of other genes.
Awhile man A woman
6% Ni in skin. 70% N M skin.

as 9 a : ww

bs
z Jez ae

©
HR-
as
tie
Ka

s not found in PaiFim- Gametic types sat keratin

Ne hak DAR
oe figures 1 the pigment, ‘producing power (in per-

gw} ion ai au Scie shines families Da t found
dian Negro- venpor:
five freque m densily of skin- pigmenti-

4
554N 462N youn

[Vou. XLIX

there s sible is
‘ack Sare pr tle uc malingo, pare paces the de

Mechanism for illustrating the of the

s manner
ape e, madre gae g ak Po tir a y mi « m marn.
Negro-While Grosses’-Davenpart

Fic, 3. Chart showing the F, Blend Associated with Multiple Factors for One
S

omatic Trait

No. 588] F, BLEND AND GENIC PURITY 747

Dr. C. B. Davenport® found, by analyzing data on the
family distribution of black skin-pigment measured quan-
titatively (by the color-top) among the mixed white-and-
black families of the Island of Jamaica, the Island of
Bermuda, and in our own Southern States, (1) that black
skin-pigment in man is the somatic working out of two
segregable genes in each gamete, and (2) that the poten-
tiality of each gene finds definite measurable somatic ex-
pression, regardless of the presence or absence in the
zygote of other genes. Now these two genes appear to
be of different valence; they appear also to lie in differ-
ent chromosomes. The scheme outlined by the mechan-
ical chart ‘‘Black Skin-Pigment in Man”’ is quite conso-
nant with the facts of inheritance which Dr. Davenport
found in nature. The facts seem to be that in white per-
sons one of these genes will develop from practically none
to about 1 per cent. of blackness in skin-color, and the
second from very little to about 2 per cent., thus resulting
in a blackness of skin-color of 6 per cent. or less in the
somas of members of the light races. He found that some
races of negroes show about 70 per cent. black in skin-
color. In such races one gene for black skin-color seems
to be potential to developing approximately 16 per cent.
of black skin-color, the other about 19 per cent. The evi-
dence that there are two such genes, and that they are
segregable, i. e., that they lie in different chromosomes,
and that their values among the strains studied are about
as described above, lies in the fact that, in the hybrid
families in Bermuda, Davenport found 5 frequency max-
ima in intensity of black skin-pigmentation, and that his
analysis of the family distribution of this trait, quanti-
tatively measured in many mongrel families of known
pedigree, demanded the existence in nature of the scheme
above outlined.

Darwin, whose method of study was essentially obser-
vational, knew that the F, generation was quite generally

5 ‘*Heredity of Skin-Color in Negro-White Crosses,’’ published by the
Carnegie Institution of Washington, 1913, by Charles B. Davenport.

748 THE AMERICAN NATURALIST [ Vou. XLIX

remarkably uniform, but among and beyond the F, gen-
eral observation found no rule of inheritance. It re-
mained for the application of the analytical or Mendelian
study to discover order in the apparent somatic tangle of
F,. The skin-color story just related is a striking case
in point.

The third class of blended inheritance—the particulate
or mosaic—is typified by the behavior in heredity of coat-
color in short-horn® cattle in which, in the F, soma, the

Black Man Yellow Man White Man

Louis 5. rie

P

«The red of these aga carmine wW which accordi
to Riòway’s Color Standards is composed 552 of spectrum
red, 45z of black.

Fie. 4. Composition of Skin-pigmentation in Representitives of Three Races.

_Jamaicans- of the Moneaque Holel, Monea Jamaica.

Mimie Webster Victer Webster David MacDonough CLlewellyn

N.162
R pas Be R.44%
Y. 292 Y.252z Y.162 y. sor
W. 272% W277 We4ex W. 192
1002 10027 “O07 hoor
Broca's Scale Brocas Scale Brocas Scale Scale Broca's Scale Broca's Scale
E T E ese TEE y SIE ORA #3.

Fie, 5. Variation in Skin-pigmentation Among Jamaicans

6 ‘‘ Inheritance of Coal-Color in Short-horn Cattle,’? AMERICAN NATURAL-
ist, December, 1911, January, 1912, by H. H. Laughlin

No. 588] F, BLEND AND GENIC PURITY

Coat Color in Short-horn
Caltle:-
is dependent upon ive ae “the”

inheritance areas
conlrol of which lie in the same@ch
mosome.

@)1Hrea cavers two flank belts, the underline, the
median line of the face, pou a mosaic, either coarse
or fine, sagas. 3 the remainder

2.Area twa cove s the neck, the sides, the back, hina
quarters and and a mosaic either coarse,
covery; remainder of the i Pony oe omen or of area
ay ee

one the breed white c
as He tae recessive alelormorph Cs te note)
> re area twe pea breed a over:
8) The, I The proof consists in the fact GEO indiridon] Shot

nd while in area twa
& pan stands for any mosaic- either fine or coarse-
of red and white.

Nete~
‘Ry is dom, is Sorina piad ti mall r for
wi

nally a as frequen ared
are coe will produce a spotted ora roan
Vif are Se oak po cae than Bme strai Pa
same aren jn another strain, fines caff produced
bya mating would the following
g Area one WR-opposing positive genes allelo-

mor, $

he this
Pis > aaee herewith vith deacrited, including th

the allelo s for o that
E morphic genes f ra-sygotic reaction i
a fluctuation thru the criti paint of somale
Servet abe te a auth feller actos oF
coat-color in Shorthorn cattle,

EER oe jy Paw) Sone ay: Tid nea e-

749

Lellers identify spools.

aa)

Fic. 6. Chart showing the F, Blend Associated with Particulate Inheritance—

a Patent Mosaic

750 THE AMERICAN NATURALIST [ Vou. XLIX

character concerned is, in its grosser aspect, clearly mid-
way between the corresponding traits of its two parents,
although a closer inspection reveals a mosaic the elements
of which are the parental traits quite unchanged. The
difference between the Andalusian fowl and the short-
horn cattle cases seems to be as follows: In the Anda-
lusian each gene influences the entire plumage-color, and
appears to be struggling unsuccessfully, as it were, for
the supremacy in somatic expression, thus resulting in a
very fine and quite generally distributed blend or mosaic;
while in short-horn cattle the controlling genes are double
the number, each pair being confined to specific coat
areas in somatic expression, and the resulting mosaic,
although quite variable in coarseness, is always relatively
coarse and is also quite definitely patterned.

Thus, normally (for the exception see the note in Fig.
6) in Area 1 the gene ‘‘W”’ is clearly dominant over the
gene ‘‘R.’’ In Area 2 the gene ‘‘R”’ is dominant over its
absence. There seems to be in Area 2 no competing or
allelomorphic gene whatever—it is simply “R?” or its
absence, i. e., albinic white; whereas in Area 1 the ‘‘W,”’
which is epistatic to ‘‘R,’’ will leave ‘‘R’’ by its absence.
The evidence for all this consists in the fact that a white
short-horn (which is evidently dominant white, always
duplex, in Area 1, and always recessive white in Area 2)
will, when crossed with a black Angus, which is dominant
black for its entire coat, give in the offspring a calf domi-
nant white, simplex, in Area 1, and black, simplex, in Area
2—the familiar ‘‘blue roan” in cattle. That in short-horn
cattle the genes ‘‘W”’ and “R” lie in the same chromo-
some is sufficiently proved by the fact that the color
pattern is never reversed, that is to say, in bi-colored indi-
viduals of whatever coarseness of mosaic, Area 1 is

(Note:—When this paper on coat-color was written it was pointed out that
coats red in Area 1 and white in Area 2 were never observed. Now the
modified interpretation, involving linkage and a variation in genic valence,

as explained in the text and Fig. 6 of the present article, accounts for prac-
tically all of the observed facts.)

No. 588] F, BLEND AND GENIC PURITY 751

always dominant white, and Area 2 is always red, and we
never find an individual red in Area 1 and white in Area
2, although solid whites and solid reds, and bi-colored
individuals of the first specified type are common. The
reversed pattern, i. e., red in Area 1 and white in Area 2,
would occur if the genes ‘‘W’’ for Area 1 and ‘‘R”’ for
Area 2 were completely segregable, i. e., if they lay in dif-
ferent chromosomes. For a further explanation of this
mode of blending inheritance see the accompanying chart,
‘*Coat-color in Short-horn Cattle.’’

THE POPULATION OF THE ‘‘BLANKET-ALGO”’
OF FRESHWATER POOLS!

' EMILIE LOUISE PLATT
CoRNELL UNIVERSITY

Tis is a study of the community of life that is bound
up with the floating masses of filamentous alge, popularly
known as ‘‘blanket-alge.’? An acquaintance with this
population is worth cultivating for the sake of the variety,
beauty and interesting peculiarities of the plants and ani-
mals found in this unique habitat. It may be of utili-
tarian value as well, for there exists a relation between
plankton production, algal growth and fish culture. Fur-
thermore, it may be a help to students and to teachers of
biology when they are in search of certain laboratory ma-
terials, which in these alge masses flourish.

Method of Collecting.—A fine silk hand net of No, 12
bolting cloth was used to lift the alge from the surface of
the water. The largest collection covered about 2,800
sq. cm.; the smallest about 10 sq. cm., but most of them
were from 200 sq. em. to 800 sq. em. in area. Doubtless,
many active and comparatively large foraging animals,
such as small fishes or adult insects, escaped while the
net was surrounding and enveloping the mass. Probably
comparatively few of the smaller forms were lost through
the fine silk mesh of the net. The volume of the mass
was then computed in cubic centimeters. As the mass
sometimes lay in thin layers and sometimes in thicker
masses, the proportion of volume to surface was seldom
the same. About 200 cu. cm. was the average. The com-
ponents of the ‘‘blanket’’ were determined and all forms,
plant and animal, were listed and their size and relative
abundance noted. The collections were made during the
fall and early winter of 1912 and the spring and early
summer of 1913.

Location and Character of the Pools.—The pools are
all located in the vicinity of Cornell University campus at

1 This study was carried on in the limnological laboratory of the depart-

ment of entomology of Cornell University under the direction of Professor
James G. Needham.

752

No. 588] POPULATION OF “ BLANKET-ALGÆ ” 753

Ithaca, N. Y. (see map). They varied from shallow,
transient collections of ditch-water to large, permanent,
usually stagnant pools. Those lettered B, C, D, G, J, M,
and N belong to the first category. Pools x, x’, x”, x, x*,

CAY.: ;
Fp / |

wer Ny = — Nas

\\
[>]
©
A
w
`
»
<
j
K
„$
“o
o
©
S
‘ \

MIL Š
CONTOUR INTERVAL £00 FEET

Pools in the Vicinity of Cornell University Campus.

and y, y', E, K, I, and L are permanent pools and measure
from four to thirty or more inches in depth. Pool H is a
quiet part of a large stream. Pools F and A are artifi-
cially enclosed and are filled from pipes. The pools of
the lowland of Cayuga Valley (about 400 ft. above sea-
level) are A, B, C, D, and E. The others are among the
hills (about 800 feet above sea-level).

The Filamentous Alge of the Floating Mass.—Although
there was such variety in seasonal conditions and in the

754 THE AMERICAN NATURALIST [ Vou. XLIX

location and character of the pools, nevertheless some
forms appeared constantly. Among the filamentous alge,
Spirogyra was almost uniformly present, appearing
twenty-eight times out of thirty. The species were not
identified until March, but in the twenty collections taken
in the spring and early summer, the most frequent species
was Spirogyra varians. Spirogyra insignis was found
five times. Other species seen less frequently were:

S. tenuissima S. communis
S. sticticum S. fluviatilis
S. grevilliana S. bellis

S. weberi S. nitida

S. quinina S. inflata

S. crassa S. decimina

S. majuscula S. rwularis

Usually the masses contained several species of Spiro-
gyra, often withalarge proportion of one species, and the
Spirogyra was almost invariably associated with other
filamentous alge. Among the most frequent of these
were Mougeotia and Zygnema. Vaucheria was found
frequently in the autumn and early winter. Oscillatoria
was quite constant after its first occurrence in early
March, but it was usually in very small quantities. Ulo-
thrix Draparnaldia and Microspora were seen occasion-
ally, but not in abundance, while Anabena oscillaroides
was found only once. In general, the large permanent
pools produced the greatest variety of genera and species
of these alge, but otherwise there was no apparent rela-
tion between the genera of alge produced and the char-
acter and location of the pools; with the possible excep-
tion of Draparnaldia plumosa, which was found four
times out of five in shallow ditches.

Diatoms, Desmids and Other Alga.—Diatoms were in-
variably present. Of these, there were four that were
constant and always in greater quantity than other kinds.
These four were Navicula, in great variety, Synedra, Coc-
conema and Gomphonema. Other diatoms were seen ir-

No. 588] POPULATION OF “BLANKET-ALGZ ” 755

regularly as to quantity and time of occurrence and in-
cluded the following:

Tabellaria Cocconets
Fragillaria Campylodiscus
Meridion Amphora
Asterionella Pleurosigma
Diatoma Nitzschia
Encyonema Odontidium
Cymbella Cyclotella

Most of these were free but often Gomphonema, Cocco-
nema and Cymbella were in colonies attached by branched
or simple stalks to larger forms. Encyonema is found
end to end in colonies enclosed in long filament-like gelat-
inous envelopes. Navicula as well as stalked diatoms
sometimes covered the bodies of larvæ and smaller crus-
taceans and also the cases in which the chironomid larvæ
spent part of their time. Variation in occurrence of dia-
toms is apparently due to seasonal changes, which will
þe considered later.

Other algæ were less constant, the most regular one be-
ing Closterium, which occurred in eight collections, show-
ing a number of species. Of the other desmids that ap-
peared, Cosmarium, Penium and Staurastrum were
usually in small quantities. Twice, however, Cosmarium
and Closterium both appeared in abundance, the first
time being in a permanent but shallow pool (I) where
Ulothrix predominated, and the second time in a shallow
but probably permanent roadside pool (G) covered with
Spirogyra. The Volvocaceæ were represented by Volvoz,
Eudorina, Pandorina, Spherella and Chlamydomonas.
Two Phæophyceæ, Dinobryon and Synura, and four Pro-
tococcaceæ, Dictyospherium, Kirchneriella, Protococcus
and Scenedesmus, added variety but did not appear fre-
quently. Peridinium, Pediastrum and Ophiocytium were
rare. :

The pools (Z and Y) that had the greatest variety in
desmids and kindred forms were also rich in diatoms.

756 THE AMERICAN NATURALIST [Vou. XLIX

These pools are large and one or two feet deep and have
thin mud overlying rock bottom. Both lie near Fall
Creek.

The pools (K, x, 21, x’, x+) near Cascadilla Creek pre-
sented the only specimens of Dinobryon that were seen.
These pools are permanent and deep and have stony
bottom. |

It may be significant that in the low-ground pools there
were few kinds of diatoms and in only one such pool (A)
were there any desmids.

The Animal Population.—The floating and entangled
vegetation of these masses supports a large animal popu-
lation. The protozoans found were particularly varied
and interesting. Ameba, Arcella and Difflugia appeared
irregularly in the upland pools. Cochliopodium and Mas-
tigameba were rare. No other Rhizopods were observed.
The ciliates were not determined before March, with the
exception of Paramecium, which was listed from the first.
In the twenty collections made since March first, fifteen
genera of ciliates have been observed. Paramecium was
constant and abundant. Among the larger representa-
tives of the group, Coleps, Chilodon, Colpidium, Stylo-
nychia and Vorticella appeared frequently and in large
numbers. Stentor, Dipleurostyla and Amphileptus were
less frequent, as were the smaller members of the group,
namely, Euplotes, Halteria and Askanasia. Coleps was
especially noticeable in pool y', while Vorticella was plen-
tiful in pools D and G. Pools D, G and J, which supplied
the largest number of genera and of individual ciliates,
are shallow ditch-pools with muddy bottom, while A and
Y in which smaller numbers were found, but still many
genera, are larger and deeper, but have muddy bottom or
muddy water. From this it seems evident that these
protozoa prefer water with inorganic material in suspen-
sion, although they are said to avoid water polluted by de-
caying organic matter. These tiny creatures forage bus-
ily among the algal filaments, some swimming and rotat-
ing smoothly, others, such as Halteria and Stylonychia,

No. 588] POPULATION OF “ BLANKET-ALGZ ” 757

moving by jerks and sudden dartings hither and yon.
The minute form, Euplotes, has a peculiar method of loco-
motion that looks like walking along a filament, though it
is merely forward progress by means of cilia.

The flagellates were represented by Euglena, Distigma
and Phacus, of which the first was fairly constant. Three
heliozoans, Actinospherium, Actinophrys and Vampyrella,
appeared infrequently. Hydra was found in one collec-
tion only, and no other ccelenterates were seen.

Various worms, mainly the microscopic nematodes as
well as unidentified planarians and turbellarians, also
three kinds of oligochetes, namely Nais, Tubifex and
Chetogaster, were frequent but not regular inhabitants of
the alga-mass.

The rotifers were regularly a part of the population,
furnishing species of eighteen genera. Determination of
genera was not undertaken until March The genus most
constantly in evidence was Diglena, especially in dirty
water, foraging industriously, nibbling and pulling at the
alge. A species of Metapidia with a broadly curved
lorica was seen several times. It clung by its toes to
debris, while the flow of water carried food-particles past
the rotating cilia into the mastax. Anuwrea, Salpina and
Syncheta were found exclusively in the pool nearest Cay-
uga Lake, pool A. Also in this pool, as well as elsewhere,
were found Polyarthra, Rotifer, Adineta, Diglena, Noth-
olca labis and a long-spined species of Rattulus. This
permanent pool and a similar pool (X!) of the highland
were richest in rotifer life. Forms found in the latter
pool and not observed elsewhere, were Notommata and a
species of Stephanops with a fan-like anterior projection
of the lorica. Other genera identified were Brachionus,
Philodina, Mytillina, Mastigocerca and Diaschiza. Cheto-
notus, a representative of the Gastrotricha, was seen a
few times.

The Gastropoda, the only mollusk group represented,
did not furnish a constant element, since only eight collec-
tions contained any snails. Lymnea appeared once,

758 THE AMERICAN NATURALIST [ Vou. XLIX

Physa five times and Planorbis four times. These snails
varied in size from two to twenty-five mm. long. Except
in one instance, they were in shaded pools or came out on
cloudy days. The exceptional case may be considered as
similar because the luxuriant growth of watercress near
the algae furnished spots of shade, although most of the
“blanket” was in sunlight. It seems fair to assume that
snails are not regular inhabitants of the surface alge, but
merely forage there when there is little or no sunlight.

Many small crustaceans were observed. Chydorus and
Bosmina were numerous, while two other Cladocera, Daph-
nia and Simocephalus, were less in evidence The ostra-
cods found were in eleven collections and quite numerous.
They have not been identified. Cyclops, Canthocamptus
and Diaptomus were the copepods identified. Cyclops
was remarkably constant and abundant. Many females
bearing paired egg-sacs and many copepod nauplii, pre-
sumably young cyclops, were among the number. The
adults were from one to three mm. long. The Isopod,
Asellus aquaticus, was found only once and then in a mass
of alge close to a mud bank. Two amphipods, Gam-
marus and Hyalella, were observed several times.

The last group of foraging animals and the one to which
the largest individuals of this population belong is the
Insecta. In this class were found larve, nymphs and
adults, representing five orders of insects. Three nymphs
of Callibetis, in pool A, one of Betis, in pool M, and ten
of undetermined genera of Ephemerida in Pool F were
the only may-fly nymphs found. The Odonata were more
frequent. There were a few Libellulids, and a number of
nymphs of Enallagma and Ischnura. The Hemiptera had
only one representative, Corisa, the water-boatman, which
was caught twice but was frequently seen swimming on the
clean surface of the pool. It can hardly be considered a
regular inhabitant of the alga-masses.

Four different larve of the order Diptera made up the
greater part of the insect population. Chironomus was
particularly conspicuous, since the larve were found con-

No. 588] POPULATION OF “BLANKET-ALGH ” 759

stantly, and were generally very numerous. Masses of
eggs of Chironomus cayuge Johannsen were found en-
closed in an oval mass of gelatin anchored to some of the
alge, also myriads of newly-hatched, almost microscopic
larve were seen, so it is reasonable to assume that, for
these pale pink or yellowish chironomus larve (1-18 mm.
long), this environment is the normal one. A few larger
species, some of them blood-red, were found also. Larve
of the ‘‘punkie’’ Ceratopogon and of the soldier-fly, Odon-
tomyia, were seen occasionally. Although mosquito-larve
are found regularly in stagnant pools, it is surprising to
note that only twice were these larve found among the
filaments of the floating alge. These larve were not
identified.

A few larval beetles and a few adults made infrequent
appearances. Undetermined Hydroporus and other dy-
tiscid larve were among these. Although known as a
dweller among filamentous alge, the Haliplid beetle larva,
Peltodytes, was seen only once, its long spiny hairs
tangled in the vegetation. _ Adults of two genera of Hy-
drophilid beetles were identified as Helophorus and
Crenophilus and a few other diving-beetles were seen but
not identified.

Although tadpoles, and once a young salamander, were
found in the collections, they can hardly be reckoned as
members of the society under consideration.

Dominant Forms.—In this diverse population the con-
stant and abundant forms have been few. Spirogyra,
especially Spirogyra varians, Mougeotia and Zygnema,
were the principal constituents of the ‘‘blankets.’’
‘Among the Diatoms, the dominating forms were Cocco-
nema, Navicula, Gomphonema and Synedra. Other alge
were best represented by Closterium, Dictyospherium
and Dinobryon. Among the animals Paramecium,
Euglena and the rotifer, Diglena, were quite constant.
The forms that appeared most regularly were Cyclops
and the larve of Chironomus. Some of the less constant
forms showed the influence of seasonal variation.

760 THE AMERICAN NATURALIST [ Vou. XLIX

Seasonal Variation.—In the autumn and early winter
Vaucheria was usually present, but appeared only twice
in the spring. Pandorina and Peridinium also appeared
late in the year. At that time fewer protozoa were seen
than in the spring, but, as has been said, variations here
seem to be more closely related to the character of the
water than to the temperature. Gammarus and the
nymphs of may-flies and dragon-flies were most numer-
ous in October, November and December.

The spring season also had its special forms. Oscil-
latoria appeared first in March and was constant there-
after. Diatom production was at its height in April and
May at water-temperatures varying between 8° and 16°
C. and there was a marked decline in diatom appearances
toward the end of June. In contrast to diatoms, desmids
seem to require higher temperatures, since most of the
Closterium and all of the Cosmarium and Staurastrum
that were seen appeared in June, in water at temperatures
between 15° and 20° C. The proportion of Dinobryon
in collections became noticeably greater during the latter
part of June. The smaller crustaceans, excepting the
ever-present Cyclops, showed marked increase in num-
bers as well as in diversity during May and June. The
Same seasonal increase was noticed for Anguillula and
the rotifers. Most of the coleopterous and dipterous
larvæ were found in May and June, except Chironomus
which was present at all seasons. °

Another point of interest in connection with seasons is
the time of reproduction. Spirogyra was found conju-
gating in October, April and June; M ougeotia in Novem-
ber, December, May and June. Young, sessile plants of
Ulothria were seen in April and May. All through the
year, copepod nauplii and female Cyclops bearing egg-
Sacs were observed. Chironomus eggs were found in
April and early in June, while very young larve were
abundant during April, May and June.

In view of the fact that floating alge were found in
large quantities in December, even under ice, it was sur-

No. 588] POPULATION OF “BLANKET-ALGA” 761

prising to find some of the pools totally devoid of this
kind of vegetation in spring. Pools K, M, 2, x', x, x, x*
showed this peculiarity. Their ‘‘blanket-algw’’ did not
reappear until May. This disappearance of surface
vegetation may have been due to spring freshets, as the
pools mentioned are in the flood-plain of Cascadilla Creek,
although not in the stream-bed.
_ The Natural Balance.—Like other societies, the popu-
lation of the ‘‘blanket-alge’’ has its producers and its
consumers, its hunters and its hunted, each readily ex-
changing rôles as occasion demands. The synthetic or-
ganisms include with the phytoplankton a few chlorophyl-
bearing organisms of the zooplankton; that is, forms like
Euglena, Phacus and Distigma, which, in sunlight, have
the holophytie method of feeding (Stokes, 1895). Dia-
toms require nitrates, silica and some salts to make their
dainty and beautifully marked shells. Since they are
comparatively heavy, they sink slowly, but are brought to
the surface during the spring and fall circulation of the
water. In spring they multiply rapidly near the surface,
since they need oxygen and sunlight.
_ Many of the tiny creatures, including ciliates, Clado-
cera, rotifers and nymphs and larve of some insects are in
search of diatoms. These animals eat other tiny food
particles as well as diatoms. The rhizopods, Arcella and
Ameba, ingest diatoms, desmids, small protozoans and
even rotifers. Vampyrella consumes the cell-contents of
alge. Actinophrys prefers the spores of alge, but takes
small protozoa. Actinospherium is omnivorous (Stokes,
1895). ‘Many of the ciliates eat diatoms and other ciliates.
The food is drawn into the oral opening by means of cur-
rents of water which are directed toward the opening by the
constant motion of cilia. One ciliate, Chilodon, has a pe-
culiar method of feeding. It protrudes a broad flexible
lip-like expansion of the anterior end and gathers up
food particles with a sweep of this organ.

Turbellarian worms feed on rhizopods, ciliates and —
rotifers. Rotifers eat diatoms and some nibble alge,

762 THE AMERICAN NATURALIST [Voi XLIX

whereas the closely related Gastrotricha, Chetonotus,
eats minute particles of decayed animal and vegetable
matter, rarely taking diatoms.

The smaller crustaceans in general and the snails are
scavengers, removing decaying algæ and bits of dead in-
sects or other animal matter. The Cladocera, however,
are said to eat diatoms and many of the smaller algæ.
Ostracoda are omnivorous and often attack their own
species.

Among the insect members of this society, the larvæ of
the may-flies and midges are the great herbivores, al-
though, in addition to algæ, diatoms and leaves of higher
plants, consuming a great variety of vegetable sub-
stances, both living and dead. The great abundance of
Chironomus larvæ make this genus an important factor,
both as a consumer, and as food for other animals.
Chironomus larvæ and pupæ are, in turn, eaten by dragon-
fly nymphs, and other predaceous larvæ. They are of
much importance as fish-food.

Dragon-fly nymphs are predatory. Some species eat
back-swimmers and water-boatmen, small crustaceans
and snails, coleopterous and dipterous larvæ and even
young dragon-fly nymphs. The larger nymphs are eaten
principally by fish, occasionally by water-birds.

This brief account of some of the feeding-habits will
serve to show how much all the members of this society
are dependent upon the others, and, at the same time,
are in constant danger of extinction. Each form acts as
a check upon too rapid multiplication of some other form.
Since the most prolific animals in this population are
Cyclops and Chironomus, each must have peculiarities
that enable it to survive in this environment and to com-
pete with other animals. Cyclops adapts itself easily to
changes. Its prolific reproduction, seasonal constancy,
and plasticity, give it great advantage over other small
crustacea. Chironomus, also constant, prolific and adapt-
able, finds abundant food and comparative shelter among
the algal filaments.

SHORTER ARTICLES AND DISCUSSION
ON PRACTICAL VITALISM

In a series of critical and polemical essays, published during
the past few years in American journals by diverse authors, par-
ticularly by Jennings, the problem of vitalism has been dis-
cussed in a manner that may seem exhaustive.

There would appear to be no possibility of adducing new ar-
guments in the matter. If in spite of this a new presentation
is here attempted, it is because the author holds a standpoint
entirely divergent from what has been thus far set forth in the
discussion.

If it is true that the argumentation of the promoter and
leader of the new scientific vitalism—Driesch—becomes at times
somewhat metaphysical, it appears to me also that the criticism,
as made by Jennings, tends at times to become dialectical and
sophistical.

I can not otherwise characterize the tendency to efface any
specific difference between the living and the non-living. By
isolating at random a feature of the living and comparing it
with an inorganic model one can indeed seem to show the iden-
tity of the two. But in this procedure we recognize the typical
method of the ancient sophists. I can find nothing of interest,
= for example, in such an argumentation as the one cited below
from Jennings.’

In a rejoinder to Lovejoy, who insists ‘‘that the same phenom-
ena occur in a given organism in spite of profound modifica-
tions of the composition and configuration of the parts’’ Jen-
nings objects that we have here

a proposition that holds for things in general. An iron body of a certain
form moves toward the earth. We may change the form in most varied
ways . . . change the material, substitute lead, brass, stone . . .; it still
moves toward the earth.”

Nothing is easier than to prove that black and white, plant
and animal, man and monkey, are ‘‘fundamentally the same.”’’
1A typical one for the antivitalistic criticism.
2 AMERICAN NATURALIST, 1913, p. 395.
763

764 THE AMERICAN NATURALIST [ Vou. XLIX

But does an affirmation of this sort annihilate in any way the
specific difference between man and monkey, or diminish the in-
terest of science in this specificity ?

The innumerable attempts of the critics of vitalism to prove
by comparison of certain isolated features that the living is
nothing more than an extreme complication of the non-living
fail, because the analysis in such cases is never exhaustive. One
may prove that living and inorganic coincide in many points;
he can not prove more.

I do not see why these points of coincidence are of more im-
portance and interest for our conception of the matter than the
points of undeniably distinctive = even though the latter
are as yet unanalyzed.

The best way to test the validity of an idea or hypothesis is to
follow it to its most extreme but logically inevitable consequences,
taking these as a statement of the proposition involved.

If we follow this method in order to obtain an objective and
exact formulation of the essence of vitalism (or of its antithesis,
mechanism), we can say that what mechanism asserts is this:
Whenever a certain configuration of matter occurs or is given,
there also what we call ‘‘life’’ is found; or in more popular
terms, the artificial production of a living organism from ‘‘non-
living’’ matter would theoretically be possible.

Vitalism, on the other hand, is a standpoint that in last in-
stance denies such a possibility.

It is clear that both the assertion and the negation are un-
provable, and as such are matters of faith, not of emprical
science,

If one attempts to give an estimate of the two from the stand- —
point of science, sympathy must, it appears to me, incline to the
vitalistie view, since scepticism is the very palladium of exact
science.®

It is generally overlooked that if one of the two opponents is
to be reproved as aggressive, that one is the mechanist rather
than the vitalist. The mechanist in asserting that he knows
more than can be proved is filled with a scientific optimism of a
somewhat frivolous character.

Yet it is the moderate agnostic standpoint, declaring no belief

3 It may be objected that a negation is dogmatic to the same degree as an
assertion. This may be true. But one can ropie the term ‘‘negation’’ by

some other less radical expression, such as ‘‘doubt,’’ without altering the
essence of the standpoint.

No.588] SHORTER ARTICLES AND DISCUSSION 765

in the possibility of artificial synthesis of the living so long as
that is not proved, that is subjected to ridicule as a dogmatic,
obscurantist and non-scientific doctrine.*

The entire problem to me falls in the domain of ‘‘ Natur-philos-
ophie,’’ that branch of our knowledge which can not directly
prove the truth or logical necessity of the results of investiga-
tions made in its field; can do no more than to make them plaus-
ible; and thus give to us a genuine sensation of mental satis-
faction.

There is no intention here of participating in the endless dis-
pute above sketched; I do not know what could be added in this
direction, from a vitalistic point of view, to the formulations, of
Driesch.” Our purpose is the defense of the right to a prac-
tical vitalism, as a method of exact empirical (although not
necessarily experimental) investigation.

e do not care whether the methods demanded by such a
vitalism are or can be proper also for inorganic investigation.

It appears that they are not, for the mechanists oppose their
‘‘veto’’ in the name of exact science to all constructions of the
vitalistie system, even though not fully analogous with that
which will be detailed below.

Practical vitalism claims the right to be restricted in formulat-
ing hypotheses only by postulates of logic and of the general
theory of knowledge, and by nothing else.

4 The same point can be made with regard to the other aspect of vitalism
—the so-called ‘‘experimental indeterminism.’’ As to this, it must be ad-
mitted that the empirical evidence seems to favor the vitalistic standpoint.
The assertion of the mechanists, that experimental indeterminism can not
hold for the living, is likewise a matter of faith, and the burden of proof
falls upon those who make it.

5I must nevertheless confess, despite my profound admiration for
Driesch’s work, that nd that his chief experimental foundation of vital-
ism, by means of his masterly analysis of certain cases of regulation, fails
to produce the desired effect; chiefly because the entire argument rests upon
certain experiments that are, as one may say, a lucky chance in biological
investigation. It would be quite possible that no organisms having the mar-
velous powers of regulation and equipotentiality shown by Tubularia, the
sea-urchin or Clavellina, should ever be discovered. Can it be admitted that
a scientific proof of vitalism as the basis of biological research would there-
fore remain inaccessible? The argument in such a capital problem must, I
think, rest on a more general basis, one resulting from an adequate analysis
of essential and genuine vital phenomena. I incline therefore to consider
Driesch’s further analysis, as presented in his ‘‘Science and Philosophy of
the Organism’’ as a no less valuable part of his work. :

766 THE AMERICAN NATURALIST [ Vou. XLIX

We hold as a justified demand of the theory of knowledge
that every hypothesis must be fruitful; that is, it must give a
number of deductions that can be verified empirically. Every
hypothesis which permits us a prediction is to be considered a
step in the progress of knowledge, until such time as it is re-
placed by a new one, more suitable or more fruitful.

Biological, and particularly embryological, investigation needs
sometimes to introduce as a hypothesis for the explanation of
certain empirical facts the idea of so-called ‘‘immaterial’’ (or
in Jennings’s terminology ‘‘non-perceptual’’) factors.

This is the chief point on which are based the recriminations
of most critics of vitalism, especially of Driesch’s vitalism.

The belief in such ‘‘non-perceptual’’ factors is in Jennings’s
mind synonymous with obscurantism or dogmatism. To Ritter
‘‘the vitalism . . . is the belief that organic phenomena can not
be fully explained by referring them to the material elements of
which organisms are composed, but that something not really
belonging to the natural order [?] ... is present in living
things’’ (italics mine).

si me it is entirely obscure why the term ‘‘non-perceptual fac-
tor,” employed by Jennings in a logical and consistent manner,
is by him rejected as nonsense.

His formulation of the non-perceptual is very clear.

Conditions subject to diverse physical tests will here be called either
perceptual or physical.®

A non-pereeptual agent would be one which though producing at a
particular time a particular physical event, was not subject to other
physical tests for its presence.”

I have given a formulation much resembling this, of what I call
the ‘‘immaterial factor’’ in my paper bearing that title, printed
in the ‘‘Festschrift fiir Schwalbe.’’

This work was to have been published the first of August,
1914, at the very moment of the outbreak of the war. Whether
it has been issued I do not know.

My definition is as follows:

Als materiell gilt uns im allgemeinen ein Objekt unserer Erkenntniss
welches eine Mehrzahl von einander unabhängiger Eigenschaften (se.
Wirkungsweisen) in sich vereinigt und sowohl in Tätigkeit als in Ruhe
befindlich wenigstens gedacht werden kann.

6 Johns i University Circular, 1914, No. 10, p. 8.

7 Ibid., p

No.588] SHORTER ARTICLES AND DISCUSSION 767

Ein zur Erklärung bestimmter Wahrnehmungen ersonnener Factor
von dem eine derartige Annahme d, h. ein Zustand der Nichtbetätigung
widersinnig wäre mag folgerichtig als “nicht materiell” bezeichnet
werden.

Neither Jennings’s ‘‘non-perceptual’’ or my ‘‘immaterial’”’
can be considered an illogical or contradictory conception.
Criticism must, so far as my own doctrine is concerned, be there-
fore concentrated solely on the strength of the empirical founda-
tion for the hypothesis of immaterial factors in any given case.

Besides the logical definition given above, an examination is
required of the question; What exactly can be meant by, or how
can one be led to assume, an ‘‘immaterial factor” as a result of
experimental investigation, or at least as a hypothesis impelled
by such a result?

To Jennings the assumption of a ‘‘non-perceptual agent’’
leads directly to, or is synonymous with, the so-called ‘‘experi-
mental indeterminism,’’ as admitted by Driesch.

He seems to neglect every other possibility of the action of an

‘‘immaterial factor.’’ I do not see that this is inevitable.

To me the essential point of the problem lies in the question
of the ‘‘bearers’’ for any sort of empirically detectible action
(induction, force, or the like).

Suppose that it were found that the factors directing the
movements of a given element of a living organism (for example,
the cell of an embryo), in a given direction m to the point n, lie
outside itself.

e will then assume at the point n a center of forces.

Suppose now that we can deduce from this assumption certain
consequences that will be subsequently verified empirically.®
Our assumption that gives us possibilities of prediction becomes
then a scientific reality. We say ‘‘reality,’’ although it may
remain somewhat hypothetical. We find the same condition of
affairs in the imperceptible but strongly inferred realities of
physies, ete.

8 I find that there is a point at which Jennings’s conception of the ‘‘non-
perceptual’’ seems to lead us wrong. It is well to say with Jennings that
such an agent is one producing at a particular time a particular physical
event but not subject to other physical tests for its presence (italics mine).
But Jennings seems not to take into consideration that a ‘‘ particular phys-
ical event’’ or a ‘‘single mode of action’’ (in my formulation) can lead to
many empirically verifiable consequences.

768 THE AMERICAN NATURALIST [Vou. XLIX

Suppose now that at the point of space where we have pro-
jected the center of forces there lies some element of the embryo,
such as a cell. The scientific routine will call this element the
‘‘bearer’’ of the forces in question.

But it is also possible that no element and no matter is to be
found at this point.

The first impulse will be to search for some other element of
the embryo, situated elsewhere, that can act as such a center, by
irradiating certain ‘‘lines of force,’ which influence in some
manner the movements of the first considered element. We will,
however, assume a case where no such element acting at a dis-
tance can reasonably be supposed. What now?

If the fundamental assumption holds true, that the factors
determining the movements of the element lie outside of itself,
we find ourselves confronted by the following alternative:

Hither the presumed factors have a bearer that is not cogniz-
able, or they have no material bearer at all!

It is clear that if we deny the existence of bearers that are
evidently perceptible, we can also exclude the possibility that
such bearers exist, but are invisible owing to their minuteness;
for the presumed center of forces lies according to our assump-
tion outside the organism; or in a district of it where there is
no formed embryonic matter at all.

Thus under the circumstances our two alternatives signify
the same thing, for to say that there is a bearer of factors
that is cognizable solely as the factors themselves involves
a tautology; an assumption of the sort so well character-
ized by the French as a ‘‘hypothése gratuite.’’ While any one
is free to make such an assumption, no scientific use can be made
of it. Methodologically it is perhaps comparable to Kant’s
‘‘Ding an sich,’’ which likewise must remain without empirical
content.

As a fundamental postulate of biological (and especially of
pele ot al research, there can therefore arise the concep-

ton of factors which, although spatial and localized in space,
e no material bearers, and as such may be denominated im-
material.

Is such an idea indeed nonsense; something that proves the
obscurantism of its promoter ?

I am well aware that the ‘‘immaterial’’ factor here presented
is far from coinciding with Driesch’s Entelechy or with any

No.588] SHORTER ARTICLES AND DISCUSSION 769

analogous agent that is e definitione not solely immaterial, but
also non-spatial.

If the entire weight of antivitalistic criticism is directed and
concentrated wholly against such ideas as that of Entelechy;
and if the mechanist will agree with me that a spatial localiza-
tion of a center of forces may be assumed without necessarily
combining this with a material bearer, I shall be much gratified.
But I fear that this is not the case. The ‘‘dynamical prefor-
mation of the morphe,’’ as I have elsewhere called the imma-
terial but spatial factors of morphogenesis,® must, I fear, fall
under the same anathema as the classical vitalism.

To resume the chief postulate of my own ‘‘vitalism’’: if mor-
phogenetic investigation is led in a rigorous inductive way to
assume a spatial factor at a definite point inside or outside the
asc no difficulty or contradiction or nonsense arises if no

‘embryonic’’ matter, or what is the same, no material ‘‘bearer’’
for this factor can be found at that point. Yet of course no
one can be prohibited from forming any sort of hypothesis as
to such functionless bearers, It may be even a psychological
necessity to form such hypothesis, for we love a ‘‘Ding an sich.’’
But such will form no part of empirical research.

The right to work with such immaterial factors, and in the in-
ductive way set forth above, is, for me, the essence of practical
vitalism.

We have now to examine consequences and postulates de-
rived from our fundamental assumption, which seem to present
very great difficulty. If we admit a dynamical factor localized
in space but not derived from a material bearer, it will be asked,
whence comes and how arises this factor?

The question of causation is based on a postulate of knowledge
that can not be eluded; it must be answered in some manner.

I will attempt to point out briefly how one can think the
origin or evolution of such an immaterial morphogenetic factor,
although it must be insisted that we have here a problem which
does not stand in immediate connection with the purely empir-
ical method of investigating the factors considered, so to say,
per se in their activity.

I see no difficulty in assuming an immaterial causality; that
is, the arising of an immaterial factor having a certain property

9 Biologisches Centrablatt, Bd. 32; Archiv. f. Entwicklungsmechanik, Ba.
39; Festschrift fiir Schwalbe, 1914

770 THE AMERICAN NATURALIST [ Vou. XLIX

(for example, configuration) from another less complicated im-
material factor, and so on.

The chain of immaterial factors could in this manner logically
be pursued backward to the beginning of the embryogenesis, or
to the egg.

As to the relation of such immaterial factors to Driesch’s
entelechy, they can be ranged solely in the category of ‘‘means’’
(Mittel) of the latter for the purpose of morphogenesis.

But I repeat that this is for me a matter belonging for the
present not to experimental investigation, but to the domain of
‘“Naturphilosophie.’’

If it appears as if I agree in this point with the ‘‘standpoint
of radically experimental analysis’’ of Jennings, this is not
really the case. The latter author seems to reject all that does
not belong to experimental investigation.

I think, on the contrary, that vigorously logical considerations,
deductive and even inductive, on the given empirical data form
a legitimate and integral part of our science of nature.

A. GURWITSCH

THE WOMAN’s UNIVERSITY,
PETROGRAD,
April, 1915

INDEX

NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS.

age Hen’s Egg, F. E. CHI-
TER
Albino ‘Series of Allelomorphs in

“TAI ket,’? of Freshwater
ia The oly arg oe the,
MILIE LOUISE PLATT
Allelomorphs, Multiple, the ‘Signi
ca of, oes E. es and

H D. FisH, 88; i "0. C.
LITTLE, 122; : The Atbing Series of,
in Guinea-pigs, SEWALL WRIGHT,

140; and Mice, T. H. MORGAN,
Er Fossil, Some Recent Stud-
ies on, Roy L. Moopie, 369; Coal
Sauk me Tage Crossopterygia,
Roy 63
Amphinixis tea Variability, L R.

WAL 6
Aa Mutationist, R. Rue-
TES, 645

Asterias tenuispina Lamk. at Ber-
muda, On the Number of Rays
W. J. rE T.

Asyhmetsy, A

Stud y of, as devel-
oped in the Psa and Families

of Desni Crinoids, AUSTIN H
CLARK,

er wea, HARLEY Harris, Muta-
tion en Masse, 129

Bean, Seed, The Influence of Posi-
tion in the Pod u upon the Weight
of the, J. ARTHUR Harris, 44;
Common, Inheritance of Habit in,

N B. Norron, 547
BELLING, Jo: On the
egregation of Genetic Factors
a Plants, 125; The E

ose Varieties of de Vries,
319; Linkage and Semi-Sterility,
582

Bermuda, On the Number of Rays
in Asteria nr Spe Lamarck at,

Bilaterality in Vertebrates, The
Origin of, A. C. EYCLESHEIMER,
504

Black-eyed White Spotting in bene
The Inheritance of, ©. C.

727

BLAKESLEE, A. F. and D. E. War-
77

NER, Correlation between Egg-

of Freshw
ools, Population of the,
EMILIE LovIsE PLATT, 752

ayester amas the Percentage of Re-
rom Incomplete Data,
W. J. SPILLMAN, 383
CALKINS, GARY N., Cycles and
Rhythms and the Problems of
‘‘ Immortality’? in Paramecium,
6
Mr. Muller

CASTLE, W., E., on the

- Mass Deena 713;
aad Th e Eng-
lish Rabbit and ee. Hinata of
Mendelian nese gar oat Con-
stancy, 23;
Black-and- Tan Rabbit and e
Significance of Multiple Allelo-
morphs, 88

Cat, chad ‘Tortoiseshell, PHINEAS W.
Wu HITING, 518

Girant “showing Sex-linked vent
delian Inheritance, Seventee
Years Selection of, cht

“re Origin of, as

PEARL,
Characters,
observed Fossil and Living
F

Chromosome, of ash geri Another
Gen the Fourth, RED A,
Hoge, 47; View of ‘Horedity . and
Its Meaning to Plant Breeders,
E. M. East,

A sTIN H., A Study o
Asymmetry, as developed in the
Genera and Families of Recent
Crinoids, 521

Coal Measures Amphibia and the
1

T12

Crossopterygia, Roy L. Moopiz,

Coc T. D. A., Diptera from
the § Seychelles, 251; Specific ani
Var iets Characters in Ann

Su aan
Correlation Toiora Egg-laying Ac-
Yellow spom in the
iien Fowl, A. F. BLAKESLEE
nd D. AR “360.
Correlations, Value of Inter- annual,
J. ARTHUR Harris, 707
RE pae S The meen
of the Best Value of the, fro
, p Set of Data, F. da E.
CUa OR
Crinoids, Recent, A Study of Asym-
metry, as developed i in the Gen-
era and Families of, AusTIN H.
CLARK, 521
Crossing, Over, None in the Female
of the Silkworm Moth H.
STURTEVANT, 42; Modification of
Characters by, R. RUGGLES GATES,
562

Crossopterygia, and Coal or

Am mobs, Roy L. Moopiz, 637
CROZIE w. dij the pn as of
Rays i in in Asterias tenuispina Lamk,

hagis pee Rhythms and the Prob-
lem of ‘‘Immortality’’ in Para-
mecium, Gary N. CALKINS, 6

Davi ; , BRADLE Y Moors, Additional

0
the Probable Top of Œnothera
Lam

a,

Dvr, ARTHUR, AEE Evolu-

pe and the Origin of Species,
149

Piskent? New Standard,
Gene e aa in, a H.

SHUL
ari from n e Seychelles, T. D.
A. COCKER Bis

Doubleness, Tnhorita e of, in Mat-
and Poteet. HOWARD B.
Fros'

Drosophila Another g in esa

me
A. Hogs, 47; ampelophila, A Pe.
culiar Mende lian Ratio in, JOSEPH
Lirr, 97; The Origin of a New
Eye-color. in, and Its Behavior in
Heredity, Ros
A wing Mutation in a S
cies o E.R. Hype, 185;
Mutations in Two Species of, c.

THE AMERICAN NATURALIST

coe R. Hypz, 183;

[ Vou. XLIX

W. and B. S. Merz, 187; repleta,
A Sex-linked Character in A.
STURTEVANT, 18 ; The Infertility

zation of a Sex-linked Mendelian
Character in, T. H. MORGAN, ;
Note on the Gonads of Gynan-
i dg of, F. N. Duncan,
Dunc F. N, A Note on the
Gon ab of Gynand
eroana pepa,
Attempt produce aa
through Hybridization, 575

Tg

HE. F. L. and G-U., ¥., The De-

termination of the Best Value of

the Coupling-ratio from a Given

Set of Data, 127

Early Portrayals of Says Opossum,
Peleg on R. EASTMAN, 585

M., The Paani of

Selt sterility, 76, 712; The Chro-

Vie of Heredity and its

Fr E
z On the ee A
the Conditions which determ
or prevent the Entrance of the
foe into the, JACQUES

Egg-laying „Activity and Yellow
the Domestice Fowl,
Correlation between, A. F.
AKES and D. E. WARNER,

360
Enchytræus albidus,

Regeneration
Posteriorly in, é
EN Ww,

UNT, 495

L., Repulsion in
eat, 127

Environment, The Role of the, in the

Realization of a Sex-li nked Char-

acter in Drosophila, T. H. MORGAN,

Varieties of de

i gnificance of
rid fon Internal Conditions of the
Organism in, F. H. and E,
L. Sco

EYcLESH

EIMER, The Origin
of Bilaterality: i in » Vertebrates, 504
Eye-color, New, Origin Drosophila
repleta, and. its Behavior in He-
redity Roscoe R. HYDE, 1

No. 588]

F, Blend accompanied by Genic
H. H. LAUGHLIN, 74

urity,
Fecundity in the feaa n,
deli

of,

Average Flock Production, as
MOND PEARL, 306

Field Experimen nts, On a Criterion
of Substratum Homogeneity sa
Heterogeneity) in, J. ARTHUR
Harris,

1sH, H. D. and W. E. CASTLE, The
Black-and-Tan Rabbit and the
Significance of Multiple Alelo-
morphs, 88

Flock Production, Average, and
Mendelian Inheritance of Fecun-
ay in ~~ omestic Fowl, RAY-

MOND PEARL, 306
Flower Pigments, M. W., 256
and ivin Anim als

A
Amphibia,
Some. Mpo: Studies in, Roy L.
Moopiz, 3
Fowl, _Domestic, Mendelian Torpe
tance of Fecundity in and are
ago F “Flock icere RAYM a
; Correlation Sarees
Ege ls laying Activity and Yellow
Pigment > i F. BLAKESLEE and

6
Freshwater Tal Population of i
‘*Blanket of

e Inheritance
Paturs, I. The Hypotheses, 623
Gates, R. Rueetes, On the Modifi-
cation of Characters by Crossing,
tay a Anticipatory Mutation-

Genk eek in the Fourth Chro-

reg ee of Drosophila, MILDRED
,4
Genetic, Definitions in the New
Standard Dictionary, G. H. S i
52; Factors in Plants, On the
Time of Se on of, Jo

gr
Deer Mice, Francis B. SUMNER,

688
Genie Ribas a Tod Blend accom-

panied by, LAUGHLIN, 741
Germ Cells a oes tie Cells, LEO

LOEB.

Gonads of ewer ongs s of Dro-
sophila ampelophila, F. N. Dun-
CAN, 455

INDEX

773

Guinea-pigs, The Albino Series of
Allelomorphs in, SEWALL WRIGHT,

140
GurwitscH, A., Practical Vitalism,

Gynandromorphs of Drosophila am-
pelophila, Gonads of, F. N. Dun-
CAN, 455

Habit, Inheritance of, in the Com-
mon Bean, JOHN B. Norton, 547
HADLEY, PHILIP an W. E.
CASTLE, The English Rabbit and
t on of Mendelian Unit-
character Constancy, 23
ing, The Resemblance of
ins in, EDWARD L

Fi
hori of Inter-annual Correla-

tio 707
Heredity, The Origin of a New Eye-
color in Drosophila hg Spe and
R

183; The romosome View of,
and its Meaning to Plant Breed-
, E. M. East, 457
Hose. MILDRED A., Another Gene =
the Fourth Chromosome of Dros

47
YRON B., Sterility in a
252

. R., Regeneration Posteri-
orly in Enchytræus albidus, 495
E coE R

ity, 183; A Wing Mu
New Species of Drosophila, 185
tt Immortality’? in Parameci
Cycles and R
Problem of, Gary N.
Sabrane Studies in, RAYMOND

PEARL
Infertility of Rudimentary Winged
ales of Drosophila ampelo-
phila, T. H. MORGAN, 240
grw, itance, Mendelian, of Fecun-
dity in the fee eg Fowl, and
Average Fi Flock Production, RAY-
, 806; of Habit in the
` Common

acter showing, RAYMOND PEARL,

7174

595; of i get in Matthiola
and Petunia, Howarp B. Frost,
ot of Black: -eyed Birer Spot-
n Mic , 127
Inte annual Chavelations, F "ARTHUR
, 107
Titersal Conditions, Certain, o e
Organi Organic Evolution,
The Signifieance of, F. H. Pike
and E. L. Scort, 321

JEFFREY, EDWARD C., Some Funda-
mental Morphological Objections
to the Mutation Theory of de
Vries, 5

LAUGHLIN, H. H., The F, Blend ac-
companied by Ge
JOSEPH, Data on
Mendelian Ratio in Drobojhila
ampelophila,
Linkage and Semi- -sterility, JOHN
a , 582
E A Note on Lo s
pere eA in Mice,
Inheritance of ae -eyed White
Spotting in Mice,

727
LOEB, JACQUES, On the Nature of

matozoon into the Egg, 2
LOEB Bo Sea Cells and Bitte
Cells,

aey ee cn co Experiments
ay

Matthiola and Petunia aces tt

ess How. B.

The English Rabbit

Sio Ost, 623.
Mendelian, Unit-character Constancy,
and the Ques-
aan of, W. E. CASTLE and PHILIP

-li ee AN, 385
-, Mutations

pecies of Drosophila, 187
sae “ad Allelomorphs, T. H. Mor-

GAN, 379 ornia Deer, Genetic
Studies graphie

of, Francis B. SUMNER,
688; The ritance of Black-

THE AMERICAN NATURALIST

[ Vou. XLIX

eyed ‘aby Spotting in, C. C.

LITTLE, 7

Some Recent Stud-
ies on Fossil “Amphib ia, 369; The
oal Measures Amphibia a and the
Crossopterygia, 637

Morean, T. H., The Infertility of
Rudimentary "Winge Females of
Drosophila ampelophila, 240; Al-

rphs

OUGH, e sap ironed
of own Mutations in Othe
Mut an Stocks, 318
orm, ’ No a h ar

Moth, Silkw
in the Female of, A. H.
VANT

Muller, Mr., on the Constancy of
Mende = Characters, W. E.

CASTLE,
Multiple Aiie orphs, The Black-

: in Mice , C. ©. LITTLE,

122
Mutant mag ie The Appearance of
Known Mutations in Other, T. H.

; Wing, in a
mise er of Drosophila, Ros-

j pe a a
Additional 1 Bvidence of, Brap.

00
Matetionist, Anticipatory, R. Ruc-
GLES GATE

Mutations, in rwo Species of Droso-
phi we C. W. and B. S. Merz >» oats
Kno a Aoi in Other Mu-
vee Stocks, T H M ORGAN and
HAROLD 31 n At-
tempt to Produce, reen Hy-

bridization, F. N. Duncan, 575

NORTON, Tpi B, y stern of
Habit e Common Bean, 547
Notes ond These "32, 127, 251

(Enothera Lamarckiana, Professor
de Vries on the Probable Mie a
of, Y Moore Davis, 59;
Additional Evidence of Mutation
in, BRADLEY Moore Davis, 702

Oran, Early Portra: rayals of,

R. EASTMAN, 585
Ges. The Significance of Cer-

No. 588]

tain Internal Conditions of, in
wii oa porna F, PIKE
d E. L. Sco , 321

Origin: of Boodlea, "i ecb. rg
Evolution, ARTHU
of a New Bye-color in " Drosophila
repleta and its Behavior in He-
ora Roscoe R. HYDE, 183

OsBo HENRY FARFE FIELD, Origin
of ‘Si ingle Characters as observed
in Fossil and Living Animals and
Plants, 193

Paramecium, Cycles and Rhythms
and the Problem of ‘‘Immor-
tality?’ in, Gary N. CALKINS, 65

PEARL, RAYMON n Inh r-

ERE of Fecundity i in a Domes-

owl, and are Flock Pro-

d-
s Selec-
ring -

nee, 595
mheritance

oubleness in: sg

Pudar 62

IKE, F. H. and E. L. Sco The

of Certain Taternal

a Heredity and its Mean-
ingt EAST, 457
LATT, p aie Lovise, The Popu-
lation of the “Blanket Alge’’ of
Freshwater Pools,
ae HAROLD and T e pager
e Appea a of K
tations in Other Mutant "ae

Practical Vitalism, A. GuRWwITSCH,
762

~~ The English, and the Ques-
of Mendelian Unit-character

Coney, STLE and
Pup B. H , 23; The Black-
and-Tan, and the ificance of

CasTLe and H. D. FISH, 88
Rays in Asterias tenuispina
at nia Number of, W. J.

e s re Method of Calculating
the Percentage of Eaa Incom-
plete Data, LMAN, 383

Regeneration Patis in Enchy-
træus albinus, H. B. Hunt, 495

gee oma in "Wheat , F. L. ENGLE-

, 127

ianiai of Young Twins in

INDEX

E.
Semi- -sterility,
, 582

775
Handwriting, EDWARD L. THORN-
DIKE, 377

Scorr, E. L. and F. H. Pixs, The
Significance of Certain Internal
Conditions of = Organism in Or-
ganic Evolution, 321

Segregation of "Genetic Factors in

On the Time of, JOHN

Be a 125

Selection, ’ Sugar- -beets — gprs
W. E. Castiz, 121; Some
Experiments in, W. E Tana,

Self-sterility, The Phenomenon of,
. East, 76, 712
and Linkage, JOHN
BELLING
Sexlinked, Character in ene TF
repleta "A. H. STURTEVAN 189;
Waridi Inheritance, Bewcubail
Years Selection of a Character
s

owing, RAYMOND PEARL,
Seychelles, i i fiom, T. D.

COCKERELL, 2
T Articles and Discussion, 121,
, 318, 455, 518, 570, 645, 702,
rts
SHULL

, G. H., Genetic ipa in

the New Standard Dictio ary, 52

Somatic Cells and Germ Cells, LEO
LOEB

Species — Sterility in, Byron
B. Horton, 252

Spermatozoon, On the Nature of the
Conditions which determine or
prevent Ponin into the Egg,
JACQUES ;

SPILLMAN, W. J., A Method of Cal-
culating the Percentage of Re-
cessives from Incomplete Data,
383

gora Black-eyed White, Inher-
of, in Mice, C. C. ' LITTE ITTLE,

Sterility, Self, The Phenomenon of,

E. M. EAST, 76; in a Species

B. Horton,

Moth, Sex-linked
Character in Dreaophiin repleta,
Substratum Homogeneity (or Het-

erogeneity) in Field Experiments,
J. ARTHUR Harris, 430
Tibia Selection and Thrips,
W. E. CASTLE
SUMNER, FRAN ae B., Genetie Stud-
ies of Races

Several Geograp hic
of California Deer Mice, 688

776

Sunflowers, Annual, hee and
Varietal. Characters in, T. D. A.
COCKERELL, 609

THORNDIKE, Epwarp L., The Re-
semblance of Young Twins in
Handwriting, 37

Thrips, Sugar-beets and Selection,
W. CASTLE, 121

Tortoiseshell Cat, PHINEAS W.

ung, Resemblance =

andwriting, Epwarp L. THO

DIKE, 377

je repai and Amphiontxi li B.

N, 649
Vertebrates, The Origin of Bilat-
lity in, A. C. EYCLESHEIMER,

Vitalis, Practical, A. GURWITSCH,

Vries, Hugo de, Some Fundamental
Merpholewies! Objections to the
Mutation Theory of, Epwarp C.
J EFFREY, 5; on the Probable Or-

igin of CEnothera Lamar ckiana,

THE AMERICAN NATURALIST

[Vou. XLIX

BRADLEY Moore Davis, 5
Evening Primrose Varieties ey
JOHN BELLING, 319

W., M., Flower Pigments, 256

WALTON, L. B., Variability and Am-
phimixis,

Warner, D. E. and A. F. BLAKES-
LEE, Correlation between Egg-lay-

R
Wheat, Repulsion in, F.. L. ENGLE-

camel ard
WH PHINEAS W., The

teisoahall Cat, 518

RIGHT, SEWALL. e Albino Series
of Allelomorphs in Guinea- “pigs,
140

Tor-

and F. L. E., The Deter-

of the Best Value of
a peer bale from a Given
Set of Data, 127

Y, & v.