BIOLOGICAL BULLETIN

OF THE

flDarine Biological Xaborator\>

WOODS HOLL, MASS.

Staff.

E. G. CONKLIN — The University of Pennsylvania.
JACQUES LOEB — The University of California.
T. H. MORGAN — Columbia University.
W. M. WHEELER — American Museum of Natural

History, New York.

C. O. WHITMAN — The University of Chicago.
E. B. WILSON — Columbia University.

Efcitor.

FRANK R. LILLIE — The University of Chicago.

VOLUME X.

WOODS HOLL, MASS.
DECEMBER, 1905, TO MAY, 1906.

PRESS OF

•HE NEW ERA PRINTING COMPANY
LAN CASTER, PA.

CONTENTS OF VOL X.

No. i. DECEMBER, 1905 PAGE

ADELE M. FIELDE : The Progressive Odor of Ants I

E. H. HARPER : Reactions to Light and Mechanical Stimuli in the EartJi-

u'orm Perichceta Bermudensis (Beddard) 17

LULU F. ALLABACH : Some Points Regarding the Behavior of Me-

fridiuw 35

No. 2. JANUARY, 1906

R. R. BENSLEY : An Examination of the Methods for the Microscopical
Detection of Phosphorus Compounds other than Phosphates in the
Tissues of Animals and Plants 49

J. F. McCLENDON : On the Locomotion of a Sea Anemone (Metridiitm

marginatuni) 66

H. H. NEWMAN: The Significance of Scute and Plate "Abnormalities"

in Chelonia 68

No. 3. FEBRUARY, 1906

H. H. NEWMAN: The Significance of Scute and Plate "Abnormalities"

in Chelonia 99

T. B. ROBERTSON: Note on the Influence of Surface -Evaporation upon

the Distribution of Infusoria 115

VERNON L. KELLOGG : Histogensis in Insect Development, and Cell

Specificity 1 20

E. A. ANDREWS : Ontogony of the Annulus Ventralis 122

No. 4. MARCH, 1906

O. C. GLASER : Correlation in tJie Development of Fasciolaria 139

CHARLES S. ROGERS: A Chameleon-like Change in Diemyctylus 165

A. C. EYCLESHYMER : The Growth and Regeneration of the Gills in the

Young Nee turns 171

W. L. TOWER : Observations on the Changes in the Hypodermis and Cuti-

cicla of Coleoptera During Ecdysis 176

iii

iv CONTENTS

No. 5. APRIL, 1906

PAGE

ESTHER F. BYRNES : Two Transitional Stages in the Development of

Cyclops signatits, var. coronatus 193

T. H. MORGAN : The Male and Female Eggs of Phylloxerans of the

Hickories 201

CHAS. W. HARGITT : The Organization and Early Development of the

Egg of Clava leptostyla Ag. 207

PHILIP B. HADLEY : Regarding the Rate of Growth of the American

Lobster 233

T. B. ROBERTSON : Note on the Influence of Temperature upon the Rate

of the Heart-beat in a Crustacean ( Ceriodaphnia) 242

L. B. SEELY : Two Distomes 249

No. 6. MAY, 1906

INEZ L. WHIPPLE : The Ypsiloid Apparatus of Urodeles 255

FRANCIS B. SUMNER : The Osmotic Relations Between Fishes and their

Surrounding Medium (Preliminary Note} 298

Vol. X. December, 1905. No. i

BIOLOGICAL BULLETIN

THE PROGRESSIVE ODOR OF ANTS.

ADELE M. FIELDE.

I. STATEMENT OF HYPOTHESES BASED ON RECENT AND
FORMER EXPERIMENTS.

1. The Specific Odor. — The mother-ant transmits to her off-
spring the distinctive odor which is identical for ants of all ages
and of both sexes within the species. This odor is appreciated
among ants by organs near the proximal end of the funicle.1

2. Progressive Odor. — Female ants, including queens and
workers, have, besides their specific odor, an odor which may be
termed progressive. Queens of different lineage have different
progressive odors. In a queen this odor is either unchanging or
changes very slowly, and it is similar to that of her newly-
hatched female offspring.

a. As worker-ants advance in age their progressive odor
intensifies or changes to such a degree that they may be said to
attain a new odor every two or three months. This progressive
odor is appreciated among ants by organs in the penultimate
joint of the funicle.2

b. Male ants have no progressive odor unless it be super-
ficially incurred through association with workers ; but the male
carries latent in his spermatozoa the progressive odor of his
mother. In other words the progressive odor is always recessive
in the male ant.

c. The progressive odor of each new generation of females is

1 "Artificial Mixed Nests of Ants," A. M. Fielde, BIOLOGICAL BULLETIN, Vol.
V., No. 6, November, 1903, p. 320.

2 "Farther Study of an Ant," A. M. Fielde, Proceedings of the Academy of
+\ 'at nra I Sciences, November, 1901, p. 531.

2 ADELE M. FIELDE.

determined by the odor of the mother latent in her egg, and the
odor of the father's mother latent in the spermatozoon. The
progressive odor therefore changes in each generation of females.

d. The progressive odor manifest in female ants is the cause of
the separation of ants of the same species into hostile colonies,
and is of great advantage to the ants in their individual and their
communal life.

3. The Incurred Odor. — An ant may incur from its associates
an odor which is not inherent in itself, and which may be
removed by washing. It may be transferred from ant to ant
through air or through water. It arises from the substances
that give the specific odor and the progressive odor and that
create the nest-aura.

II. RECENT EXPERIMENTS WITH THE PROGENY OF A SINGLE
QUEEN, COMPONOTUS PENNSYLVANICUS.

In 1901 I found that the odor of working-ants of Stcnainma
fulvmn piccuin changes with their age,1 forty days being the mini-
mum of time in which there occurs a change so great as to effect
the behavior of ants of the same colony toward one another at
their first meeting.

In 1902 my further experiments indicated that a cause 2 for the
hostility of one colony toward another of the same species and
variety is a difference in odor coincident with difference in the age
of the colonies.

In 1904 my observations on several species of ants,3 represent-
ing three subfamilies, gave further evidence of their change of
odor with advance of age, and indicated that the odor of the queen
is unchanging, or that her odor changes much more slowly than
does that of the workers.

1 have now had under observation for more than two years a
colony of Caniponotus pcnnsylvaniats, in which the assertion of a

1('A Study of an Ant," Proceedings of the Academy of Natural Sciences of
Philadelphia, July, 1901, p. 449.

2 " Notes on an Ant," Proceedings of the Academy of Natural Sciences of Phila-
delphia, September, 1902, p. 609 ; " Cause of Feud Between Ants of the Same
Species," BIOLOGICAL BULLETIN, Vol. V., No. 6, November, 1903, p. 328.

3" Power of Recognition Among Ants," BIOLOGICAL BULLETIN, Vol. VII., No.
5, October, 1904, p. 244.

THE PROGRESSIVE ODOR OF ANTS. 3

progressive odor in the workers is definite and indisputable, the
five successive broods included in the experiment being the issue
of one queen.

The N Queen.- -This queen was captured on Nonamesset
Island, July 28, 1903. She was then deflated and was probably
the mother of the hundred workers seen in her wild nest, and
also of the ants that afterwards hatched from the many cocoons
brought with her to the laboratory. She remained under my
care and, unless another is indicated, she is the queen referred to
in the herein recorded experiments.

The Ni Group of Workers. — Some of the captured workers
were transferred to Dr. Irving A. Field, and they remained segre-
gated in his care, usually at Harvard University, until the time
of the experiment in which they appear. As no other than male
offspring had appeared in this group during the two years of its
separation from the queen mother, the workers composing it in
August, 1905,, were certainly acquainted with the queen pre-
vious to her capture in July, 1903. Of the age of these workers
of course nothing more was known than that it exceeded two
years.

On August 6, 1905, I introduced into this nest,1 where there
were six major and five minor workers and about thirty larvae
from their own eggs, the queen-mother from whom these eleven
workers had been separated for two years. The queen showed
instant hostility, seized a major worker by one of its mandibles,
braced herself on the sponge and held her prisoner there during
the ensuing seven hours. All the other workers, sometimes six
at a time, examined the queen meanwhile. They patted her with
their antennae, nabbed her gently, and licked her back and legs.
Two of them, touching her body with their antennae, appeared to
dance for joy, shuffling their feet with great rapidity during
several consecutive minutes. The queen then began to drag the
worker that she had seized, and upon my releasing the latter, took
a position near the larvae-pile, as if to claim her incipient grand-
sons as her exclusive property, opening her mandibles at every
worker who approached. Then followed a most curious and pro-

1 All the artificial nests referred to in this paper were of the Fielde pattern. See
"Portable Ant-Nests," BIOLOGICAL BULLETIN, Vol. VII., No. 4, September, 1504.

4 ADELE M. FIELDE.

longed effort on the part of the workers to placate the queen-
mother. They surrounded her at all times, offering her regurgi-
tated food. Whichever way she turned, there stood a humble
servant with a proffered mouthful of pabulum. As many as
seven workers simultaneously offered nourishment to her. Every
worker of the eleven seemed bent upon wooing and winning her,
and she was not for a moment left without attention. These
efforts were unceasingly continued, and were meeting with a fair
degree of success, when I removed the queen on the following

morning.

This experiment showed that the workers all recognized the
odor of their queen after two years of separation from it, and that
the segregated workers had during the same interval acquired an
odor unfamiliar to the queen, who had meanwhile met none of
her daughters who were over fifteen months old. It also showed
that major workers, having in this species nearly the same form
and sometimes nearly the same bulk as has the queen, are like
minor workers in having a progressive odor.

On August 7 I introduced into this Ni group a marked major
and a marked minor worker, daughters of the N queen, but
many months younger than any of the ants in this group. The
visitors were received with signs of curiosity, but with perfect
amiability, though no younger sisters had been encountered
within two years by these Ni ants. The odor of the younger
sisters was perfectly recognized by the eleven residents, and I
removed the former.

I then introduced a young winged queen of the same species,
Camponotns pennsylvanicus, but of an alien colony. The resident
ants attacked her instantly and with exceeding virulence. In an
instant she lost an antenna, one worker was pulling out her
remaining antenna, and three others were dragging her by her
legs. The scrimmage was fierce, and before I removed the in-
truder four of the residents had received injuries that resulted
in their deaths. The residents had given to the alien queen
a reception strongly contrasted with that accorded to their own
queen-mother ; while the havoc wrought by the alien queen
indicated that, if unable to escape from the nest, she might have
destroyed all the workers and have remained a fostering mother
to their larvae.

THE PROGRESSIVE ODOR OF ANTS. 5

T/ic JV2 group. — During the first week in August, 1903, the
queen deposited about a hundred eggs, and from these were
reared five minor workers, denoted here as the N2 ants. These
workers hatched between April 24 and May 10, 1904, and were
therefore some fifteen months old at the time of the experiment
here recorded. These workers had never met other ants than
those of their own segregated group, and were therefore unac-
quainted with the odor of ants in any wise unlike themselves.
They had never lived with the queen, had laid no eggs, and had
the care of no young. On July 16, 1905, I put these ants into
a new and very small nest where I had isolated the queen-mother
without young. The five workers were wholly at ease with the
queen, and hastened to evince their devotion in ant fashion; but
the queen opened her jaws whenever they approached her, and
was somewhat querulous in her behavior during the ensuing two
days. The queen had lived during the previous five months with
daughters, all minor ants, less than five months old, and her be-
havior indicated a difference in the odor of her younger and her
older daughters. Her memory was manifestly less tenacious
than that of the workers, who, on their part recognized in their
queen the odor that had been their own in their infancy, fourteen
months earlier.

The Nj group. — This group consisted of two major workers
hatched in July, 1904, and four of their younger sisters, minor
workers, over five months and less than a year old, all the issue
of the N queen. The two majors were acquainted with sisters
older than themselves, while the minors knew no sister older
than these two majors. The members of this group had all lived
with the queen, and had been separated from her and living in
segregation since February 14, 1905. They had deposited no
eggs, and they had the care of a few introduced larvae. On July
1 8, 1905, I introduced into this group a sister nine months older
than the oldest in the group. The majors, who had had acquaint-
ance with sisters much older than themselves, did not attack the
newcomer at all, while every one of the minors, never having
met a sister so old as was the visitor, attacked, dragged and
finally killed her.

It appeared that the behavior of the two major workers was

0 ADELE M. FIELDE.

dictated by memory, while that of the four minor workers was an
effect of hostility created by the presentation of an unfamiliar
odor. The major workers were either wanting in compassion,
or else they lacked means of communicating with their younger
sisters, for although they were each double the size of any minor
ant in the group, they did not interfere in behalf of the victim.

The Nj. Group. - - The queen was transferred without eggs or
young to a new nest on July 14, 1904. She laid no eggs there-
after until December, 1904, and from the eggs then deposited the
five minor workers constituting the N4 group were hatched
between February 19 and March 23, 1905. These workers were
therefore four or five months old at the time of the experiment.
On July 16, 1905, I removed the queen from the nest, leaving the
five workers in charge of twenty larvae, the issue of the queen's
December eggs. Into this group of five minors, who had
never met older sisters, I introduced one of the majors from
group IS 3, now just a year old, and twice or thrice the bulk
of any of the five residents. The introduced ant was instantly
and violently attacked by three residents. This attack indicates
that the major, like the minor ants, like in shape and size as they
are to the queen, change their odor with advance in age, as do
minor workers.

Having removed this visitor, I introduced a marked large
minor worker, fourteen months older than the residents, a sister
of theirs, hatched from eggs deposited by the queen in August,
1903. This visitor was likewise violently attacked, every one of
the five residents manifesting hostility to her, and the next day I
found her mangled body on their rubbish pile.

The N$ Group. — This group consisted of two minor workers,
the issue of the queen's December eggs, sequestered in their
cocoons and hatched on September 8, 1905. They were at
once placed in segregation in a new nest, with a few larvae and
cocoons from their mother's eggs. Ten days later these ants
drove away from their pile of young any member of the N4
group of sisters six months older than themselves.

In these experiments it appears that it is the age of the work-
ers, not the age of the queen at the time when she deposits the
eggs from which the workers issue, that determines at any date

THE PROGRESSIVE ODOR OF ANTS. 7

their progressive odor. All the ants engaged in this last men-
tioned experiment were certainly the issue of the eggs laid by the
queen early in December, 1904. That there is similar progress
in odor among ants of the same age and species is indicated by
an immediate and amicable association of ants that are reunited
after a period of separation so long as two years.1

Whether two mutually hostile groups could be created from
among the worker-progeny of a single queen would depend on
power of memory in the older workers. By segregating from
the pupa-stage the broods of different summers, it would be
found that the younger sisters would always be hostile to the
older sisters, because the older sisters would present an unfamiliar
odor to the younger. The hostility of the older sisters toward
the younger would be nullified by their memory of the odors by
which they had themselves been characterized at earlier periods
in their own lives. If the younger sisters bore an odor which
the older sisters, through the lapse of many years, should have
forgotten, then the hostility would become mutual. It is certain
that worker-ants can remember for years an odor with which
they have once become familiar, and it is probable that they
remember such odors as long as they live.

When ants of different groups meet amicably, either the mem-
bers of these groups have the same odor, or else they have at
some time in their lives been familiar with ants bearing the pre-
sented odor. If one group recognizes a familiar odor, while the
other group discerns a strange odor, then thos'e finding them-
selves among strangers will try to escape, or will make attack.
There is no love at first sight between ants of different odors.

III. THE ODOR AND THE SENSE OF SMELL IN MALE ANTS.

Male ants apparently bear a specific odor, beside the odor that
may be incurred during their residence with nurses in the home
nest. I have introduced males of different species into the nests
of Steitamma fu/vni/i, Cremastogaster lineolata, Mynnica ritbra,
Formica sanguinea, Formica Schaufnssi, Camponotus pennsyl-

1 " A Cause of Feud Between Ants of the Same Species Living in Different Com-
munities," A. M. Fields, BIOLOGICAL BULLETIN, Vol. V. , No. 6, November, 1903,

P-

8 ADELE M. FIELDE.

vamcns, Camponotus pictus, Camponotus americanus, Lasius
latipes, and Lasius umbratns, and all these males have invariably
been killed within a day or two. If hybridization is to be
effected among ants it will be necessary to cause the males and
females to become acquainted with one another within a few hours
after hatching. When hatched in the same nest, males of Sten-
ammafiilvum pursue queens of Cremastogaster liucolata with the
same ardor that they show in pursuing queens of their own
species. In my mixed nests the failure of individuals of these
two genera to mate was manifestly due to physical and not to
psychic incompatibility.

In the summer of 1905 I had material in my stock of ants for
experiments giving evidence that the male ant has at hatching
the specific odor of his virgin worker-mother. My E mixed
nest consisted of workers of Cainponotus pictus, Formica ncoga-
gates, Formica subscricea, and Stcnanima fulvinn, all hatched
during the last week of July, 1904, and kept in the same nest
until the first day of January, 1905, when the Stenammas were
segregated apart. They remained in segregation until August
22, 1905, when I put into their small nest, where there were ten
workers and a few eggs, a fine male Camponotus pictus, the
offspring of a virgin worker-mother who had shared the nest of
these Stenammas until she was five months old. This young
male was received by the resident Stenammas with evident pleas-
ure. They licked him, regurgitated food to him, and rode on his
back. He continued to live happily with them for many days.
He bore a familiar specific odor, although hatched among segre-
gated workers of his species, eight months older than any that
these Sfenammas had known ; and this familiar odor made him
welcome. His fate was in strong contrast to that of some of his
brothers or cousins introduced into another nest. At the time
of these experiments I had also a nest, marked D, of eleven
Stenamma fuhnun workers, that had hatched in a mixed nest
during the last half of August, 1903, and had lived for several
months with Camponotus pennsylvanicns, and Formica subsericea,
but had never met a Camponotus pictns. These eleven Stenammas
had lived in segregation since July 17, 1904, and were destitute
of young, when on August 22, 1905, I introduced into their nest

THE PROGRESSIVE ODOR OF ANTS. 9

a newly hatched Cainponotus pictJis male, the offspring of a virgin
worker, a brother or cousin of the one in the E nest above
mentioned. These ants of the D nest at once began to harry
him, and although he was eleven millimeters long and very
sturdy, while none of the Stenammas were more than five milli-
meters in length, they harried his life out within two days. Repe-
titions of this experiment gave similar results in every case.

The eggs from which these Cainponotus pictns males were
produced were deposited by their virgin worker-mothers in May,
1905, five months after the said mothers were separated from the
Stenammas that the said mothers had lived with during the first
five months of their lives. It therefore appears that the male
progeny of virgin workers have not the progressive odor which
characterized their mothers. The males have, however, a specific
odor, an odor recognized by the ants through certain joints of
the antennae, and this odor is doubtless the stimulus calling forth
the exceeding care given by the workers to young males with
whose specific odor they are familiar.

On August 26 I put into each nest, the E and the D nests
above described, two males of Stenamma fnlvnni. These males,
the first of their species ever encountered by these workers, wery
treated alike in the two nests. They were so eagerly grasped be
several residents at once that it seemed as if they must lose their
lives through the determined efforts of the workers to retain
them. They were not left free for several hours ; but so judi-
cious were their virgin captors that no injury was done to the
captives, and they lived in health and honor many days in these
nests. In the E nest the Camponotns pictns male continued to be
their associate.1 In both the E and the D nests newly hatched

1 A dealated queen Camoponolus pictus captured alone in the open on July 5, w
kept in isolation till August 15, 1905, when she received amicably into her small
artificial nest two young males of her species, the offspring of virgin worker ants.
She licked them, regurgitated food to them, and during the several days that they
remained under my observation, remained in close companionship with them. Later
on this queen also received in amiable fashion the virgin mothers of these males, the
worker-mothers having been kept by me in segregation during their whole lives. As
this queen was captured near the spot in which the workers had their origin a year
earlier, these ants may all have been of one colony. This queen killed young males
of Formica argentata and Stenamma fitlvum introduced into her nest.

IO ADELE M. FIELDE.

Stenamma workers, from the same colony as were these males,
were immediately killed.

Since the males avoid, or are indifferent to, ants of other spe-
cies than their own, unless hatched among such species in arti-
ficial nests, it appears probable that they discern the specific odor
of other ants. But they probably lack the sub-nose that per-
ceives the progessive odor of workers. Male ants of various
species placed under observation in one of my artificial nests,
grouped themselves according to species, but did not quarrel
with males of species unlike their own. I infer that the only
inherent odor of males is that of their species ; but that they are
the medium through which the progressive odor of their female
progenitors is transmitted to the egg that produces a female, the
progressive odor being latent in the males and reappearing in
their female descendants. Only the egg receiving a spermato-
zoon would produce a female, and this female would be endowed
with her paternal grandmother's tendencies in progressive odor,
the progressive odor thus manifesting itself only in the female
line of descent. The fact that the worker progeny of a queen,
sequestered from the pupa stage, will receive their queen-mother
or the queen-mother's sister with equal pleasure, indicates simi-
larity of odor in the product of the same queen's impregnated

eggs.

I venture to predict that there will be found in female ants
secretory glands or organs that are wanting or are rudimentary
in the male, and that these organs are the producers of the pro-
pressive odor. There must be in both males and females secre-
tory glands or organs producing the specific odor which is com-
mon to both sexes. These diverse organs might be identified
through the possession of both sets by the female and of a single
set in the male. It is also probable that the male lacks the
glands that secrete the scent whereby the female lays down her
individual path from the nest, and he may also lack the sub-nose
which discerns this path-scent. The male seems to be unable to
lay a path, and, in a change of domicile by the colony, he is car-
ried bodily by the females to the new nest. It is through his
appreciation of the specific odor and his lack of perception of the

THE PROGRESSIVE ODOR OF ANTS. I I

progressive odor that the male is best fitted for his distinctive
office in the ant world.

IV. THE PROGRESSIVE ODOR OF QUEEN ANTS.

The change in the inherent, transmissible, progressive odor in
a line of queens is probably slow and cumulative, but that such
a change occurs is evidenced by the behavior of any segregated
group of Stcnaunna fithnui workers, a species in which the
queens generally remain in the colony in which they are pro-
duced. When workers from such a colony are segregated from
the pupa-stage upward, it becomes difficult to find, in the wild
nest, any queen that these segregated workers will accept as their
own. In this species, I have reared worker-offspring from queens
that were sequestered from all males except those of their own
colony,1 and these workers willingly associated with their worker-
cousins. That the change of odor is but slight in a single gene-
ration is also shown by the fact that the worker-daughters of a
queen, after having been segregated from their pupa-stage upward,
and with no criterion of odor save that of their own bodies, will
affiliate with their queen-mother at a first meeting, though they
always examine her with exceeding care before rendering com-
plete homage.

The gradual change in odor, through the introduction of the
male element, from generation to generation, may be crudely re-
presented by the use of letters as symbols of the odor of queens
of the same species and variety.

The Roman numerals at the left denote successive generations
of mated queens.

The letter a is used as a symbol of the odor characterizing two
sister queens ; the other letters as symbols of the odor inherited
from the paternal grandmother.

I. a, I. a,

II. rt + 3, II. a-\-l,

HI. a + b + c, III. a + l+m,

IV. a+b + c + d, IV. a + l+m + n,

V. a + b -f c -f d-\- e, V. a -+ I -f m + « + o,
etc. etc.

1 " Notes on an Ant," Proceedings of the Academy of Natural Sciences of Phila-
delphia, December, 1902, p. 605.

12 ADELE M. FIELDE.

The female descendants of sister queens would thus become
more unlike in odor with every generation.

An odor providing the means of recognizing a maternal ances-
tor, or another descendant of that ancestor, may be dominant
through more than one generation of females.

The fact that worker ants who have never met any queen will
as joyfully associate with their queen-mother's sister as with their
own mother, indicates that sister-queens have the same odor after
mating that they had before mating, and that the first divergence
in odor becomes apparent to the ants only in the offspring of
sister-queens that mated with males capable of imparting unlike
odors to their respective progeny. The worker ants, having
attained the distinctive progressive odor characterizing their
mothers' worker- offspring for the current year, may produce
males who will each impart to his progeny the distinctive odor
borne by all the female issue of the queen with whom he mates.
Each generation in the line of queens would then depart farther
from the odor of the queen ancestor, and we should find, as we
do, colonies in which all the female inhabitants are inimical to all
the female inhabitants of another colony. There would also be
produced, as in colonies of Stenammafulvum, where queens mate
within the nest and remain to increase its population, the phe-
nomenon of callows that, if segregated from the pupa-stage,
refuse to affiliate with queens from the nest in which were depos-
ited the eggs from which these callows issued.

During several years I have been interested in ascertaining
whether adult, queenless workers would willingly accept a queen
who was indisputably of another colony of their own species, and
among many experiments I have never seen such an acceptance.
If forced into association, escape of either party being made im-
possible, the workers may after a longer or shorter interval live
peaceably with the alien queen, as they also may do with alien
workers. But such forced alliances do not result in normal pros-
perity, even when a whole year is allowed for the cementing of
friendship. So exacting are the ants concerning adherence to
their standard of odor that they prefer a queenless state to the
presence of an unknown ant-oclor. Observations made by me
in the summer of 1905 accord with my earlier ones. Eleven

THE PROGRESSIVE ODOR OF ANTS. 13

workers of Stenamma fulvum piccuin had been inmates of one of
my mixed nests, with Camponotus pennsylvanicus and Formica
subscricca, all hatched between August 14 and September 3,
1903. The Stcnammcts had been removed from the mixed nest,
and kept in segregation since June 23, 1904, and had never met
a queen. On August 13, 1905, I introduced into this nest a
young, winged queen of the same variety as these workers, on the
twenty-fourth day after I had isolated her to ensure her freedom
from incurred odor. The queen fled from the group of workers
and constantly tried to escape. She was attacked whenever I
forced her into the group of workers, and was caught and killed
by them on the ninth day of her sojourn. A dealated queen intro-
duced later, from the wild colony to which these workers origi-
nally belonged, was also killed by them.

Since ants possess so discriminating a sense of smell, and are
so exacting concerning an adherence to the criterion established
for their nest, and since even those ants who have had an ex-
tended experience in ant-odors, and who have been queenless for
two years, refuse to affiliate with a queen of an alien odor though
of their own variety, we may hardly expect that they will volun-
tarily associate with queens of another species. During the sum-
mer of 1905 I introduced queens of other species into segregated,
queenless groups of adult Stenamma fulvum^ of Formica ncoga-
gatcs and of Formica Schaufussi, that had had their sense of smell
highly educated by long association with workers of two or three
different species of ants, but in every case the introduced queen
was killed within a few days, in spite of her constant efforts to
keep aloof from the workers.

In no species of ant have I found workers that would tolerate
the presence of any queen of unfamiliar odor, nor any queen that
would willingly remain among workers of unfamiliar odor.
Although all species of ants have not been thus tested we may
well assume that what is shown to be a fundamental trait in a few
species will manifest itself in all species of the tribe.

14 ADELE M. FIELDE.

V. EFFECTS OF THE PROGRESSIVE ODOR IN THE COMMUNAL

LIFE OF ANTS.

Since the queen is ordinarily the earliest occupant of the ant-
nest, and since her callow young have the same odor as herself,
the odor of her earliest nest must at first be the same as is that
of the queen. Probably this odor is at all times dominant in the
permanent nest ; but as the progressive odor of the workers is
gradually added thereto, the nest-aura would be thereby modi-
fied. The change in the nest-aura, cumulative with the age of
the colony and the increase of the inmates, would be so gradual
that all habitants of the nest would at all times find it familiar
and therefore congenial. The greater dominance of the queen's
odor in the earlier nest may be the cause of the persistence with
which many workers cling or return to the old habitation even
after the majority of the colony has for sound reason removed to
a new abode.

It appears probable also that diffused ant-odor is in direct ratio
to bulk of ant-body, and that a cause of the common activity of
workers in adding the lesser to the larger pile of brood, some-
times even against the inhibitory effects of light, is due to the
more manifest odor of the larger pile.

I have at different times during several years observed in my
artificial nests a most curious phenomenon among ants that had
long lived amicably together. Several or many workers were
seen standing around one ant as if holding a court of inquiry con-
cerning this associate. Sometimes the associate is proscribed,
sometimes rent limb from limb. This extraordinary behavior is
probably due to the victim having attained a progressive odor that
is obnoxious to many other inmates of the nest because unknown
to them. This might happen to an aged ant whose horde of com-
panions were all young. It might also happen that in prowling
for food, or in raids made on the nests of aliens, the worker ants
would bring in alien young for food, and that this much licked
and tended young would incur the progressive odor of the nurses.
At a later period the introduced ant might produce a progres-
sive odor unlike that of the multitude inhabiting the nest, and it
would therefore be doomed to destruction. Ostracisms or violent

THE PROGRESSIVE ODOR OF ANTS. I 5

deaths, such as sometimes occur in nests where amity has long
prevailed, are probably to be explained by the attainment by some
of the inmates of a new and therefore an alarming progressive
odor.

There may be seen among ants of the same variety, and even
in the same individual, all degrees of attraction and repulsion
towards other ants at a first meeting with them. Such manifes-
tations range all the way from cuddling, caressing, cherishing de-
votion through indifference and inattention, distrust, suspicion,
animosity and enduring, ferocious enmity. The inciting cause
is doubtless the progressive odor of the visitor, and the prac-
tical end is the preservation of the chemical standard of the nest.

Whatever the action of the ants, it is always more obvious
when there are numerous young in the nests, and when the nest-
aura is well established.

During five years of fairly constant study of ants I have seen
no evidence that their antennae are the organs of any other sense
than the chemical sense, and I am convinced that any statement
concerning the behavior of ants may well be distrusted if it ignore
the dominance of the olfactory sense over the conduct of the ant,
the ant's almost inconceivable minuteness of discrimination in
odors, or the ant's marvelous memory of odors that have been
encountered. Only when ants are segregated from the pupa-
stage, and full record kept of every experience of theirs in meet-
ing other ants, can the investigator truthfully declare that ants
behave in a certain manner in the presence of other ants. More-
over, as every ant acts on personal experience and individual
memory, the ants should be considered singly as well as in
groups and communities, when a theory of their behavior is to be
enunciated. But when the total history of an ant is known, the
investigator may accurately predict the behavior of that ant
toward other ants. There is ex-ceeding uniformity of behavior
among ants having an identical history.

The progressive odor of the worker ants is manifestly an ad-
vantage in their communal life, since it furnishes the means
whereby every ant can recognize its home and its fellow-citizens,
avoiding nests and communities other than its own. The uses of
this odor within the colony may also be numerous, and it may
determine the distribution of labor in the ant community.

1 6 ADELE M. F1ELDE.

Through the male element, it probably differentiates the odor
of queens of the same species, enabling the workers to find, to
defend, and to cherish their own queen.

It differentiates ants otherwise alike ; determines their distribu-
tion in separate communities ; dictates the behavior of members
of one colony toward those of another colony, and in connection
with an acute sense of smell and a powerful memory, is a domi-
nant factor in the life of the individual ant and in the structure of
the ant-colony.

REACTIONS TO LIGHT AND MECHANICAL STIMULI

IN THE EARTHWORM PERICH/ETA BER-

MUDENSIS (BEDDARD).

E. H. HARPER.

Recent work concerning the behavior of earthworms has related
chiefly to their reactions to light. Since the contributions of
Hofmeister and Darwin, and that of Hesse ('96) there have been
a group of recent papers by Parker and Arkin, Miss Smith,
Adams and Holmes, which have been devoted chiefly to the
directive influence of light. In the present state of the discussion
of this subject the current theory of tropisms has been called in
question, according to which the earthworm is oriented directly
by light. Holmes has shown that light induces a general state
of activity leading to random movements of which those toward
the light are checked and those away from it continued, this
resulting in final orientation.

This paper aims to show that random movements are a feature
of less strong light, tending to disappear with the increase of
intensity, and are replaced by direct orientation in very strong
light. It is also shown experimentally that the earthworm is
more sensitive in the extended than in the contracted state, and
that this has an important bearing upon the production of random
movements. The explanation given of this is that when extended
the sensitive elements of the skin are expanded over a greater
surface. This is shown to have a bearing upon the production of
random movements as follows : Locomotion consists of a succes-
sion of extensions and contractions and as each extension begins
in a state of lower sensibility the anterior end may be projected
toward the light, only to be checked when its increase of sensi-
bility with extension makes the stimulus appreciated. Movements
away from the light are not so checked. In stronger light the
sensibility of the worm when contracted is sufficient to suppress
movements toward the light at the outset. In such light the
worm appears to be orientated without trial movements. It is
important that the worms be kept in the dark before all experi-

17

1 8 E. H. HARPER.

ments, as their discrimination diminishes and random movements
begin again when this is the case.

It is shown that the reactions which are typical of the life in
the burrow are more definite and controlled by weaker stimuli
than reactions in the open, and this may be expressed by saying
that the earthworm's organization is more highly adapted for life
in the burrow. Reactions in the axial direction are definite and
more sensitive to stimuli than lateral movements in response to
light.

The genus Pcrichceta is noted for its agility, and of its special
reactions the leaping movements are the most notable.

DESCRIPTION OF THE SPECIES.

Perichata, the eel- worm, as it is called by gardeners, is an
exotic genus of earthworms which is said to be quite commonly
established in greenhouses in the old world, and also in gardens
in parts of France, where they have been introduced, it is said,
from the east. The only mention of Pcrichceta having been
found in this country, that has come under the writer's notice, is
that of Garman, who reported a species of Pcrichatta as becom-
ing established in greenhouses in Urbana, 111. The writer found
Pcrichceta bermudensis (Beddard) in a greenhouse in Evanston,
111. In suitable conditions of soil these worms flourish in great
abundance.

The genus Pcricliceta is noted for its activity. The squirming
movements which have given it its name of eel-worm are a strik-
ing exhibition of agility. This sort of movement is not confined
to Pericliceta, but is developed in the genus to an extent not
found elsewhere. By alternate contractions of the longitudinal
muscle bands it makes a series of leaps, by which it may waltz
about for quite a distance. It reacts in this way when handled
or disturbed, as when uncovered from its burrow.

The worms are of rather large size. They are found often
measuring nine inches in lengh when killed fully extended.
They are rather pointed at both ends. The continuous circles of
setae on each segment give the name to the family. The clitel-
lum is a complete band or girdle encircling segments 14—16. A
large pair of spermiducal glands shine through the opalescent

REACTIONS TO LIGHT IN THE EARTHWORM. 19

skin behind the clitellum, making a conspicuous mark. The
dorsal pores are very prominent, exuding an abundant yellowish
mucus. The everted buccal cavity is used as a proboscis, and
is thrust out constantly in its feeling movements. The blood
vessels are prominent, shining distinctly through the skin. The
very numerous, minute, diffuse nephridia are a feature which,
along with the continuous circles of setae, have caused consider-
able discussion as to whether these conditions are primitive for
earthworms or secondarily derived.

THE THEORY OF TROPISMS.

The orientations to light and other stimuli, which are among
the most striking phenomena in the behavior of the lower ani-
mals, have received various explanations. After the first anthro-
pomorphic explanations of these movements, based upon likes and
dislikes, there came an apparent revolution of ideas bringing in
explanations of seemingly great simplicity. As the physiology
of plants, particularly of the higher plants, had made consider-
able progress towards a solid physico-chemical basis, there was
a transference of conceptions based upon plant physiology to
the realm of animal behavior and the orientations of the lower
animals were illuminated by analogies drawn from plants. For
example, we find the assertion of identity between heliotropic
phenomena in plants and animals. The mechanism of the
tropism was not a reflex according to this conception, but was
a unique form of movement to be added to the classification of
animal movements into reflex, instinctive and voluntary.

The current theory of phototropic or -tactic phenomena as ap-
plied, for example, to the earthworm, was that when light strikes
one side of the animal so as to cause unequal stimulation of the
two sides, it changes the tone of the muscles on the side affected.
The muscles of one side are thus either relaxed or their tension
is increased according as the animal happens to be positively or
negatively phototropic. It is bent away from or toward the
source of stimulation by the direct action of the environment
upon the protoplasm. The tropism is accordingly regarded as a
peculiar kind of forced movement, dependent upon the chemical
nature of the protoplasm.

2O E. H. HARPER.

Jennings has shown in the case of the Protozoa and also the
Rotifera that the tropism theory gives an untrue explanation of
the mechanism of orientation. These animals are not directly
swerved away from or toward the source of stimulation, but they
have their peculiar methods of reaction and orientation in the
direction of the stimuli is effected by a sort of " trial and error "

method.

REACTIONS OF EARTHWORMS TO LIGHT.

Since Darwin's account of the habits of earthworms there has
been a series of papers devoted chiefly to the directive action of
light upon these forms. Parker and Arkin, Miss Smith, Adams
and Holmes have studied the reactions of earthworms crawling
over surfaces, exposed to light stimulation from one side.

Parker and Arkin observed the head movements of worms
placed at right angles to the direction of the light and determined
that 65 per cent, of the movements were indifferent, /. i\, straight
ahead, 30 per cent, were away from the light and 4 per cent,
toward it. They regarded the various head movements in differ-
ent directions as due to a variety of chiefly undefined causes in
addition to light and since 4 per cent, were toward the light they
assume that as many of the negative responses would be due to
other causes than light. So subtracting 4 per cent, from 30 per
cent, the remaining 26 per cent, they regard as the measure of
the negative phototactic response. Adams showed in addition
that the earthworm is positive to very weak light.

The observers mentioned did not consider the question of the
mechanism of orientation. Holmes takes up the current tropism
theory and questions its explanation of the mechanism of orien-
tation for these animals. He shows that the various extension
movements appear to be of a simply random character, due to a
general stimulation by light. The way in which orientation is
effected he describes as follows. Movements that are toward the
light are checked and the animal draws back and usually moves
in the opposite direction. Movements away from the light do
not lead to further stimulation and so are prolonged farther, and
as a final result of such random movements, the worm gets into
the direction of the rays, in which position the stimulation of the
sensitive anterior end is least, and it then continues to move

REACTIONS TO LIGHT IN THE EARTHWORM. 21

straight ahead. Any swerving from this path leads to an increase
of stimulation and hence is corrected. Holmes regards none of
the movements as forced by light. All are random in direction
but certain favorable ones are followed up and unfavorable ones
checked by the increase of stimulation resulting from them.

Holmes proposes his theory of the " selection of random move-
ments " only as one factor in phototaxis, not wishing to exclude
the possibility of a slight amount of directive influence in the
light. His reason for so doing is based on the observation of
himself and the other experimenters alluded to that there is an
excess of negative turnings over positive ones. Of course if the
movements of the animal are random there should be an equal
number of movements in the positive direction as in the negative,
when one considers only the first movements occurring after
stimulation. Holmes counted a number of first movements and
found them about as equally divided between the positive and
negative side as could perhaps be expected (23 : 27). Parker
and Arkin found an excess of negative movements over positive
of 26 per cent. Miss Smith (on the same basis of reckoning)
found an excess of 39 per cent, and Adams, using different inten-
sities of light, found that the excess was greater with an increase
in the intensity. If the observers did not count only the first
movements after stimulation but also many subsequent move-
ments, the excess of negative movements is not against the sup-
position of their random character. It may be well for clearness
to suppose a case. Of one hundred first movements after stimu-
lation (when the worms are placed at right angles to the light)
there should be an equal number of positive and negative, if they
are purely random. But according to the theory, negative move-
ments tend to be continued while the positive ones are checked
and may be followed by negative movements. This would give
rise to an excess of negative movements in any large number that
were counted. Holmes says that the excess of negative move-
ments may be due to one of three causes — accident, failure to.
count many of the slight positive movements which are easily
overlooked, or to a slight orienting tendency of the light.
Holmes undoubtedly has in mind first movements only, when he
assumes that an excess of negative movements is against the
supposition of their random character.

22 E. H. HARPER.

Holmes's theory of the "selection of random movements as a
factor in phototaxis " is thus based upon observational evidence
which is easy to verify. It is" easy to observe that the move-
ments toward the light are apt to be checked and the movements
away to be more prolonged. It is less easy to note in weak light,
as the final result of orientation takes longer in that case.

ABSENCE OF RANDOM MOVEMENTS IN NEGATIVE PHOTOTAXIS

IN VERY STRONG LIGHT.

All of the experimenters referred to used artificial light except
Miss Smith, who used diffuse daylight. Since all of them but
Holmes took for granted the direct orienting power of light, they
did not care to put the matter to a crucial test. It would seem
that a test of the orienting power of light would require the use
of lights of various strength, and especially of very strong inten-
sity, since the perceptive power for light is so poorly developed
in the earthworm. A test of the orienting power of direct sun-
light is a very easy thing to make. Place the earthworm upon a
sheet of wet paper in a beam of direct sunlight from a window.
The light may be passed through a water chamber. The results
are sufficiently obvious as to leave no doubt of their general
nature. Pericliceta is oriented directly away from the light, when
placed at right angles to the rays. The first effect is a turning
of the anterior end away from the light and by a series of turns
the worm gets into the oriented position and crawls directly
away. Usually the result is produced without a false move-
ment. It is immaterial whether heat effects are excluded by
passing the light through water or not. A species of Lmnbricus
was experimented with and behaved in the same manner.

If the sheet of paper is turned as the worm turns, so as to keep
it at right angles to the rays the worm will travel in a circle con-
tinuously. To show the difference between the orienting effect
of sunlight and that of an ordinary artificial light the following
experiment was tried. By using a sort of searchlight consisting
of a tube of asbestos paper surrounding a 32 c.p. incandescent
light and narrowed to a small aperture, the light was so manipu-
lated by the hands as to keep it constantly directed upon the
anterior end of the worm, with the worm at right angles to the

REACTIONS TO LIGHT IN THE EARTHWORM. 23

rays. In this way the worm was kept under constant stimulation
and caused to turn through one complete revolution and the
time required was noted. The process of turning was slow and
was effected by a series of readjustments involving many trial
movements in the opposite direction. Most commonly about
two minutes was necessary. In twenty such cases the average
time required was five minutes, the greatest time, twenty
minutes.

In the beam of sunlight as before stated the worm turns con-
tinuously without trial movements. The difference in behavior
in the two cases is so striking that the occurrence of an occa-
sional positive random movement in the sunlight is plainly seen
not to affect the general result. When the worm is exposed to
the sunlight, if a passing cloud obscures the sun, random move-
ments begin to appear. Miss Smith, who used diffuse daylight
from a north window, observed that the worm moves in a general
direction away from the light, but in an uncertain manner.
Adams, using a graded series of artificial lights, showed that the
per cent, of negative movements increased with the intensity.
Adams did not observe the whole process of orientation since he
placed the worms in an illuminated box and observed the direc-
tion of their movement after an interval of stimulation. Holmes
used artificial light of only one strength. A Welsbach burner
was also used to give an intermediate intensity between those
before mentioned. Worms were used that had been kept in the
dark and they were brought suddenly into this powerful light.
They all moved away from the light with very little appearance
of random movements. At each forward extension they would
turn a little away from the light so that their path appeared like
a curve. It is not meant to be stated that there were no random
movements. But there could be no hesitation in saying that
there was a decided difference in their reaction under the stronger
light. Fresh worms would by a series of turns get into the ori-
ented position frequently without a noticeable random movement.
If the worms were kept in the light for some minutes, they lost
sensitiveness and their random movements began to be evident.

24 E. H. HARPER.

OCCURRENCE OF POSITIVE PHOTOTROPISM.

When using a 3 2-candle-power incandescent light it was noticed
that some individuals behaved positively. About 6 per cent, of
200 worms tested showed the positive reaction. But at a few
inches distance from the light these worms would apparently
become negative. Heat effects were not excluded however.
The following is a typical instance. An earthworm crawling on
a table moved straight toward a 3 2-candle-power incandescent
light until within a few inches, when it began to swerve and
without pausing moved in a continuous curve away from the
light until it was in the line of the rays, when it continued to
move in a straight line away from the light.

DIFFERENCE IN THE SENSIBILITY OF EARTHWORMS TO LIGHT IN
THE CONTRACTED AND EXPANDED STATE AND THE
BEARING OF THIS FACT UPON THE PRODUC-
TION OF RANDOM MOVEMENTS.

The conclusion reached is that earthworms are oriented
directly by light, but owing to their low degree of sensitiveness
their movements are uncertain except in very strong light. The
influence of light produces a number of noticeable effects upon
the behavior. First, there is a state of general stimulation or
restlessness inducing locomotion. Second, in light not strong
enough to produce direct orientation the worm projects its
anterior end in any direction. If toward the light, the worm after
stretching out its anterior end will again retract it as if stimulated.
If the worm is checked only after making an extension move-
ment toward the light, the conclusion would seem to be that the
anterior end is more sensitive when extended than when in the
contracted condition. One may test this conclusion by further
experiment. If a light is flashed suddenly upon a contracted
worm the influence of the stimulus seems to affect it gradually,
leading after an interval to movements. The extended anterior
end responds far more quickly to sudden changes of stimulation.
The basis for this difference in reaction must be in the fact that
when the head is extended the sensitive elements in the skin are
spread out over more surface than in the contracted state. A
simple experiment will illustrate this fact. If an earthworm is

REACTIONS TO LIGHT IN THE EARTHWORM. 25

crawling on a moist paper it may be shaded by the hand or
otherwise. When the worm crawls to the edge of the shadow
and thrusts out its anterior end into the light it is jerked back
suddenly. But if the light be thrown upon the worm when con-
tracted, there is no sudden response, but only a gradual awaken-
ing to stimulation, as evidenced by subsequent movements. The
bearing of this observation upon the movements of the worm would
seem to be as follows : The worm contracted is like an animal with
its eyes partly closed. It extends its head at random, thus grad-
ually receiving the full stimulation upon its surface. If the
movement is toward the light, this causes it to contract more or
less and so check stimulation. If the movement is away from the
light, the oblique illumination produces less stimulation and the
movement is more prolonged. An animal with eyes, as a crusta-
cean, or an insect, is of course so organized that movements
toward the light may be checked, as it were, at the outset, in the
case of negatively phototactic animals.

It is to be observed that the earthworm begins these random
movements while in the contracted state. After extension it
draws up its body by means of the longitudinal muscles and is
therefore in the contracted state. It then advances again, and at
each advance there may be a random change in direction. Thus
the worm begins these random movements when in the con-
tracted state and under minimum stimulation. The nature of
its locomotion and of the sensitive elements in its skin necessitates
the alternation of states of low and high sensitiveness. The
random movements of an earthworm under light stimulation are
consequently of an entirely special character, due to causes in-
herent in its structure.

To recapitulate, three situations in regard to light have been
described, with their characteristic reactions. First, in weak
light, second in strong light and third in a situation involving
change of light intensity.

The stimulus of a change in intensity causes the animal to
draw back its anterior end slightly and it then usually alters its
course. When crawling under the influence of sufficiently strong
light, it bends its head away from the light at each successive
advance, until it gets into the oriented position. In light not

26 E. H. HARPER.

strong enough to have the directive effect its extension movements
are random, an advance toward the light being checked and
orientation being brought about by following up of favorable
random movements. There are only two responses in reality, the
checking or drawing back of the head involving the symmetrical
use of the longitudinal muscles of both sides, and the turning
response, involving the longitudinal muscles of only one side,
that opposite to the source of stimulation. The two responses
may also be combined.

THE ANATOMICAL BASIS FOR THE DIFFERENCE IN SENSITIVE-
NESS TO LIGHT IN THE EXTENDED AND IN THE
CONTRACTED STATE.

The text figures introduced are intended to make clear the
reason for the difference in the sensibility to light of the anterior
end in the contracted and extended state. Hesse, who has
worked on the organs of light perception of the lower animals,
has shown the structure of the light cells in many species of
earthworms and has worked out their distribution segmentally.
He shows that these cells are most numerous on the first seg-
ment, and especially on the prostomium (which is fused with the
first segment in Perich&td) and that their number diminishes
rapidly on each segment as we go farther back. It is conse-
quently the very tip of the animal ( the posterior tip as well )
which is most important for the perception of light, although
light cells are found in small numbers over the whole length of
the body.

The sections of PcricJiceta (Figs. I and 2) show that the first
segments are subject to great extension and contraction. It was
not possible to get the worm fixed in the fullest state of either
extension or contraction. In Fig. 2 it is seen that the first seg-
ment is partly inrolled into the buccal cavity in the state of con-
traction. For further demonstration of this point the epithelial
layer alone, of the first segment, is represented in the extended
and the contracted state in Figs. 3 and 4. It is seen to be greatly
thickened as well as inrolled when contracted. The effect of
this on the light cells is seen by comparison of Figs. 5 and 6.
The light cells are on the basement membrane. The thickening

REACTIONS TO LIGHT IN THE EARTHWORM.

The figures illustrate certain features of the worm in the extended and in the con-
tracted state. The states of extension and contraction represented are not the most
complete possible.

FIGS. I and 2 are sagittal sections of the first five segments of an extended and a
contracted worm respectively. (<?) body wall, (£) everted buccal cavity slightly
protruding, which is used as a proboscis. It is to be noted that the first segment,
which is the most sensitive to light, is partly inrolled in Fig. 2

FIGS. 3 and 4 are sections of the dorsal epithelium of the first segment of an ex-
tended and contracted worm respectively in the same plane as the last.

FIGS. 5 and 6 give sections of epithelium in the extended and contracted condi-
tion. (/. f.) light cell.

28 . E. H. HARPER.

of the epithelial layer must of course tend to cut off light from
these cells. The inrolling of the most sensitive region is another
important factor.

THE BEARING OF EXPERIMENTAL RESULTS UPON THE HABITS

OF THE EARTHWORM.

It is a truism that in all experiments upon animals the relation
of the experimental results to the normal life of the animal should
be kept in mind. The behavior of the earthworm has not been
systematically studied as a whole except by Darwin. It is
obvious that all the experimenters mentioned have studied the
reactions of the earthworm in only one phase of its activity, and
that phase is not what we should call the normal life of the worm.
It is as if the experimenters had chosen the situation of the earth-
worm as we find it crawling on the sidewalks after a heavy rain
as being its typical mode of life. None would probably admit
this sooner than themselves, and doubtless they have regarded
certain facts as too obvious to require mention. Does the fact
that the normal life of the earthworm is carried on in a burrow
affect our view of the experimental results obtained ? Now the
earthworm does spend a portion of its life, during the night time,
crawling on the surface of the ground in search of leaves, and also
during sexual activity it is less mindful of the light, as is stated.
The earthworm leaves its burrow rather reluctantly. Darwin
describes the earthworm as retaining the posterior end in the
burrow while making searching movements in all directions in
search of leaves. In drier weather we know that worms burrow
deeply and seldom are found near the surface, depositing their
castings in old burrows instead of on the surface.

If the worm is at home almost exclusively in the burrow we
should expect those responses which are typical of the burrow
life to be better organized and more definite than its activities
when crawling in the open. The movements which are typical
of the life in a burrow are mainly in the line of the axis of its
body. Are these movements and the responses which control
them of a more definite nature than its lateral movements ? We
may first consider the typical burrow movements in response to
light. These may be imitated easily by using a screen to shade

REACTIONS TO LIGHT IN THE EARTHWORM. 29

the worm or portions of it while crawling on a moist surface,
preferably covered with a thin layer of dirt.

If a worm which has been kept in the dark is placed on the
moist surface, and the screen is suddenly moved so as to expose
the anterior end to light, it contracts the anterior segments slightly,
sometimes so slightly as to be barely noticeable, and crawls back-
ward into the shadow. If the posterior end be illuminated in the
same way the worm crawls forward into the shade, but after a
noticeably longer interval. A slight twitching of the posterior
end may be noticed at first, if the light is suddenly turned on.
The worm always crawls forward when stimulated posteriorly,
If a worm is crawling backward it can always be reversed by
stimulating the posterior end. Crawling backward is of course
the method by which the worm comes to the surface to eject
castings. The two sorts of responses described are of the kind
called " photopathic " as distinct from phototactic, and they serve
of course to inhibit the worm from leaving its burrow in the light.
These " photopathic " responses are very definite and the stimulus
calling them forth may be quite weak light. Adams has shown
that in very weak light AllolobopJwra fcetida is positive and he
suggests that the worm leaves its burrow in response to the
stimulus of very weak light upon its anterior end. These ob-
servations show that the movements of the worm in its burrow
are very definitely controlled by the light, so far as they may
come in contact with it by their more"sensitive anterior and pos-
terior ends. The middle of the worm is less sensitive to light
but its sensitiveness may be shown in the following way.

If the worm is placed on the moist surface exposed to full
light from overhead (a 32-candle-power incandescent was used,
at a distance of 15 inches) and a screen is then brought over the
posterior part of the body leaving the anterior end exposed, the
worm does not draw back as when the anterior end was suddenly
illuminated. Instead it begins to make random movements in
various directions. It may crawl farther out into the light, thus
bringing the middle portion into stimulation. This movement is
however checked before the more sensitive posterior end is ex-
posed. After a noticeable latent period showing the lower sensi-
tiveness of the middle portion, the worm crawls back under the

3O E. H. HARPER.

shade rather quickly, but usually not completely. The com-
monest way in which the worm gets under the shade is as fol-
lows : It makes all sorts of random movements in every direction,
and tries to burrow into the thin layer of dirt, until it acci-
dentally gets the tip of the anterior end under the shadow of the
screen. It then at once is oriented, so to speak, and crawls com-
pletely under the glass. It may crawl under as if circling around
a post. The imaginary post may be exposed to the light so that
the posterior part has to crawl forward into the light to get
around the post. Usually, however, the anterior end travels
faster so as to jerk the middle part under the screen at once.

These so-called " photopathic >: reactions are consequently
very definite and predictable because they are adaptations im-
portant in the normal life of the worm. As compared with the
random lateral movements we see that they are controlled by
weaker stimuli and are more definite. The anterior and posterior
ends are more sensitive than the middle for the obvious reason
that the ends alone come into contact to any great extent with
light stimulation.

The lateral movements, which are typical of life outside the
burrow, are as we have seen of a random nature and less defi-
nitely controlled. The worm " dashes back like a rabbit into its
burrow," to use Darwin's expression, under a weak stimulus.
But when crawling on the surface the same strength of stimulus
produces only a general irritation and swaying random move-
ments occur which lead to orientation away from the light only
after many trials. With a higher intensity of light the worm is
oriented more quickly. Thus we see that a very high stimulus
is required to produce a direct sidewise movement away from the
light while a very weak stimulus will cause it to move back into
its burrow away from the light. The random lateral movements
are aptly described by Holmes as " inconsequential vermicula-
tions." But this description does not apply to the movements
which are typical of its burrow life. The worm is as definitely
adapted to the burrow and as little adapted for life in the open
as some other burrowing animals of higher rank that could be
mentioned. However this statement must be modified when we
consider that a worm exposed to the light on the ground does not

REACTIONS TO LIGHT IN THE EARTHWORM. 3!

trouble to make random movements but begins to burrow into
the soil immediately. After heavy rains we see them washed out
of their burrows, and crawling in unwonted places when they are
unable to burrow.

REACTIONS TO MECHANICAL STIMULI.

Pcric/ueta goes through its peculiar jumping movements only
under mechanical or similar stimulation, never under the influence
of light, so far as we have observed. When touched with a
needle on the anterior end it contracts the anterior segments
slightly and may begin to crawl backward or it may go forward,
lifting its head and making various random movements before
settling on any direction. With a slightly stronger stimulus the
anterior end turns away slightly from the stimulus. Increase the
stimulus and the worm may contract the longitudinal muscles of
the opposite side so as to jerk the body around 90 or even 180
degrees, and so give it a new direction. Or the worm may go
off into a whole series of jerks, so that there is a complete grada-
tion between the extent of the responses, depending upon the
stimulus. More important as determining the extent of the
reaction is the condition of the worm. Well-fed worms in fresh
condition, when just dug out of their burrow, spring around in
the liveliest fashion. If handled they give a series of movements
which must make it difficult for an enemy, a bird, for instance, to
pick them up before they get a chance to crawl under cover.
When stimulated they exude an abundant yellowish mucus.
Whether this is an offensive secretion to its enemies is not known
to the writer. When a point in the middle of the worm is stimu-
lated the body recoils away from the stimulus at that point and
there is a slight swelling due to contraction of the longitudinal
muscles, like the contraction and shortening of the anterior end
under stimulation. Occasionally the worm may move violently
toward the stimulus, but this seemed to be due to an overstimu-
lation producing a complex of effects rather than a simple reflex.

The leaping movements of Pericli<zta are certainly the best
examples of random movements that are afforded. They are
exclusively adapted to those chance circumstances when the
worm gets out of its burrow. They lead in no definite direction,

32 E. H. HARPER.

though they may carry the worm to a considerable distance and
enable it to distract the enemy. They are a conspicuous example
of the character of most movements of the earthworm, which
belong to its limited life outside the burrow.

CONCLUSIONS.

The method of reaction to light of the earthworm is far
removed from the sort of "trial and error method" of the Infu-
soria, as analyzed by Jennings. Its avoiding reaction in strong
light is of the nature of a definite reflex which causes it to turn
directly away from the stimulus, if the whole body is in the light,
or to retreat into its burrow, if only the anterior end is stimu-
lated, or go forward if the posterior end alone is stimulated.
Methods of trial and error in reaction to light and other ordinary
stimuli have clearly been supplanted by more definite responses
in all but the Protozoa and certain other low types of animal
life. The earthworm's reactions to stimuli, mechanical, thermal,
chemical, are in general such as its nervous system and muscu-
lature would lead us to expect. The occurrence of random
movements in response to all but very strong light is the out-
come of the undeveloped condition of its organs of light per-
ception, not to the want of a nervous system and musculature
adapted for such simple reflexes. Diffuse organs of light per-
ception may not respond definitely to a localized stimulus unless it
is a very strong one. The trial and error method of its responses
to relatively weak light are exceptional in character in com-
parison with its reactions to other ordinary stimuli. Its archaic
type of end organs for light gives rise to a type of behavior which
is to be regarded as primitive. For the trial and error method
is clearly supplanted in the ascending scale of animal life, by
reactions of a definite nature, in the case of the simple responses
to the ordinary stimuli.

SUMMARY OF RESULTS.

1. PcricliiCta bcrmudensis (Beddard) is an exotic earthworm
found sometimes in greenhouses. Its active habits are one of its
chief characteristics.

2. The body is less sensitive to light when contracted than

REACTIONS TO LIGHT IN THE EARTHWORM. 33

when extended, owing to the fact that when extended the sensi-
tive elements are spread out over a greater surface and become
more susceptible.

3. In locomotion, as there are alternate extensions and con-
tractions, there is an alternation of the condition of lower and of
higher sensibility. This is important particularly in the sensitive
anterior end.

4. As the worm begins each extension in a condition of
lower sensibility, it may project its anterior end toward the source
of light. This movement is checked as soon as the increased
sensibility of the extended anterior end appreciates the stimulus.
Movements away from the light do not meet such a check and
so are prolonged farther. Orientation is the result of a trial and
error method.

5. In strong enough light, random movements toward the
light are suppressed altogether, and the worm appears to move
djrectly away from the light without noticeable trial movements.
This applies to worms which have been kept in the dark and are
in a perfectly fresh condition, as after a time they lose their dis-
crimination and begin to make random movements.

6. Movements in the longitudinal direction are typical of the
normal burrow life of the animal, and the axial movements
initiated by the anterior and posterior ends are more definitely
controlled by the stimulation of light and by a weaker stimulus
than are the lateral movements. Lateral movements tend more
to be random and are directed only by stronger stimuli because
the organization of the worm is chiefly in adaptation to a burrow-
ing life and not to an open air life.

7. The characteristic leaping motion of Perichcefa is a con-
spicuous example of random lateral movements, adapted to life
outside of the burrow. All gradations may be observed between
the ordinary reaction to a slight local stimulus by jerking back,
and also bending the body away, if the stimulus be stronger, up
to a complete series of leaping movements.

8. Reactions to mechanical stimuli, as well as to other stimuli,
chemical, thermal and electric, show that the worm is like other
animals as highly organized as itself in responding to a local
stimulus of an injurious nature by contracting and bending away

34 E. H. HARPER.

in a definite " avoiding reaction." In this respect the effect is
like the response to very strong light. Consequently we see that
the random reactions to weaker light have a special explanation
and are only an apparent exception to the general form of nega-
tive response.

BIBLIOGRAPHY.
Adams, George P.

'03 On the Negative and Positive Phototropism of the Earthworm Allolobophora
foetida (Sav. ) as Determined by Light of Different Intensities. Am. jour.
Physiol., Vol. IX., No I.

Gar man, H.

Bulletin of Illinois State Lab. of Nat. Hist., Vol. III., Art. IV. Note
p. 74.

Hesse, R.

'96 Untersuchungen iiber die Organe der Licht-Empfindung bei niederen
Thieren. Zeit. wiss. Zool., 6l, p. 393.

Holmes, S. J.

'05 The Selection of Random Movements as a Factor in Phototaxis. Journ.
Compar. Neurology and Psychology, Vol. XV., No. 2, pp. 98-112.

Jennings, Herbert S.

'04 Contributions to the Study of the Behavior of Lower Organisms. Published
by the Carnegie Institution of Washington.

Parker, G. H. and Arkin, L.

'01 The Directive Influence of Light on the Earthworm Allolobophora foelida
(Sav.). Amer. Journ. Physiol., IV., pp. 151-157.

Smith, Amelia C.

'02. The Influence of Temperature, Odors, Light, and Contact on the Move-
ments of the Earthworm. Amer. Journ. Physiol., VI., pp. 459-486.
ZOOLOGICAL LABORATORY,

NORTHWESTERN UNIVERSITY,

EVANSTON, ILLINOIS, August 14, 1905.

SOME POINTS REGARDING THE BEHAVIOR
OF METRIDIUM.

LULU F. ALLABACH.

It has recently been shown that the reactions of many sea
anemones are modifiable, in dependence on a variety of internal
conditions (Jennings, 1905). The purpose of the work here pre-
sented was to determine how far similar relations hold for Mctri-
dinni niarginatinn. Since jMctridimn is the commonest of our
sea anemones, and the one most used in investigation and in-
struction, it is important that its behavior should be well known.
The work was suggested by Dr. H. S. Jennings, and carried out
under his direction at the Marine Biological Laboratory at
Woods Hole.

I. CHANGES IN THE REACTIONS TO CERTAIN SORTS OF

FOOD BODIES.

The point to which experimentation was first directed was the
interpretation of the results of certain experiments of Nagel
(1892) and Parker (1896). In Parker's experiments alternate
pieces of meat and filter paper (soaked in meat juice) were given
to the tentacles of one side of the disk of Mctridiuin. It was
found that while the meat was swallowed each time with equal
readiness, the time taken in swallowing the paper increased, and
after three or four trials the animal no longer ingested the paper,
though the latter contained each time the same amount of meat

o

juice as at first. After reaching this result with the right side
of the disk, the same series of experiments was performed on the
opposite side of the disk of the same specimen. It was found
that the left side* had not become modified by the experience of
the right side. It at first took the paper, then by the same
gradual change seen previously on the right side, it came to
refuse the paper. A series of records of the times required for
swallowing the meat and paper in such an experiment by Parker
are given in the following table :

35

LULU F. ALLABACH.

Right Side.

Left Side.

Right Side.

Left Side.

'

I. Meat

85 sec.

45 sec.

9. Meat

70 "

35 sec.

2. Paper

80

90

10. Paper

—

85 «

3. Meat

5°

45

II. Meat

40 "

30 "

4. Paper

90

—

1 2. Paper

95 "

5. Meat

40

45

13. Meat

35 "

6. Paper

I°5

55

14. Paper

—

7. Meat

5°

35

15. Meat

35 "

8. Paper

105

16. Paper

—

What is the cause of the change of behavior in these experi-
ments ? Several possibilities suggest themselves :

1. Jennings (1905) found that in Aiptasia similar effects were
produced, and in this case the result was evidently due mainly to
changes in the state of hunger. The very hungry Aiptasia took
meat and paper readily, but after feeding a -short time it refused
paper, and later it came to refuse meat also. The animal could
be caused to refuse paper more readily by feeding it meat alone
than by feeding paper alone or by feeding the two in alternation.
It was evident that the changed reaction toward paper was due
to loss of hunger. We must inquire whether this factor plays a
part in Metridium.

2. Nagel (1892) referred the changes in reaction toward paper
to a process corresponding to what we call judgment in higher
animals ; he held that the animal discovers by experience that
the paper is unfit for food, and thereafter refuses to take it. Such
a process in so low an animal would of course be of great
interest, and the evidence for it needs to be examined carefully.
Nagel held that it was owing to the lack of close nervous inter-
relation of parts in these animals that the experience of one side
is not transmitted to the opposite side.

3. Parker sums up the phenomena shown in these experiments
as follows : " The successive application of a very weak stimulus
is accompanied, not by the summation of the effects of stimula-
tion, but by a gradual decline in these effects, till finally the
response fails entirely" (Parker, 1896, p. 116).

In my experiments I attempted to test these different possi-
bilities, and to work out in a systematic way the various factors
which modify reactions in Metridimn.

The taking of food has been well described by Parker (1896).
It is important to note that ciliary action plays a large part in the

BEHAVIOR OF METRIDIUM. 37

taking of food by Mctriduim. Muscular movements of the ten-
tacles, disk, and oesophagus also plays a part, but a less impor-
tant one than in the anemones studied by Jennings (1905). The
tentacles bearing bits of food are bent toward the mouth, and
the cilia of the tentacular surface carry the food toward the mouth
opening. The cilia of the oesophagus are usually beating out-
ward, but when food enters the mouth the stroke of the cilia of
the part in contact with the food becomes reversed, so that the
food body is conveyed inward (see Parker, 1896, 1905). I have
found that the reversal of the cesophageal cilia is frequently
caused in JMctridiuui by indifferent solids, such as filter paper, so
that such bodies are ingested.1 There is much variation in regard
to this matter, some individuals take filter paper readily, others
slowly and only at times, others not at all. As we shall see
later, this depends largely on the degree of hunger.

We will now take up experimentally the various possibilities
of modification above distinguished. In studying these matters,
it is most important that specimens which are fresh and in good
condition should be used, otherwise clear results will not be

7

obtained.

I. Hunger. — Conditions of hunger and satiety affect the food
reactions of Metridinin in a most decided way. When the animal
is very hungry (and in good condition otherwise), the column is
extended, becoming long and slender, while the disk is widely
spread, and the tentacles extend a considerable distance beyond
its edge. If the animal remains contracted, it can often be in-
duced to extend by placing a piece of clam meat or some meat
juice on the infolded disk. Two or three applications of food
will often cause the most obstinately contracted specimens to ex-
pand beautifully. If now a piece of mussel of considerable size
(having an area equal to the cross section of the column, with a
thickness of three or four millimeters) is brought near the edge
of the disk, so as to come in contact with the tips of two or three
tentacles, a decided reaction is produced. The tentacle tips ad-
here to the meat, and the tentacles and adjacent parts of the disk
contract quickly, so that the piece of meat is drawn inward. It

1 This was possibly due, as Parker (1905, 1905*7 } has set forth, to the paper's
having been touched by the fingers ; this matter was not tested.

38 LULU F. ALLABACH.

thus comes in contact with many other tentacles, these bend
down upon it. Then all bend over toward the mouth, while at
the same time this portion of the disk contracts. Thus the food
is brought nearer the center of the disk. The mouth meanwhile
opens, and the food is passed into it, partly by the ciliary action
of the tentacles, partly by muscular contractions of tentacles,
disk and mouth. The latter factors play a more important part
in the reaction of a hungry specimen to a large piece of food than
does ciliary action.

After ten or a dozen good-sized pieces of meat have been
swallowed, the reaction becomes much slower. If the meat is
brought in contact only with the outer tentacles, these no longer
react, and such food is not taken. If placed on the inner ten-
tacles, the meat is slowly transferred to the mouth, where it is
swallowed. The animal is frequently in this condition when
brought into the laboratory ; food placed on the outer tentacles is
refused, while that on the inner tentacles is taken.

The reaction of the tentacles becomes slower as more food is
taken, so that the process of ingestion takes a much longer time
than at first. Finally, all the tentacles cease reacting to food,
and it is not carried to the mouth. But if the meat is placed by
the experimenter directly on the mouth, it is ingested. This ap-
pears to take place almost alone through the action of the cilia ;
the reversal of the cilia seems to be more nearly independent of
the physiological states of the animal than are the contractions
of the muscles. The mouth never reaches a condition where it
rejects pieces of mussel placed directly upon it. So much food
may be taken that the body becomes puffed out to form a swol-
len sack, yet new pieces are forced inward. Large pieces of meat
may however be refused even when placed directly on the mouth,
the lack of assistance from muscular contractions appearing to
make it impossible for the cilia to draw them inward. The feed-
ing may be carried so far as to cause internal disturbance, result-
ing in the disgorgement of the food. But immediately after such
disgorgement the mouth will take new food. Sometimes when
the animal is nearly filled, a large piece of meat placed on the
mouth is partly swallowed, then partly disgorged by a convulsive
movement, then the swallowing is resumed. This may happen
repeatedly.

BEHAVIOR OF METRIDIUM. 39

The loss of reaction on the part of the tentacles after much
food has been taken is not due to fatigue resulting from their
activity in taking food. This is demonstrated by the following
facts : (i) The animal may be fed from one side of the disk till
it is satiated. Now meat given to the opposite side of the disk
is not taken, though the tentacles of this side have not been
active, and so cannot have become fatigued. (2) Seven hours
after the animal has been fed all it will take, the tentacles still
refuse to take food, though they have had this period for recuper-
ation. Actual fatigue, as we shall see later, lasts but a few
minutes.

The effects of hunger and satiety are further seen in the reac-
tions of Mctridinin to indifferent bodies, such as bits of filter
paper. These are commonly taken readily by hungry specimens
of Mctridinui. The tentacles react to them, just as to meat, so
that they are carried to the mouth. Here they cause the
reversal of the ciliary movement in the same way as does meat,
so that they are carried inward. But after the animal has been
fed a considerable quantity of meat, it will no longer take filter
paper. First the outer tentacles refuse it, later the inner ten-
tacles, and finally the mouth. A piece of filter paper placed
squarely on the mouth no longer causes the reversal of the
stroke of the cilia of the oesophagus, so that it is not carried
inward.

The fact that the reversal of the cilia under such stimuli
depends on the physiological state of the animal is one of much
interest. It shows that the cilia are not entirely independent of
such states, as some other facts would seem to indicate.

Seven hours after the animals had been fed an abundant meal
of mussel meat, they still refused to take filter paper, though
before the meal paper was taken readily.

Thus it is clear that the state of the processes of metabolism is
in j\lctridinin, as in other sea anemones, a most important factor
in determining behavior under mechanical and chemical stimuli.
But it is equally clear that this will not explain the results of
Parker's experiments, described in the first paragraphs of this
paper. Parker found that after one side of the disk has refused
to take paper, the other side still accepts it. In repeating

4O LULU F. ALLABACH.

Parker's experiments, I fed in one case six successive regions of
the disk, each till it had rejected food ; the next would then take
it as readily as at first. This refusal then cannot be due to a
general lack of hunger, and we must examine the other explana-
tions that have been given.

2. "Judgment." — If the animal comes to reject paper through
experience of the fact that the paper is not good for food, there
must be some way in which this experience is obtained. It
might be supposed that this comes through swallowing the food ;
being indigestible, its effect after swallowing might cause the
animal thereafter to reject it. This was tested by preventing the
swallowing of the paper. The animal was fed meat and paper
in alternation, as in Parker's experiments. But after the paper
had been carried to the mouth and was passing down in the
oesophagus, it was removed with a fine pair of tweezers. This is
easily done without disturbing the animal. Thus the bits of
paper never reach the digestive cavity. Yet the animal comes
to reject them as quickly as before. After a few alternations of
meat and paper, only the meat being completely swallowed, the
animal ceased to take the paper, while it still accepted the meat.
Hence the effect of the paper after it reaches the digestive cavity
is not the cause of its rejection.

Furthermore, nothing like a contrast or comparison between
the meat and the filter paper is necessary in order to induce the
rejection of the paper. If successive pieces of paper were fed
alone to the anemones, they soon came to reject these as before.
In such cases it is noticeable that the animal takes a larger num-
ber of pieces of paper than when the paper is fed in alternation
with meat. The number of pieces required is about the same
as the number of pieces of meat and paper together that result in
rejection of the paper when the two are given in alternation.
This fact throws some light on the cause of the rejection, as we
shall see later.

3. Repetition of Weak Stimuli, till Effect Fails. — Is the loss
of the positive reaction to the paper due to the general fact that
weak stimuli, when repeated, gradually lose their effect ? This
was tested by excluding this factor from the experiments, in the
following way. A given specimen was first tested and found to

BEHAVIOR OF METR1DIUM. 4!

accept filter paper (in some cases plain, in others soaked in meat
juice). After this first test, the same region of the disk was fed
successive pieces of meat, which were all readily taken. After
eight to twelve pieces of meat had been accepted, a piece of filter
paper like that originally accepted was given to the same region
of the disk. // tvas not accepted. This experiment was repeated
with many specimens, always with the same result. This result
was likewise reached if the animal was not allowed to complete
the swallowing of the meat, the latter being removed after it had
passed into the oesophagus. This, of course, shows conclusively
that loss of hunger is not the cause of the change of reaction
toward the paper.

The placing of meat juice on a certain region of the disk
causes the food reaction, as Parker has shown. If this experi-
ment is tried successively a dozen times in the same region of the
disk, the animal comes to reject filter paper in this region, as in
the experiments described in the foregoing paragraph.

Thus it is not necessary that weak stimuli should be repeated
in order that the animal shall reach a state in which it fails to
react to them. Repetition of strong stimuli (meat) causes failure
to react to weak stimuli just as readily as does repetition of the
latter. Repetition of strong stimuli alone, of weak stimuli alone,
and of the two in alternation, all have the same effect ; the animal
ceases to react to weak stimuli.

In all cases in which meat is fed to a given region, moreover,
the reaction to strong stimuli ceases some time later than that to
weak stimuli. After giving a certain region sixteen to twenty
pieces of meat, meat is no longer accepted here, though other
regions of the disk take it readily.

4. Fatigue. — The facts brought out in the foregoing para-
graphs seem to make possible a clear interpretation of the rejec-
tion of the paper. It is evidently a case of plain fatigue. After
stimulating a certain region of the disk a number of times, it
ceases to react — first to weak stimuli, then to strong stimuli -
though other parts react as before. The same results are pro-
duced whether the successive stimuli are all strong or all weak,
or partly strong and partly weak. It appears evident therefore
that it is the reaction of the animal, not the precise character of

42 LULU F. ALLABACH.

the stimulus, that causes the fatigue. This is perhaps what
should be expected when the nature of the food reactions is
taken into consideration. In taking food the region in contact
with the food produces a very large quantity of mucus, envelop-
ing the food body. It is not surprising that successive imme-
diate repetitions of this excessive production of mucus gradually
exhausts the region. As is usual in fatigue, strong stimuli may
produce reaction for some time after weak ones have failed.

The fatigue thus caused usually lasts only two to five minutes.
After this period has elapsed the fatigued region is frequently as
ready to take food as before - - provided the animal is still hungry.

Nagel and Parker have held that the result of their experi-
ments " illustrates the extreme looseness, or even independence,
of the nervous activities of the two sides of the animal " (Par-
ker, 1896, p. 1 1 6) --since the effects of the experience of one
side are not transmitted to the other side. With the recognition
that these results are a simple matter of fatigue, they perhaps
cease to have any bearing on the question of the closeness or
looseness of nervous interconnection. In the highest organisms,
as man, fatigue induced by repeated contractions of a finger of
the left hand, in erogographic experiments, is not transmitted ap-
preciably to the right hand. But the experience gained by touch-
ing a hot iron with the left hand would nevertheless later prevent
the right hand from touching it.

II. OTHER MODIFICATIONS IN BEHAVIOR.

A very peculiar modification of behavior is seen in the follow-
ing : A specimen refuses to take filter paper, though it still takes
meat. After it has thus refused paper, two or three pieces of
meat are given in succession, and taken readily. Now the bit of
paper is again placed on the disk, and it too is swallowed.
Clearly, the uninterrupted taking of a number of pieces of meat
changes the physiological condition of the animal in some way,
preparing it for the taking of any object with which it comes in
contact. (After a larger number of pieces of meat, the paper is
refused, as we have before seen.)

Acclimatization to weak stimuli is readily demonstrated in
fresh, active specimens of Metridium. If a light stream of water

BEHAVIOR OF METRIDIUM. 43

is directed with a pipette against the expanded disk, the animal
contracts strongly. Waiting till it has again expanded, the
stream of water is directed upon it as before. This time it does
not react. In specimens that are not in good condition, this
change of behavior cannot be seen. The animal does not con-
tract at all save under strong stimuli, and if such are repeated, it
contracts as at first.

PAPERS CITED.
Jennings, H. S.

'05 Modiriability in Behavior. I. Behavior of Sea Anemones. Journ. Exper.
Zool., II., No. 4.

Nagel, W. A.

'92 Das Geschmacksinn der Actinien. Zool. Anz., XV., 334-338.
Parker, G. H.

'96 The Reactions of Metridium to Food and Other Substances. Bui. Mus.
Comp. Zool. Harvard College, XXIX., 102-119.

'05 The Reversal of Ciliary Movements in Metazoans. Amer. Journ. Physiol.,
XIII., 1-16.

The Reversal of the Effective Stroke of the Labial Cilia of Sea Anemones by
Organic Substances. Amer. Journ. Physiol., XIV., 1-5.

THE INFLUENCE OF THE NERVE ON THE RE-
GENERATION OF THE LEG OF DIEMYCTYLUS.

CECIL SHEPARD HINES.

The following experiments were carried on at the suggestion
of Professor Morgan, for the purpose of ascertaining whether
regeneration in the leg of Dicinyctyhts is dependent on its con-
nection with the nervous system, as has been found in the case
of other urodeles, or whether the supposed result may not have
been due to unintentional injury to the blood supply. The hind
limb was chosen for operating on account of its larger size. The
general course of procedure was to cut the nerve in the upper
part of the leg without injury to the artery, and then amputate the
leg at the knee joint. After a period of a little more than three
weeks the new part can be clearly recognized as a dark protu-
berance sharply contrasting with the lighter color of the sur-
rounding skin.

In the first lots the nerve was cut as near the proximal end of
the femur as possible. A longitudinal slit through the skin was
made with a sharp knife. The muscles were then separated
until the nerve was brought into view. Care was taken not to
injure the blood vessel which closely adheres to the nerve and is
almost inseparable from it. If the operation were performed
without injury to the blood vessel and the leg showed a result-
ing paralysis it was amputated at the knee. For comparison an
equal number of salamanders had the leg cut off without injuring
the nerve or blood vessel. The results obtained from the first
series seemed to show that regeneration in Dicmyctylus was in no
way dependent upon the nerve. The proliferation of new ma-
terial began as soon in those in which the nerve had been cut as
it did in the checks. Nor did the amount of material regener-
ated seem to be affected in any way. There were, it is true,
great variations in the rate of regeneration, but these seemed to
arise from purely individual differences, and to bear no definite
relation to the presence or absence of the nerve connection. In
two checks operated upon on the same day and kept in the same

44

INFLUENCE OF NERVE ON REGENERATION. 45

aquarium, so that external conditions could play little, if any,
part in the result, an interval of ten days or more might occur
between the earliest and latest appearance of proliferation. The
results obtained after cutting the nerve at this level may have
been due to the presence of collateral nerve-connection sufficient
to give the required stimulus to the tissue ; for as was later clearly
shown, the nerve is an important factor in the regeneration.

In the succeeding series the nerve was cut through the pelvic
girdle close to the backbone, in the hope of more completely
cutting off the nerve supply of the limb. The incision necessarily
went through into the body cavity as the nerves given off from
the spinal cord lie close to the inner wall and are covered only
by ccelomic epithelium. The operation is by no means as serious
as would be imagined, since the wound heals completely in three
days. During this time the animal acts in a perfectly normal
manner except for the injured leg. Instead of using separate
individuals for checks as before, both legs of the same individual
were amputated, but the nerve was cut on one side only. The
males are much more difficult to operate upon than the females,
on account of their greater muscular development, whereas in
the female the pelvic girdle stands out prominently and the body
wall is thin. In the males the girdle is not visible externally and
is overlaid by a thick musculature which renders operating upon
the animal at this point difficult.

On November 26 a set of seven salamanders was operated
upon as described above. They were not again observed until
the forty-fifth day. At that time both legs of each individual
had proliferated material to a greater or less extent. The side
used as a check could be identified in every case but one, by its
far greater amount of proliferated material.

In the exception the animal assumed a peculiar green color,
evidently from a disease, and at a later period both legs were
entirely sloughed off. This set was continued under observation
until the eighty-third day, at which time there was still a decided
difference between the appearance of the two sides, although by
no means as pronounced as before. The check showed in each
individual the foot plainly differentiated, while in only two in-
stances was this the case upon the other side, and even in those
cases to a much less extent.

46 CECIL SHEPAKD HIKES.

Later I operated upon at least thirty salamanders in several
lots. Some of these were starved while others were well fed.
The abundance or lack of food did not seem to be a factor in the
rate of regeneration. Salamanders which were reduced almost
to "skin and bones" showed the same comparative amount of
regeneration as well-fed individuals. The influence of food did
show itself, however, in the amount of material proliferated.
During starvation an individual shrinks greatly in size and pro-
liferates much less material for the same relative amount of
regeneration as a well-fed companion.

To ascertain whether the blood supply was a factor in regen-
eration the artery of several individuals was cut just above its
entrance into the leg, the nerve being left intact. The result
showed that the leg regenerated at the normal rate. However,
not much stress can be laid upon this experiment, owing to the
rapidity with which a sectioned blood-vessel heals.

That the circulation in the leg may have continued to some
extent after the operation in those individuals whose nerve as
well as artery had been sectioned near the backbone was shown
conclusively by the following experiment. A salamander was
taken and a cut made in the pelvic region as before. Then a
vein was severed in the lower part of the leg. This continued
to bleed freely for a considerable time, as would not have occurred
had the total blood supply been cut off. The collateral blood
supply probably still brought blood to the limb. Similarly the
collateral nerve supply may in the first series of experiments
have sufficed to keep the regeneration up to the same rate as in
a limb in which the nerve was not cut.

In a number of cases regeneration did not set in at all on the
side on which the nerve had been cut. At least, after a period
of two months and a half there was not the least sign of prolifer-
ation, while in the normal course of regeneration the new part
appears in about twenty-five days. This lack of regeneration is
probably due to the distal end of the nerve being displaced, and
in consequence the regenerating nerve was unable to grow down
its old path along the degenerated nerve, but was turned aside.
Consequently not even a retarded regeneration occurred.

The most important work bearing on the question of the rela-

INFLUENCE OF NERVE ON REGENERATION. 47

tion between the nerve and the regeneration of the leg is that of
Wolff.1 He found that if the nerve-cord were destroyed in the
region of origin of the leg nerve that the leg regenerated at the
normal rate. Since the spinal ganglion was left after this opera-
tion, its presence may have sufficed to produce the result. In
fact, nerves were found in the new foot. In order to remove the
ganglion also, a piece of the spinal column was cut out including
cord and ganglion. Six individuals that survived this operation
showed that after the first proliferation of new material had taken
place growth came to a standstill for a time and then began
again. The result suggests, Wolff thinks, that the standstill was
due to the lack of nerve connection, while the renewal of growth
was due to the reestablishing of a new nerve connection. In
fact the disabled leg showed some signs of having regained its
power of motion.

Wolff discusses the question whether the period of standstill
may not have been due to the lack of function or activity of
the leg while its later growth was due to its regaining its loco-
motor function. He argues that this is probably not the case,
but that the nerve connection is directly responsible for the
result.

A student of Busfurth's, R. Rubin,2 has obtained similar
results. What part the nervous connection plays in these cases
is still obscure. Morgan and Davis3 have found that for the
regeneration of the tail of the tadpole, the presence of the noto-
chord and not the nervous system is the important factor.

COLUMBIA UNIVERSITY.

1 Virchow1 s Arthiv, CLXIX., 1902.

2 Archiv Enfw. Mec/i., XVI., 1903.
* Archiv Ent-w. Mech., XV., 1903.

Vol. X. January, 1906. No. 2

BIOLOGICAL BULLETIN

AN EXAMINATION OF THE METHODS FOR THE
MICROCHEMICAL DETECTION OF PHOS-
PHORUS COMPOUNDS OTHER THAN
PHOSPHATES IN THE TISSUES
OF ANIMALS AND PLANTS.

R. R. BENSLEY.
(From the Hull Anatomical Laboratory, University of Chicago. )

The microchemical reaction for the detection of phosphorus
in the tissues of animals and plants introduced in 1898 by Macal-
lurn ('98) is a modification of that devised in 1893 by Lilienfeld
and Monti ('93). These investigators attempted to demonstrate
the distribution of phosphorus in tissues by subjecting the latter
for some time to the action of a solution of ammonium molyb-
date in nitric acid, after which they were treated with a solution
of pyrogallic acid. The nitric molybdate reagent was supposed to
liberate the phosphorus from its organic combinations, to convert
it into orthophosphoric acid, and finally to precipitate the latter
as the yellow phosphomolybdate of ammonium. The further
treatment of the tissues with pyrogallic acid had for its object the
reduction of the ammonium phosphomolybdate to a lower oxide
of molybdenum, which, according to Lilienfeld and Monti had a
brown or black color. In this way the pale yellow precipitate
containing the phosphorus was converted into a dark-colored pre-
cipitate which could be easily studied under the microscope.

The importance of a microchemical reaction which would ena-
ble us to determine accurately the distribution of the compounds
of phosphorus not only in the tissues but also in the parts of the
cells of the tissues can hardly be overestimated. It is therefore

49

5O R. K. BENSLEY.

not surprising that the results of Lilienfeld and Monti have been
subjected to careful experimental examination by a number of
other investigators.

Raciborski ('93) in 1893, in his review of Lilienfeld and Monti's
article, showed that the reaction of ammonium phosphomolybdate
with pyrogallic acid resulted in the production of a green com-
pound, while ammonium molybdate gave by reduction with the
same reagent a brown compound. He concluded that the brown
reaction of Lilienfeld and Monti was due to ammonium molyb-
date mechanically imbibed by the section, and not to ammonium
phosphomolybdate.

Heine ('96), also, showed that phosphorus-free histone, pre-
pared from the thytnus, formed with the nitric-molybdate reagent
compounds from which the ammonium molybdate could not be
removed by washing in water, and in which it could be detected
by the use of reducing compounds. For this purpose he em-
ployed stannous chloride.

Macallum confirmed Raciborski's observations as to the color
compounds produced by the reduction of ammonium phospho-
molybdate and ammonium molybdate respectively, and showed
that ammonium molybdate could not be removed from tissues
which had been treated with the nitric-molybdate reagent even
by washing for several months in changes of distilled water.
Macallum perceived the necessity of substituting for pyrogallic
acid, which gives colored compounds with both the phospho-
molybdate and the molybdate of ammonium, some reagent which
would discriminate between these two compounds, and give a
color reaction with phosphomolybdate alone. This condition he
found to be fulfilled by zinc chloride, previously introduced for
this purpose by Polacci ('94), which gave a green color with the
phosphomolybdate but did not act on ammonium molybdate.
Owing, however, to the fact that zinc chloride acted very slowly,
he finally adopted as a reducing agent phenylhydrazin hydro-
chloride, which, according to him, made a very marked distinc-
tion, in the absence of alcohol or of caustic alkali, between the
molybdate and the phosphomolybdate compounds. It gave with
the former in powder the brown oxide at once, in solution, a
brownish precipitate, which appeared at once, or later, according

DETECTION OF PHOSPHORUS COMPOUNDS. 51

to the strength of the solution. On a solution of the molybdate
containing nitric acid, c. g., that used as the reagent for phos-
phoric acid, it had no apparent effect on the molybdenum com-
pound, although, in a few minutes, a soluble, reddish, aromatic
compound might be formed in the solution. On the other hand,
with phosphomolybdates, either in the presence or in the absence
of ammonium molybdate, or of nitric acid, or of both, it gave at
once the dark green oxide of molybdenum.

Concerning the use of these reagents on tissues Macallum
says: " On the molybdate and phosphomolybdate compounds
distributed in animal and vegetable tissues, the phenylhydrazin
hydrochloride acts as it does on these in the test-tube. It is not
necessary to free the tissue preparations from ammonium molyb-
date." He recommends washing the preparations for a minute
or two in a dilute solution of nitric acid after which they are
transferred to the reducing solution, which in less than two min-
utes, brings out the green color where the phosphomolybdate
compound occurs, but a faint yellow reaction where ammonium
molybdate alone is present.

The technique of the reaction is as follows: Fresh tissues or
tissues hardened in alcohol were used. Pieces of tissue or thin
sections in the case of hardened material, were placed, for a
period varying from ten minutes to forty-eight hours, in a solution
of ammonium molybdate in nitric acid, prepared by dissolving
one part of pure molybdic acid in four parts of strong ammonia,
and adding thereto, slowly, fifteen parts of nitric acid, sp. gr. 1.2.
After the nitric molybdate reagent has acted for a sufficient
length of time, the preparations are washed in water or in dilute
nitric acid, and treated with a ^ per cent, solution of phenyl-
hydrazin hydrochloride, which reduces the phosphomolybdate to
a -green-colored oxide of molybdenum.1 The tissues may be
then dehydrated, cleared in. oil of cedar and mounted in
balsam.

According to Macallum, inorganic compounds of phosphorus

'Although Macallum speaks of a green oxide of molybdenum being formed by this
reaction, it is probable that the bodies formed belong to the blue oxides. The green
color obtained at the beginning of the reduction of phosphomolybdate of ammonium
in vitro is due to the yellow background of unreduced molybdate, that obtained in
the tissues to the associated xanthoproteic reaction (vide infra).

52 R. K. BENSLEY.

are first affected, then lecithins, and finally the organic compounds
of phosphorus. Where it is desired to demonstrate the distribu-
tion of the latter he recommends the preliminary removal of the
lecithins by repeated extraction with hot ethyl alcohol in a
Soxhlet apparatus.

In the original form devised by Lilienfeld and Monti or in the
form of Macallum's modification this reaction for phosphorus has
been extensively employed. Macallum's original paper dealing
with the reaction contains a considerable number of contributions
dealing with the distribution of organic compounds of phosphorus
in various tissues, and the reaction has been applied to the solu
tion of special problems of this nature by Sherrington ('94),
Gourlay ('94), Scott ('99), Held ('95), Bensley ('03), Wager
('05), Richter ('05), and many others. It is therefore of the
utmost importance that every detail of the reaction should be
carefully tested to exclude all possible sources of error.

Recently I have obtained results which have led me to suspect
that the reaction obtained by Macallum's method is not wholly
due to the formation in the tissue of ammonium phosphomolyb-
date, but that other compounds of molybdenum may be present
which are capable of reduction to the blue oxide by means of
phenylhydrazin hydrochloride. For example, I observed that
the peripheral portions of sections gave uniformly a deeper and
more diffuse reaction than the central portions. This result was
first noted in the tips of the villi in sections of the small intestine,
and was ascribed to the presence of inorganic phosphates ab-
sorbed from the food. Later it was noted that the same result
was obtained in sections of the liver, pancreas, and other organs.
Furthermore, I observed that freshly prepared solutions of the
nitric molybdate often gave a strong reaction in the tissues after
very short periods of immersion. For example, sections of the
fundus region of the stomach of the rabbit, treated with warm,
freshly prepared nitric molybdate reagent for ten minutes, then
washed in water and reduced by means of a one per cent, solu-
tion of phenylhydrazin hydrochloride gave a diffuse bluish green
reaction together with a strong reaction in the nuclei and in the
granules of the parietal cells. This result was clearly not due to
phosphorus compounds of any sort, because the same nitric

DETECTION OF PHOSPHORUS COMPOUNDS. 53

molybdate reagent after having been kept several days, during
which it had deposited copious crusts of molybdic acid, gave,
when applied to sections from the same source, no such result,
the reaction proceeding in the slow progressive manner character-
istic of the phosphorus reaction as described by Macallum.
Again, in his investigation of the nature of the 'granule cells of
Paneth, Mr. Klein, working under my direction, found it neces-
sary to employ formalin in the fixation of the tissues in order to
preserve the granules. In preparations of this material treated
by Macallum's method, we were surprised to obtain a strong reac-
tion in the fibrils of the collagenic tissue of the tela submucosa.

Clearly, the intense marginal reaction obtained in sections and
the early reaction obtained when freshly prepared solutions of
the nitric molybdic reagent were used could not be due to phos-
phorus. These anomalous characters of the reaction as applied
to sections could only be explained on the assumption that,
after the treatment with the nitric molybdic reagent, there existed
in the tissue compounds of molybdenum other than phospho-
molybdates, which gave the blue reaction with phenylhydrazin
hydrochloride.

On account of the fact that the difference between freshly pre-
pared and older solutions of the nitric molybdic reagent is to be
found in the amount of molybdic acid contained, it seemed prob-
able that the extraordinary results of the reaction, as described
above, were due to the absorption of this substance by the tissue
elements. Accordingly, I undertook experiments to determine
the behavior of solutions of molybdic acid in reaction with
phenylhydrazin hydrochloride, as well as the capacity of the
tissue elements for absorbing it from its solutions. Later it was
found necessary to reinvestigate the reaction obtained by treating
ammonium molybdate in solution, and phosphomolybdate of
ammonium suspended in water, respectively with solutions of
phenylhydrazin hydrochloride.

Although the experiments were started with the expectation
that a portion of the reaction would be found to be due to ab-
sorbed molybdic acid, I still thought at that time that the funda-
mental assumption was true, upon which the reaction was based,
namely, that the organic phosphorus was liberated from its com-

54 R- R- BENSLEY.

binations by the reagent, converted into orthophosphoric acid and
immediately precipitated in situ as ammonium phosphomolyb-
date. As a result of the experiments, however, I have been
forced to the conclusion that the whole of the reaction obtained
by Macallum's method is due to compounds of molybdenum
other than phosphomolybdate and that the phosphorus of the
tissues is not concerned in the production of the reaction at all.
For the purpose of testing the reactions of molybdic acid with
phenylhydrazin hydrochloride, I prepared two soluble molybdic
acids. The first of these was prepared by the method recom-
mended by Ullik ('67). Barium molybdate, prepared by pre-
cipitating a warm solution of ammonium molybdate with barium
chloride washing thoroughly with hot distilled water, and dry-
ing the precipitate on a water bath at 100° C., was suspended
in water and decomposed with its equivalent of sulphuric acid.
The solution was then filtered, tested for barium, sulphuric acid
and chlorides, from which it was found to be free, and the total
acidity was determined by titration with a normal solution of
sodium hydroxide, using phenolphthalein as indicator. Assum-
ing that the solutions contained a molybdic acid having the for-
mula H2Mo2O7 the concentration of the solutions obtained by the
method described above was in the case of one preparation 5.25
per cent., in another 7.52 per cent. The second molybdic acid
was colloidal molybdic acid prepared by the process recommended
by Graham (64), except that ammonium molybdate was employed
instead of sodium molybdate. A solution of ammonium molyb-
date in hydrochloric acid was dialyzed for several days against
distilled water, until free from chloride. The resulting solution
was then titrated against normal soda solution using phenolphtha-
lein as indicator. According to Sabanejew ('90) the molecular
weight of the molybdic acid prepared by Graham's method as
determined by lowering of freezing point is 620 corresponding to
the formula H2Mo4O13 . Assuming that the same compound was
obtained by the dialysis of the solution of ammonium molyb-
date in hydrochloric acid the solution obtained contained 6.73
per cent, of colloidal molybdic acid. From these solutions were
prepared the various solutions mentioned in the succeeding ex-
periments.

DETECTION OF PHOSPHORUS COMPOUNDS. 55

Solutions of both molybdic acids give, when treated with
phenylhydrazin hydrochloride an immediate blue reaction which
gradually deepens in color and a blue precipitate forms.

Sections of tissues fixed in alcohol, cut in paraffin, and fastened
to the slide by the water method were placed in each of the solu-
tions of molybdic acid. From time to time sections were re-
moved from the solution, rinsed in water, and tested with a I per
cent, solution of phenylhydrazin hydrochloride. It was found
that the molybdic acid was taken up by the tissues from both
solutions and was detectable in them by the blue reaction obtained
by reduction with phenylhydrazin hydrochloride. In sections
treated with pure solutions of soluble or of colloidal molybdic
acid the strongest reaction was obtained in the collagenic fibrils
which were deep blue. A slight diffuse reaction was obtained in
the cytoplasm of cells, and a somewhat stronger reaction in the
nuclear chromatin. The amount, however, of molybdic acid
taken up from dilute pure solutions was not great except as
regards the collagenic tissue. The experiments show, however,
that molybdic acid may be taken up from its solutions by tissues
and may be detected in these by the blue reaction produced by
treatment of the sections with phenylhydrazin hydrochloride.

Under the conditions of the Lilienfeld-Monti-Macallum reaction
molybdic acid occurs in the solution associated with nitric acid as
well as with ammonium molybdate, ammonium nitrate, and the
products of dissociation of all these compounds. Accordingly,
the effect of the presence of acids on the absorption of the
molybdic acid from its solutions by tissues was tested. Sections
were placed in solutions of soluble and of colloidal molybdic acid
to which five per cent, of nitric or of hydrochloric acid had been
added, and were tested from time to time with phenylhydrazin
hydrochloride. I found that the addition of either nitric acid or
hydrochloric acid to the solutions of molybdic acid produced a
remarkable increase in the capacity of sections for combining
molybdic acid, which was again detectable by the blue or green
reaction obtained by reduction with phenylhydrazin hydro-
chloride. With the mixture of nitric and molybdic acids the
reaction obtained after reduction was a deep greenish blue color
in the nuclear chromatin, a faint greenish blue in the cytoplasm,

56 R. R. BENSLKY.

and a deep blue in the collagenic fibers. Except for the strong
reaction in the connective tissue, the result obtained by treatment
of sections with a solution of molybdic acid containing nitric acid,
followed by reduction in one per cent, phenylhydrazin hydro-
chloride was exactly similar to the so-called phosphorus reaction
obtained by the procedure recommended by Macallum. It may
be noted that the color obtained by the use of molybdic acid con-
taining nitric acid followed by reduction differed from that pro-
duced by reduction of molybdic acid by phenylhydrazin in the
test tube, inasmuch as the former gives a greenish blue color, the
latter a pure blue. This difference was obviously due to the
yellow background afforded by the xanthoproteic reaction. As
in the phosphorus reaction, the absorption of the molybdic acid
was progressive, the reaction after eighteen hours being much
stronger than after three hours.

Similar results were obtained with solutions of molybdic acid
containing hydrochloric acid, except that the molybdic acid was
taken up much more rapidly from the hydrochloric solution than
from the nitric solution, and that the resulting reaction was blue
rather than greenish blue, owing to the absence of the yellow
zanthoproteic reaction. A further difference exhibited itself in
the fact that sections left for some time in the solutions developed
the blue color witJiout the use of any reducing' agent, the organic
compounds of the tissue evidently acting as reducers. In the
presence of nitric acid this, of course, could not occur because of
the strong oxidative action of this compound.

Thus, in sections treated with solutions of molybdic acid con-
taining either hydrochloric, or nitric acid, followed by phenyl-
hydrazin hydrochloride, results were obtained which were the
exact counterpart of the results of the so-called phosphorus
reaction, although there could be little possibility of the form-
ation of precipitates of ammonium phosphomolybdate in the
tissues. Curiously enough, the anomalous characters occasion-
ally observed in the reaction obtained by Macallum's method
wene to be found in the molybdic acid material, that is to say,
the more intense diffuse reaction of the outer portions of the sec-
tions and the deep reaction in the connective tissue. It seemed
clear from these experiments that a portion, at least, of the result

DETECTION OF PHOSPHORUS COMPOUNDS. 57

obtained by the procedure of Macallum was due to absorption of
tnolybdic acid from the nitric molybdate solution. There was,
however, some possibility that even this reaction with molybdic
acid solutions depended on the presence of phosphorus and its
liberation as phosphoric acid. It might be supposed that this
phosphoric acid reacted with the molybdic acid to produce
phosphomolybdic acid which was in turn precipitated by the
albumens of the tissues. Or it might even be supposed that
ammonium phosphomolybdate was formed, the ammonium ions
necessary to the reaction being furnished by the albumens. That
this is not the case, and that the reaction obtained by the use of
molybdic acid solutions is in no way dependent on the phosphorus
content of the tissue, I think the following experiments will
show.

In studying the reaction of solutions of molybdic acid with
phenylhydrazin hydrochloride, I found that the addition of nitric
acid to the mixture retarded the reaction and if a sufficient quan-
tity were present prevented it altogether. Accordingly, experi-
ments were undertaken to determine the limits of this reaction
with molybdic acid, ammonium molybdate, and ammonium
phosphomolybdate, respectively, in the hope that a sufficient dif-
ference in the behavior of these compounds would be discovered
to enable one to employ a solution of phenylhydrazin hydro-
chloride containing enough nitric acid to inhibit the reduction of
molybdic acid and of ammonium molybdate while permitting the
reduction of ammonium phosphomolybdate, and thus discrim-
inate between these compounds occurring in tissues treated by
Macallum's methods.

I found that with a constant concentration of phenylhydrazin
hydrochloride, the amount of nitric acid required to prevent the
reduction to the blue oxide of molybdenum varied directly with
the concentration of the molybdic acid but was constant for any
given concentration.

I found, moreover, that the blue reduction was invariably
obtained in solutions of ammonium molybdate containing nitric
acid, when they were treated with solutions of phenylhydrazin
hydrochloride, provided the amount of nitric acid did not exceed
a certain amount, which, as in the case of molybdic acid, was

58 R. R. BENSLEY.

constant for a given concentration of the other two constituents
but varied directly with the concentration. This fact made it
necessary to determine the conditions of this reaction with
ammonium molybdate if solutions of the hydrochloride con-
taining nitric acid were to be used for the reduction of the sec-
tions, because in trying to eliminate the reduction of the molybdic
acid by using a solution of the hydrochloride containing nitric
acid, a new source of error might be introduced, inasmuch as the
nitric acid would make ammonium molybdate available to the
blue reduction.

The experiments to determine the limits of these reactions
were made in test-tubes in much the same way as experiments
in haemolysis are carried out. In each of a series of test-tubes
was placed, from a pipette graduated in fiftieths of a cubic centi-
meter, a measured quantity of the solution of molybdic acid.
To this was added a quantity of nitric acid solution of known
strength increasing by increments of one tenth of a cubic centi-
meter from tube to tube. Sufficient distilled water was then
added to make the contents of each tube up to 9. 5 c.c., and finally
0.5 c.c. of a two per cent, solution of phenylhydrazin hydro-
chloride was added to each tube. In such a series of tubes, if a
sufficient range of concentrations of the nitric acid was included
the tubes at one end of the series would give the blue reduction,
those at the other would at first show no signs of reaction, but
after several hours would develop a brownish color. At some
point in the series two tubes, side by side, differing from one
another only in the nitric acid content, would present the one a
blue, the other a brown color. When solutions of nitric acid
containing 32 grammes of nitric acid per looc.c. were employed,
that is, when the difference in the contents of two tubes amounted
to .032 gr. of nitric acid, the contrast in color between the tubes
which marked the limit of the reaction was a very striking one.
Attempts to define this limit more accurately by the use of more
dilute solutions of nitric acid, and accordingly smaller increments,
of nitric acid from tube to tube did not result more satisfactorily.
For example in a series of tubes in which the increment of nitric
acid was 0.012 gr. from tube to tube, the transition was distrib-
uted over several tubes, those immediately preceding the first

DETECTION OF PHOSPHORUS COMPOUNDS. 59

tube free from the blue precipitate containing a slight blue pre-
cipitate which subsided after several hours leaving a brownish
supernatant fluid. Thus, the possible error in determining the
proportion of nitric acid necessary to prevent the blue reaction is
considerable, although the results are accurate enough for the
purposes of this investigation, as subsequent statements will show.

REACTIONS OF SOLUBLE MOLYBDIC ACID WITH PHENYL-

HYDRAZIN HYDROCHLORIDE IN THE PRESENCE

OF NITRIC ACID.

Percentage of Phenyl- Percentage of Nitric

Strength of Molybdic Acid hydrazin Hydro- Acid Necessary to Pre-

in Fractions of Ncrmal. chloride. vent Blue Reaction.

O.O2 O.I 1.38

0.04 o.i 2.26

0.06 o.i 2.94

0.08 o.i 3.27

REACTIONS OF SOLUTIONS OF AMMONIUM MOLYBDATE WITH
PHENYLHYDRAZIN HYDROCHLORIDE IN THE PRESENCE

OF NITRIC ACID.

Solutions of ammonium molybdate when treated with phenyl-
hydrazin hydrochloride give, as stated by Macallum, a brown
color, and a brown precipitate slowly forms in the solution. In
the presence of nitric acid, however, provided the latter does not
exceed a certain amount which varies with the concentration of
the molybdate, the reaction consists in the production first of a
blue color and finally of a deep blue precipitate. If the amount
of nitric acid exceeds this maximum no reaction occurs at first,
but a brown color slowly develops in the solution. The follow-
ing table gives the concentration of nitric acid necessary to pre-
vent wholly the blue reaction with several concentrations of
molybdate.

Percentage of Ammonium Percentage of Phenylhydrazin Per Cent. Nitric Necessary to

Molybdate. Hydrochlotide. Prevent Blue Reaction.

0.5 o.i 3.14

I.O O.I 4.40

1.5 o.i 5.03

2.0 O.I 5.66

6O R. R. BENSLEY.

REACTIONS OF PHOSPHOMOLVBDIC ACID AND PHOSPHOMOLYB-
DATE OF AMMONIUM WITH PHENYLHYDRAZIN
HYDROCHLORIDE.

Phenylhydrazin when added to a solution of phosphomolybdic
acid or to crystals of ammonium phosphomolybdate suspended in
water gives at once the reduction to the blue oxide. The reac-
tion under these conditions proceeds so rapidly that it is difficult
to follow the steps. The resulting body is not green in color as
described by Macallum but blue, the green color which is first
seen when phenylhydrazin hydrochloride is added to the crystals
of ammonium phosphomolybdate being simply due to the yellow
background afforded by the unreduced phosphomolybdate.
After a few minutes the reaction proceeds still further and a solu-
ble blue-violet compound is formed.

In the presence of nitric acid, the reaction proceeds somewhat
more slowly, although it is less sensitive to the presence of nitric
acid than the corresponding reactions with molybdic acid and
ammonium molybdate. With crystals of phosphomolybdate sus-
pended in water and o. I per cent, of phenylhydrazin the concen-
tration of nitric acid may be increased to 36 per cent, without the
reduction to the blue oxide being prevented. Even in the pres-
ence of this great amount of nitric acid the reduction of the
phosphomolybdate reaches a maximum within ten minutes after the
phenylhydrazin hydrochloride is added. The reaction between
phenylhydrazin and phosphomolybdic acid proceeds in much
the same way, and is similarly much less sensitive to the
presence of nitric acid, than the corresponding reactions with
molybdic acid and ammonium molybdate. For purposes of
comparison the experiments were made the results of which are
presented in the subjoined table.

Percentage of Phosphomolyb- Percentage of Phenylhydrazin Per Cent. Nitric Acid Necessary
die Acid. Hydrochloride. to Prevent Blue Reaction.

o.i o.i 6.55

O.2 O.I 9.82

0.3 o.i 13.10

0.4 o.i J4-74

On comparison of this table with the preceding one it will be
seen that the reduction of phosphomolybdic acid having a con-

DETECTION OF PHOSPHORUS COMPOUNDS. 6 1

centration of o. I per cent, will proceed in the presence of nitric
acid having a concentration sufficient to prevent reduction in a
solution of ammonium molybdate of 2 per cent, strength.

While it would have been difficult to draw from these experi-
ments conclusions as to the probable behavior of the compounds
of molybdic acid in the tissues when treated with solutions of
phenylhydrazin hydrochloride containing nitric acid, yet they
suggested a possibility which was capable of being proved experi-
mentally that the compounds of molybdic acid and molybdates
found in the tissues would fail to react to phenylhydrazin hydro-
chloride in the presence of an amount of nitric acid which would
have no effect on the reduction of ammonium phosphomolybdate.

In order to test this question, sections of the liver of Necturus
prepared after fixation in alcohol and fastened to the slide by the
water method were treated with Macallum's nitric molybdate re-
agent, a solution of soluble molybdic acid in 10 per cent, nitric
acid, and a ten percent, solution of phosphoric acid, respectively.
The two first mentioned solutions were allowed to act for three
hours at 37.5° C. followed by eighteen hours at ordinary room
temperature. They were then tested with a o. I per cent, solution
of phenylhydrazin hydrochloride and found in each case to give a
strong reaction corresponding in its characters and distribution
to the phosphorus reaction of Macallum. The reaction obtained
in the sections treated with the solution of molybdic acid was much
the stronger. Other sections from the same lot were then treated
with solutions containing o. I per cent, of phenylhydrazin hydro-
chloride and varying known quantities of nitric acid, in each case
for a period of fifteen minutes. The sections from the molybdic
acid solution and from the nitric molybdate reagent were treated
side by side in the same solution, and for purposes of control a
section which had been soaked in phosphoric acid and then treated
with the nitric molybdate reagent was also put at the same time
in the solution, so that it was possible to observe the effect of
different concentrations of nitric acid on the reduction of sections
treated with molybdic acid, or with the nitric molybdate reagent,
and of sections containing ammonium phosphomolybdate arti-
ficially introduced. It was my expectation that a low concentra-
tion of nitric acid would suffice to abolish that portion of the reac-

62 R. R. BENSLEY.

tion which was due to ammonium molybdateand to molybdic acid,
and that a considerable residuum of the reaction would be found
unaffected by even high concentrations of nitric acid and could
thus be interpreted as a true phosphorus reaction. I was quite un-
prepared for, and greatly disappointed at the actual result of these
experiments, namely, that relatively low concentrations of nitric
acid abolished the reaction altogether.

With a concentration of 3.27 per cent, of nitric acid, phenyl-
hydrazin hydrochloride o.i per cent., the molybdic acid sections
and the nitric molybdate section showed no reaction after three
minutes' treatment, although the section containing ammonium
phosphornolybdate artificially introduced gave a maximum reac-
tion in less than one minute. After fifteen minutes' action, a
very faint reaction was obtained in the nuclei, both in the mo-
lybdic acid section and in the nitric molybdate section. When
the concentration of the nitric acid reached 16.37 Per cent., the
phenylhydrazin remaining the same, the reaction was not recog-
nizable after fifteen minutes' treatment although sections contain-
ing ammonium phosphomolybdate artificially introduced reduced
to a maximum depth of color in the same solution in five minutes.
I have repeated these experiments many times, always with the
same results. It is significant that the reaction disappeared at
exactly the same point as regards concentration of nitric acid in
the molybdic acid section and the nitric molybdate section.

Only one conclusion is possible from these experiments,
namely, that sections after treatment with Macallum's reagent
for this length of time did not contain appreciable quantities of
ammonium phosphomolybdate. Thus the fundamental assump-
tion on which the reaction of Lilienfeld and Monti and of Macal-
lum is based falls to the ground. It is obvious that if the phos-
phorus of the organic compounds is liberated at a point short of
the destruction of the recognizable structures of the cell, it is not,
at all events, precipitated in situ by the nitric molybdate reagent.

As a result of these experiments I am of the opinion that the
reaction obtained by Macallum's procedure is entirely due to the
formation of compounds of molybdic acid with the albumens of
the tissue and not in any respect to the formation of ammonium
phosphomolybdate at the expense of the organic phosphorus.

DETECTION OF PHOSPHORUS COMPOUNDS. 63

The facts on which the conclusion is based are, briefly, as
follows :

The essential conditions of a successful phosphorus reaction
are, first, that the phosphorus may be liberated from its organic
combinations at a point short of the destruction of the recogniz-
able structure of the cell ; second, that the liberated phosphorus
be precipitated at once at the point of origin as ammonium phos-
phomolybdate ; third, that the reducing substance employed to
make the phosphomolybdate visible for microscopic study act on
phosphomolybdate and on no other compound of molybdenum
which may be present in the tissue.

Phenylhydrazin hydrochloride does not meet the third condition
because it reduces to the blue oxide of molybdenum, soluble
molybdic acid in the test tube as well as molybdic acid combined
with the tissue constituents in sections.

Phenylhydrazin hydrochloride also produces the blue oxide
when treated with ammonium molybdate in the presence of nitric
acid, provided that the latter does not exceed a certain concen-
tration which is constant for constant concentrations of the mo-
lybdate.

Nitric acid affects the reduction of molybdic acid, ammonium
molybdate, and ammonium phosphomolybdate, by phenylhydra-
zin hydrochloride in the same way, namely, retards the reduc-
tion, but to different degrees, inasmuch as low concentrations of
nitric acid prevent the reduction of the two former to the blue
oxide, while high concentrations of nitric merely retard the blue
reduction of the phosphomolybdate. Accordingly, if phospho-
molybdate is formed at the site where a reaction is obtained by the
method of Macallum, the reaction ought to be elicited by treat-
ment of the sections with solutions of phenylhydrazin in having
a high content of nitric acid. This, however, the experiments
show is not the case. Even low concentrations of nitric acid
eliminate the greater portion of the reaction, and the reaction is
entirely abolished by a nitric acid content which has little effect
on the reduction of phosphomolybdate of ammonium artificially
introduced into sections for purposes of control. Furthermore,
the reaction is abolished at the same concentration of nitric acid
with sections treated with Macallum's nitric molybdate reagent

64 R. R. BENSLEY.

as with sections treated with a pure solution of molybdic acid.
These facts dispose finally of the first and second essential con-
ditions of a successful microchemical reaction for organic phos-
phorus, for it is clear that if the sections after treatment with
the reagent contain no phosphomolybdate of ammonium, that
the organic phosphorus has either not been liberated from its
compounds, or that, if it has, it has not been precipitated at
the moment and at the point of liberation. If these conclu-
sions are correct, it is also obvious that there is no hope of a
real phosphorus microchemical reaction being obtained by the
employment of the nitric molybdate reagent.

These conclusions do not, of course, apply to the identifica-
tion of phosphates by the nitric molybdate reagent, in cases
where the characteristic crystal form of the ammonium phospho-
molybdate can be recognized under the microscope.

In conclusion, it may be mentioned that in making the experi-
ments to determine the effect of nitric acid on the reduction of
the molybdenum compounds by phenylhydrazin hydrochloride
it is important to employ solutions which are free from nitrous
acid, which reacts with the phenylhydrazin and reduces its con-
centration.

BIBLIOGRAPHY.

Bensley, R. R.

'03 The Structure of the Glands of Brunner. The Decennial Publications, Uni-
versity of Chicago, Vol. X., 1903, pp. 279-326.

Gourlay, F.

'94 Proteids of the 1'hyroid and Spleen. J. Physiol., Cambridge, Vol. 16,1894,

PP- 23-33-
Graham, T.

'64 On the Properties of Silicic Acid and other Analogous Colloidal Substances.
J. Chem. Soc., Lond., n. s., Vol. 2, 1864, pp. 318-327.

Heine, L.

'96 Ueber die Molybdansaure als mikroskopisches Reagenz. Ztschr. f. physiol.
Chem., Strassburg, Bd. 22, 1896-97, pp. 132-136.

Held, H.

'95 Beitrage zur Structur der Nervenzellen undihrer Fortsatze. Arch. f. Anal. u.
Entwicklngsgesch., Leipzig, 1895, pp. 396-414.

DETECTION OF PHOSPHORUS COMPOUNDS. 65

Lilienfeld, L., u. Monti, A.

'93 Ueber die mikrochemische Lokalisation des Phosphors. Ztschr. f. physiol.

Chem., Strass., Bd. 17, 1893, pp. 410-425.
Macallum, A. B.

'98 On the Detection and Localization of Phosphorus in Animal and Vegetable
Tissues. Proc. Roy. Soc., Lond., Vol. 63, 1898, pp. 467-479.

Polacci, G.

'94 Sulla distributzione del fosforo nei tessuti vegetali. Malpighia, Vol. 8, 1894,

PP- 36I-379-
Raciborski, M.

'93 Review of Lilienfeld and Monti's Article (vide supra). Botanische Zeitung,
Jahr. 51, 1893, pp. 245-247.

Richter, 0.

'05 Die Fortschritte der botanischen Mikrochemie seit Zimmermanns Botanischer
Mikrotechnik. Ztschr. f. wissensch. Mlkr., Leipz., 1905, Bd. 22, pp. 194-
261.

Sabanejew.

'90 Quoted after abstract in Ber. d. deutsch. Chem. Gesellsch., Berl., 1890,
Bd. 23, Referate, p. 87.

Scott, F. H.

'99 The Structure, Michrochemistry and Development of Nerve Cells. Tr.

Canad. Inst., Toronto, Vol. 5, 1899, PP- 4°5~433-
Sherrington, C. S.

'94 Note on some Changes in the Blood of the general Circulation consequent
upon certain Inflammations of acute and local Character. Proc. Roy. Soc.,
Lond., Vol. 55, 1894, pp. 161.

Ullik, F.

'67 Untersuchungen iiber Molybdansaure und deren Salze. Annalen d. Chem. u.

Pharm. herausg. v. Wohler, Liebig, u. Kopp, Heidelb. , Bd. 144, 1867, pp.

320-351-
'70 Ueber Molybdansaure und ihre Verbindungen. Ibid., Bd. 153, 1870, pp.

369-376.
Wager, H.

'05 On some Problems of Cell Structure and Physiology. Sectional Address,

Section K, Botany, British Association for the Advancement of Science,

English Mechanic and World of Science, No. 2111, 1905, pp. 102-108.

ON THE LOCOMOTION OF A SEA ANEMONE
(METRIDIUM MARGINATUM).1

J. F. McCLENDON.

Last winter while studying some animals in the marine aquaria
of the University of Pennsylvania I noticed that the anemones,
after being placed at the bottom of an aquarium, would creep up
the side of the glass to a more favorable position. Their method
of progression is similar to the ordinary creeping of a snail,2 con-

FlG. I. AL'tridiiun inai^ino/iini seen through the glass, up which it is creeping.
The lower side of the photograph has been outlined.

sisting of a succession of waves that travel from behind forward,
but in the anemone the waves are larger and not so rapid or
regular. The accompanying photographs were taken of an
anemone creeping up the side of an aquarium, with its distal
end inclined forward, probably to test the water into which it
was advancing. The undulations of the foot progress in the
direction of locomotion. If the functionally posterior end of the

1 Contribution from the Zoological Laboratory of the University of Pennsylvania.

2 For a more complicated form of locomotion in some snails see A. J. Carlson :
"The Physiology of Locomotion in Gasteropods," BIOL. BULL., Vol. 2, January

1905, pp. 85-93.

66

LOCOMOTION OF A SEA ANEMONE. 67

foot be watched closely it will be seen to let go at several points
(below in Fig. i ) and slip forward. This contraction is carried
forward and on reaching the center of the foot, the contracted
portion rises up from the glass, forming a wave that deepens as

FlG. 2. Metridiiim marginatum about half a minute later than Fig. I.

it approaches the "anterior" end (Fig. 2, above). On reaching
the " anterior " edge the wave is retarded by the firmer attachment
of the edge, which releases locally, breaking the wave into seg-
ments (Fig. i, above). A wave requires about a minute to trav-
erse the foot of the anemone, and before it has disappeared,
another commences.

I threw a number of anemones into an aquarium to observe
their actions. They threw out acontia, which caught hold of
any solid near them and contracted until some portion of the foot
touched the object and caught hold. One anemone sinking to
the bottom and resting on its tentacles contrived to rierht itself

o

by suddenly contracting and expelling water from its mouth. I
observed this once or twice subsequently, but rather think it a
coincidence than a common reflex.

BIOLOGICAL HALL, UNIV. OF PENN.,
December 5, 1905.

THE SIGNIFICANCE OF SCUTE AND PLATE
"ABNORMALITIES"1 IN CHELONIA.

A CONTRIBUTION TO THE EVOLUTIONARY HISTORY OF THE
CHELONIAN CARAPACE AND PLASTRON.

H. H. NEWMAN.
CONTENTS.

PART I.

PAGE.
I. Introduction.

1. Statement of the Problem 68

2. Nomenclature. The Normal Plate and Scute Pattern: (a) Plates;

(/;) Scutes 69

II. Discussion of the Present Status of the Question Concerning the Morph-
ology of the Chelonian Armor.

1. Paleontological Data 71

2. Embryological Data 72

3. Data Derived from Comparative Anatomy 74

III. Description and Discussion of Abnormalities.

1. Inframarginals 77

2 . In terpl astrals 80

3. Supernumerary Scutes in a Row on the Carapace 81

4. Supernumerary Scutes in a Row on the Plastron 93

5. Correlated Abnormalities in Scutes and Bony Plates, with Discussion

Regarding the Relationship Between these Two Structures 94

6. Correlation in Scute Abnormalities 9&

I. INTRODUCTION.

I . Statement of the Problem.

During a residence of several years on Lake Maxinkuckee,
Marshall County, Indiana, my attention was repeatedly attracted
to the large numbers and variety of species of tortoises that
are found in the lake and in its accessory streams, swamps
and pools.

In the spring of 1903 I began a study of the habits, variations,

'The word "abnormalities" in the title is used for lack of a better one, and
includes supernumerary scutes and plates, deficiencies in these structures and cases
of fusion. The word " diversities " might have been used with equal appropriateness.

68

"ABNORMALITIES" IN CHELONIA. 69

etc., of these species, involving the collection of large numbers of
individuals of all sizes. One of the most striking phenomena
that came to light was the prevalence of many kinds of scute
abnormalities, consisting for the most part of supernumary scutes
on the carapace and plastron. As examples multiplied I became
aware of a marked degree of regularity in these abnormalities,
the same supernumerary scutes occurring in exactly the same
locations time after time.

Diversity in scutes had been noted by two observers. Gadow
('99) studied Tlialassochelys carctta (L.), a species with no fixed
number or arrangement of scutes, and Parker ('01) found two
abnormal specimens of Chelopus insculptus (Le C.) on the basis
of which he published a paper on correlated abnormalities in the
scutes and the bony plates.

It seemed, then, that this phenomenon needed further investi-
gation and the collection of large numbers of abnormal speci-
mens was begun in the hope of reaching a rational explanation
of this very prevalent diversity. Careful study has convinced
me that these abnormalities are to be considered not as meaning-
less anomalies but as examples of systematic atavism in the sense
of deVries. From this standpoint it seems possible to throw
some light on the phylogeny of Chclonia.

The color patterns are intimately associated with the scutes
and throw much light on their phylogeny. Consequently a brief
consideration of chelonian coloration has been appended.

2. Nomenclature. The Normal Plate and Scute Pattern.

The following description and appended drawings (Plate I.,
Figs, i and 2), although referring particularly to an adult female
specimen of Graptemys geographica, will apply to any genus of
the Emydidae. Fig. I represents the dorsal and Fig. 2 the ventral
aspect.

The armor of tortoises consists of two elements, bony plates
and horny scutes, which for brevity will be referred to as plates
and scutes. Dotted outlines are used for the plates and solid
outlines for the scutes. In labeling, small letters are used for
plates and capital letters for scutes.

A. Plates. — There are in the carapace (Fig. i) five longitudinal

7O H. H. NEWMAN.

rows of plates, a single median and two paired rows. The median
row has been variously designated as dorsal, vertebral and neural.
In this paper the term neural will be invariably used. The neural
row consists of the following elements : an anterior plate of large
size called the nuchal (;/?/.), eight neurals (n. i-S), two procaudals
(pr. I and 2), and posteriorly the pygal (/>.).

Lateral to the median row are the paired costals (c. I— 8),
directly overlying the eight pairs of ribs.

Bordering the carapace on both sides and extending from
nuchal to pygal are the marginals (;//. i — 1 1).

The plates of the plastron (Fig. 2) are nine in number — the
paired epi- (V.), hyo- (ho.}, hypo- (hp.}, and xiphi- (.r.) plastrals,
and the unpaired endo-plastral (en.}.

The hyo- and hypoplastrals articulate directly with the fourth,
fifth and sixth marginals and form the so-called " bridge ' be-
tween the dorsal and ventral armor.

B. Scutes. - - On the carapace (Fig. i) there are, as in the case
of the plates, five longitudinal rows of scutes that receive the
same names as the plates. The median row, neurals, consists
of a small anterior element, the nuchal (NU.}, and five large
neurals (N. 1-5). There are four pairs of large costals (C. 1-4).
Twelve pairs of marginals (M. 1 — 12) completely surround
the carapace with the exception of the small space occupied by
the nuchal.

The scutes of the plastron (Fig. 2) are twelve in number, con-
sisting of six pairs of large flat elements named from anterior to
posterior end as follows : gulars (6".), numerals (H.}, pectorals
(P.}, abdominals (A.}, femorals (F}, and anals (AN}.

At the angles made by the junction of the pectorals and ab-
dominals with the marginals are two pairs of small triangular
scutes called respectively axillaries (A^.) and inguinals (/.). These
constitute all that remains in the Emydidae of the inframarginals,
a row much more prominent and complete in more primitive
families.

No other plates or scutes occur normally among the Emydidae,
but for the sake of completing the nomenclature, it should be
mentioned that one species, Macrochelys teunnincki, possesses an
additional pair of rows of scutes between costals and marginals,

ABNORMALITIES IN CHELONIA. /I

called SHpraiiMrginals. Traces of a median ventral row of scutes
are found normally in some species - - and I have given the name
" inter pldstral* to this row. A single median scute occurs
normally in the anterior part of the plastron of certain special-
ized groups and receives the name intergular.

II. DISCUSSIONS OF THE PRESENT STATUS OF THE QUESTION
CONCERNING THE MORPHOLOGY OF THE CHELONIAN ARMOR.

The frequent abnormal occurrence of traces of the inframargi-
nals and interplastrals in Grapteinys gcograpliica, Chryscinys mar-
gi/iataznd Chclydra scrpcntinaled me to review the literature re-
lating to the evolutionary history of the chelonian carapace and
plastron.

For nearly a century the chelonian armor has offered to mor-
phologists a problem of unusual difficulty, and, although much
has been written on the subject, its derivation is still unsettled.
The question has been attacked from the three standpoints of
paleontology, embryology and comparative anatomy.

I. Paleontological data are far from conclusive. It is not pos-
sible to go into this phase of the subject at all fully. Baur in
1887 published a brief summary of the more valuable paleonto-
logical data in an article entitled " On the Morphogeny of the
Carapace of the Testudinata." A brief statement of the substance
of this paper will, perhaps, serve to show the inadequacy of the
paleontological evidence in this case.

The condition seen in the Dermochelydae is considered to be
the most primitive. Fossil remains of this gro^ip agree closely
with the existing Dermochelys coriacca in the possession of " a
pavement of small osseous plates extending over the whole shield,
jointed to one another by more or less fine sutures. The num-
ber of these plates is much larger than that of the other Testudi-
nata, which is never more than 70." This pavement of osseous
plates is not united with the internal skeleton, as are the plates
of other Testudinata, but has an independent dermal origin.
" That the carapace of the Dermochelydae is homologous to the
carapace, without internal skeleton, of the rest of the Testudinata,
there is no doubt." The fusion of the dermal pavement bones
with the ribs and vertebras is, according to Baur, proved by a

72 H. H. NEWMAN.

specimen of Eretnwclielys imbricata, a fossil species in which are
found " small polygonal plates of the same shape as those of
Dermochelys, suturally connected with the third, fourth, fifth and
sixth costal plates." "A form between the Dermochelydae and
" Thecop/iora" (Dollo) is represented by the oldest known turtle,
Psephodenna alpinuin, H. v. Meyer, from the Triassic of the
Bavarian mountains, preserved in Munich. In this highly in-
teresting specimen, never mentioned in monographs on the
Testudinata, we have certainly not less than 193 plates suturally
united." According to Zittel's Paleontologie, Baur later ex-
pressed the opinion that Psephodenna may not be a chelonian
at all, but perhaps a nothosaurus. Thus doubt is cast upon
the best link in the chain of evidence. That all the principal
groups of Chelonia were in existence in the earlier Mesozoic ages
and that Palaeozoic Chelonia are entirely unknown are familiar
facts. So our attempts to reconstruct an ancestral condition
must be made largely on the basis of embryology and compara-
tive anatomy.

2. Nor is cinbryological evidence of chelonian phylogeny at
all conclusive. The best and most recent study of the develop-
mental history of the chelonian carapace and plastron was made
by Goette in 1899. He summarizes the previous literature on
the subject and shows that the main question at issue is that of
the character of the neural and costal plates. Some authors,
principally paleontologists, have maintained that these structures
have a dermal origin and hence arise independently of the inter-
nal skeleton. Others hold that these plates are mere outgrowths
of the ribs and spinal processes of the vertebras. Goette favors
the latter view and presents as evidence of its correctness a series
of very careful embryological studies.

Suspecting that there might be some flaw in Goette's work, I
repeated much of it, using the embryos of Chelydra serpcntina
and Grapteinys geograpJiica, and have satisfied myself that the
neural and costal plates actually do originate as outgrowths of a
differentiated tissue that surrounds the neural and rib cartilages.
Whether this differentiated tissue be true periosteum, as Goette
affirms, or simply a somewhat denser portion of the connective
tissue that fills the space between the epidermis and the cartilagi-

ABNORMALITIES IN CHELONIA. 73

nous skeleton, is not certain. Haycraft ('99) maintains the lat-
ter view, but his paper is far from convincing.

As to the remaining plates of the carapace — nuchal, pro-
caudals, pygal and marginals — there is no difference of opinion.
All agree that they are of true dermal origin.

Thus it would seem that the plates of the carapace have a dual
origin — the neurals and costals being periosteal ossifications
while the nuchal, procaudals, pygal and marginals are dermal
ossifications.

The carapace, then, as it exists to-day is not a simple struc-
ture but consists of a complex of at least two independent systems
of bones.

Accepting the evidence of embryology as to the origin of the
neural and costal plates, it remains to determine whether the
dermal ossifications are, as Goette believes, mere supplementary
structures that have come in to supply the deficiencies of the
periosteal system, or are remnants of a once more or less complete
dermal carapace that has in large measure been rendered super-
fluous by the broadening-out of the ribs and neural processes.
The latter view would involve the former existence of complete
rows of dermal bones overlying the vertebrae and ribs. Embry-
ological evidence seems contrary to this view, as no dermal ossi-
fications are found in the costal or mid-neural regions. It is
possible that we may in this case overestimate the evidence of
embryology as a guide to phytogeny. The great antiquity of
the chelonian carapace is undoubted and in highly specialized
structures that have attained a marked morphological fixity we
should not be surprised to find great condensation in develop-
ment, so that two structures formerly independent in origin —
such as dermal and periosteal plates — may originate simulta-
neously so as to form only one inseparable structure. It seems
quite plausible, then, that the rapid secondary broadening of
ribs and neural processes has crowded out or appropriated the
primordia that formerly went to form the dermal carapace and
that only in places where the ribs and neural processes fail
to reach the dermis do the true dermal bones have a chance
to appear.

1 he fact that the nuchal plate appears before the ribs and

74 H. H. NEWMAN.

neural spines have even commenced to broaden out and that the
procaudals and the marginals follow before the neurals and costals
are completely organized, points to the antiquity of these dermal
structures and indicates that the neurals and costals are of more
recent origin.

3. Comparative anatomy furnishes us much valuable evidence.
In the family Trionychidae, for example, we have a series of
forms that show a gradual reduction of a portion of the dermal
armor. Fossil Trionychidae are well known in which are shown
a nuchal, a procaudal and a nearly complete set of marginal
plates. Such a form was figured by Dollo in 1884 and named
by him Pseudotrionyx. Another fossil species discovered by the
same palaeontologist and named by him Emyda granosa, lacks
the procaudal and the marginals from the anterior half of the
carapace. A third form, Emyda Ceylonemis, possesses a nuchal
and several marginals at the posterior part of the carapace. The
extreme limit of reduction is seen in Aspidonectes spiuifcr, which
possesses only the nuchal plate as the last remnant of the dermal
carapace.

It will be noted that the order of the appearance of these
dermal ossifications in ontogeny is just the inverse of the order
of disappearance in phylogeny. The latest elements to be
formed in ontogeny are the first to disappear in phylogeny. This
is just what we would expect if we consider that there has been
a gradual shortening of the developmental process, a gradual
elimination of the latest stages. The Trionychidae show clearly
that there is a marked tendency to reduce the system of dermal
bones and it is not difficult to imagine that earlier reduction has
taken place in which the dermal ossifications of midneural and
costal regions were lost.

What evidence have we that such dermal ossifications over-
lying neural processes and ribs actually existed ? O. P. Hay
('97) in an important paper dealing with the evolution of the
chelonian carapace and plastron, describes and pictures an in-
complete carapace of a fossil form named Toxochelys serrifer.
Three ossicles occur above and overlapping the neural plates and
occupy positions coincident with the keels of the second, third
and fourth neural scutes. These ossicles have the general form

"ABNORMALITIES" IN CHELONIA. 75

of the tubercles seen on the dorsal ridge of the tail of Clielydra
scrpcntina, and this suggests that the ossicles of Toxochelys are
merely a continuation forward of a series of tubercles that must
have been present on the tail.

Hay suggests that the keels seen especially in the young of
modern Chdonia are the representatives of ancient dermal tuber-
cles that formed the chief armor of ancestral forms. That in
most cases these dermal ossicles have ceased to form indepen-
dently of the deeper and more vigorous bony layers is perhaps
to be expected as the result of condensation in developmental
processes.

The degree to which modern species exhibit keels is extremely
varied. Some highly specialized forms show none, or at most
one, even in very young specimens, while one very primitive spe-
cies, Macrochclys tcmmincki, possesses seven distinct keels on the
carapace and four rows of flat scutes on the plastron. This mul-
tiplicity of keels is evidently a very primitive condition and natu-
rally suggests to Hay the condition seen in Dermochelys coriacea
in which twele well-marked keels are found, each keel consist-
ing of rows of dermal ossifications that are lager and more
prominent than the remaining intermediate ossicles that form the
continuous pavement of the test. This peculiar aberrant che-
lonian is taken by Hay, following Baur and others, as the hypo-
thetical ancestral type from which our modern chelonians have
been derived by a process of simplification.

A survey of the field reveals the fact that the nearest approach
to this condition of twelve rows of keels is seen in Macrochelys
tcmmincki, which possesses seven distinct keels on the carapace.
The four rows of flat scutes on the plastron may once have been
keeled, for keels on the plastron are known in both extinct and
living groups. The total number of keels or keel equivalents in
Macrochelys is then eleven, one short of the supposed ancestral
condition. The missing keel is the mid-ventral one and is repre-
sented in certain groups by intergulars. Thus all of the ancestral
keels find representatives among modern species.

Hay seems to have been the first observer to suggest the im-
portance of the scutes as factors in the evolution of the carapace.
Previous authors have confined their attention to the bony struc-

76 H. H. NEWMAN.

tures, considering the scutes as of little significance. Hay's view
of the role of the scutes may be stated briefly as follows : The
probable ancestral condition is that seen in Dermachelys, the skin
of which is found to be broken up into small polygonal areas,
larger in the keels than elsewhere. These areas coincided with
the osteodermal plates that are or will be developed in the skin.
As the deeper elements of the carapace (neural and costal plates)
increased in protective efficiency, the dermal structures were in
many regions rendered superfluous and disappeared. In some
cases the scutes were lost with their corresponding plates, in
others the lost plate left its trace in the keel of the scute. The
direction of growth of each of the existing series of scutes shows
the direction of encroachment on other rows now lost.

This exposition of Hay's seems to me to be the most rational
yet advanced, yet I believe that he fails to appreciate the evidence
of embryology and thus introduces undue complexity. In the
first place, he considers the nuchal plate as a fascia bone instead
of an ordinary dermal plate. In the second place he states that
the neural and costal plates are of the same character as the
nuchal. Embryology shows that the nuchal plate is as true a
dermal bone as are the marginals, while the neurals and costals
are true periosteal expansions. It seems to me more rational to
suppose that the dermal ossifications of the mid-neural and costal
regions have undergone a complete suppression identical with
that indicated by the series of Trionychidae described above, rather
than that they have become indistinguishable by fusion with the
rib and neural fascia bones, as Hay calls them.

If we remove the scutes and underlying dermis from the cara-
pace of a specimen of CJielydra we find that the long tubercles on
the neural and costal plates bear no constant relation to the
plates themselves, but are nevertheless clearly of a piece with
them. It was natural for Hay to suppose that these bony
tubercles were produced separately and then fused with the
underlying plates. I have been able to trace this matter to a
conclusion in the young of Chclydra, with the result that I have
seen all the stages of ossification in the carapace and know that
the tubercular keels on the neural and costal plates are produced
by gradual thickenings of the growing plates. These thickenings

"ABNORMALITIES IN CHELONIA. 77

send out branching processes that gradually displace the dermal
connective tissue of the tubercles and fill the space with bone.
Complete ossification of these tubercles does not occur until the
animals are several years of age.

III. DESCRIPTION AND DISCUSSION OF ABNORMALITIES.

That a process of reduction both in the number of rows of
scutes and in the number of scutes in surviving rows has taken
place seems highly probable. From this standpoint I made a
systematic study of all the abnormal specimens that showed traces
of these lost rows or lost scutes. Inframarginals of all grades of
prominence were found in specimens of Graptcmys geograpkica
and Chrysemys marginata, while interplastrals were found more
rarely in the same two species. It will be noted that both of
these recurring rows are plastron rows which probably means
that the carapace has reached a high degree of fixity with refer-
ence to number of rows. Yet many abnormalities are found that
indicate that the reduction in the number of scutes in a row was
of comparatively recent occurrence.

These abnormalities will be discussed under three heads : (i)
Inframarginals, (2) interplastrals, (3) supernumerary scutes in a

row.

i . Inframarginals.

The occurrence or non-occurrence of inframarginals has formed
the basis for separating the Thecophora into two great groups.
Gadow in his volume on Amphibia and Reptiles gives Boulanger's
key for classifying Chelonia. In this the two groups are charac-
terized as follows :

1. Pectoral shields separated from the marginals by inframar-
ginals— Chelydridae, Platysternidae, Cinosternidae.

2. Pectoral shields in contact with the marginals — Testudini-
dae, Chelydidas, Pelomedusidae.

It is evident that the more primitive families possess as normal
factors this row of scutes while the more specialized families
normally lack this row. When, however, dozens of specimens
of Graptcmys and Chrysemys possess this row in more or less
perfect form, I am forced to consider this phenomenon as a well-
marked case of systematic atavism. In view of the fact that no

78 H. H. NEWMAN.

such anomalies have been previously described, it seems worth
while to tabulate those in my collection - - an easy task in view
of the fact that the scutes occur in definite places. In any species,
such as Chelydra, that possesses this row normally, there are
typically three scutes in the row, one in contact with the axillary,
one at the angle of contact of the pectoral and humeral and the
marginals, and one abutting on the inguinal scute. These three
scutes may be designated respectively as I., II. and III. Out of
476 specimens of Grapteuiys geographica examined, I found 3 1
with traces of inframarginals varying all the way from three large
scutes on each side to one small one on one side. The tabula-
tion below gives the number of the specimen, the sex, the length
and breadth of carapace in millimeters, the occurrences of infra-
marginals on the right and left sides separately. Three general
sizes are distinguished, which although quite arbitrarily laid down
may serve to give a more definite idea of the amount of variation
that occurs. These sizes are designated as large, medium and
small.

At Woods Hole this summer I found two specimens of Nan-
ncinys gnttata and three specimens of Chrysemys picta with well
marked inframarginals.

It will be readily seen that in both species the middle scute is
much the commonest recurrence, and this is natural if we consider
that in Chelydra, and other species with well developed infra-
marginal rows, the middle scute is always the largest. The
largest and most vigorous scute would probably persist longer
and hence be most likely to recur as an atavistic reminiscence.
The fact that no. III. is next in prevalence in Grapteuiys and
no. I. in Chrysemyjs, indicates that the order of suppression of the
other scutes of the row was subject to individual and group
variation.

In the species Chelydra serpentina the inframarginal row is in
a highly variable condition. Many stages in the reduction of
numbers of scutes are to be seen in different individuals. The
middle scute, corresponding to no. II., is always the largest, and
the adjoining ones are next in size and would correspond to
no. I. and no. III. Frequently there are two or three smaller
scutes both in front of and behind the large central scutes, but

ABNORMALITIES IN CHELONIA.

79

GRAPTEMYS GEOGRAPHICA.

No.

Sex.

Length in Breadth in
mm. mm.

Right Side.

Left Side.

I

F

98

81

Small II.

Medium II.

2

F

189

139

Small II.

3

F 200

171 Small II.

Large III.

4

M 109

82

Medium I.

5

F 75

63

Large II.

Medium II.

6

63 57

Medium III. Medium III.

7

M

82 70

Small II.

8

?

63

56

Large II.

9

F

75

63

Small III.

10

?

60

54

Small II.

Medium II.

ii

?

57

Large II.

Large II.

Large III.

Large III.

12

F

195

147 Large III.

Large 11.

Large III.

13

F

1 60

118

Medium III.

Large III.

14

F

99

82

Medium I.

Medium I.

Medium II.

Medium II.

Medium III.

Medium III.

'5

F

94

82

Large II.

Large II.

Large III.

16

F

134

112

Large I.

Large I.

Large II.

Large II.

17

F

101

76

Medium I.

Large 11. ( See

Fig. 44. )

Large II.

Large III.

Large III.

18

M 88 69 Medium I.

19

M 95 73 Large II.

Large II.

Large III.

20

?

56

51 Small II.

Small II.

Small III.

21 ?

55

50 Medium II.

Medium II.

22

?

57

53

Large II.

Large II.

23

?

62

54

Large II.

Large II.

Large III.

24

F

So

65

Medium II.

25

M

60

50 Large II.

Medium I.

Medium II.

26 ?

51

48 Medium II.

Medium II.

Medium III.

27

?

58

53 Small III.

28

M

66

54 Medium II.

Medium I.

29

• M

1 10 78 Medium III.

30 M 97 77

Small II.

they shown signs of suppression and in the majority of specimens
are of insignificant size. The axillary and inguinal scutes of the
Emydidae, etc., correspond, I believe, to two of these smaller
scutes that are undergoing suppression in Chelydra. They have
persisted in the Emydidae probably because they were needed to
fill in the angles between the plastrals and marginals. Aroma-

So

H. H. NEWMAN.

CHRYSEMYS MARGINATA (188 specimens examined).

No

Sex.

Length in
mm.

Breadth in
mm.

Right Side.

Left Side.

I

F

IO4

80

Medium II.

Medium III.

2

F

116

88

Small II.

3

F

IO2

81

Large I.

Large II.

Large II.

4

M

85

67

Medium I.

Small III.

Medium II.

5

F

90

73

Small II.

Small II.

6

M

96

72

Large I.

7

F

98

80

Large I.

8

?

55

50

Medium I.

Large 11.

Medium II.

9

M

72

58

Small II.

Small II.

10

F

"3

86

Large II.

Large II. (See
Fig. 43 )

A tabulation of the above results shows :

Right Side.
I. 4 scutes

II. 20 "
III. 13 «

Total : 37 "

Right Side.
I. 4 scutes
II. 8 "
III. I "

Total : 13 "

Grapheniys •

Ch rysemys

Left Side
5 scutes

20 "
10 "

35 "

Left Side.
2 scutes

4 "
I "

chclys odorata (Fig. 53) shows a curious survival of inframarginals,
having invariably only two scutes, one large and the other very
small and vestigial. From its position, separating the pectoral
shields from the marginals, I would homologize this large scute
with no. II. and the vestigial scute with no. I. Complete sup-
pression of the inframarginal row has occurred in the terrestrial
genera of the Emydida?.

2. Interplastrals.

The occurrence of traces of the interplastral row are not
nearly so frequently found as those of the inframarginals. Yet
they are sufficiently numerous and definite to note in this con-
nection. Traces have been found in Chelydra, Grapteuiys and
Cliryseinys. In preparing a list of these occurrences it will be
convenient to number the places where such scutes might occur,

"ABNORMALITIES" IN CHELONIA. 81

A,B, C, D and E, beginning at the anterior end. Two specimens
of Chrysemys marginata have extra scutes at A (Fig. 47). Two
specimens of Chclydra have extra scutes at C (Fig. 45). One
specimen of Graptcuiys has extra scute at D (Fig. 48). One
specimen of Graptcmys has a pair of extra scutes at E (Fig. 46).
The primitive condition was probably one in which a scute was
present at each point of union of four plastron scutes, but the
fact that even in the tail of Chelydra this row is either partially
or wholly wanting indicates the rather uncertain character of
the row. In the specimens listed above scutes are found occur-
ring in four places out of a possible five. No doubt a larger col-
lection would serve to fill in this gap.

I consider these recurrences as true reversions to ancestral
conditions ; and that they come under the head of systematic
atavism I see no reason to doubt.

How the typical number of scute rows seen in our modern
tortoises has been acquired has, perhaps, been sufficiently dis-
cussed and it now seems necessary to consider the processes
that have brought about the reduction of the number of scutes
in a row — for it is beyond dispute that such a reduction has
taken place.

3. Supernumerary Scutes in a Row on the Carapace.

The literature on this subject is limited to one paper, Gadow's
much-discussed " Orthogenetic Variation in the Scutes of Che-
Ionia," that was published in Willey's Zoological Results in
1899. The author gives a very interesting account of the con-
ditions found in the common loggerhead turtle, Tlialassoclielys
carctta. He has gathered together a miscellaneous assortment
of some sixty-nine specimens of various sizes, principally new-
born, from many parts of the world. On the basis of this col-
lection he comes to the conclusion that scute reduction proceeds
along certain definite lines. His observations, however, are
limited to reductions in the neural and costal rows. According

t>

to Gadow, the ideal ancestral condition is one in which the neural
and costal bony plates determine the number of scutes. The
author's idea is that there was originally a scute for each of these
plates.

82 H. H. NEWMAN.

Starting with this ideal condition as stage I., he finds the
nearest approach to it in specimen I, that has 8 left costals of
which 2 are vestigial, 8 right costals of which I is vestigial, and
8 neurals. The greatest reduction is that seen in specimen 26,
which has 5 left costals of which I is vestigial, 4 right costals,
and 7 neurals of which I is vestigial. This latter specimen is
reduced below the normal for the species, which is arbitrarily
said to possess 6 neurals and 5 pairs of costals. This condition
is said to be the goal toward which every young Thalassochelys
carctta is striving.

The following stages are mapped out in diagram, following
Gadow, to show the sequence in scute reduction in the chelonian
carapace :

Stage I. - - Hypothetical, eight neurals and eight pairs of
costals. Neurals and costals lie in the same transverse plane
and coincide with neurals and costal plates.

Stage II. - - Eight neurals and eight pairs of costals, the latter
fitting with their inner angles dovetailed between two successive
neurals. Rearrangement probably brought about by the partial
reduction of one pair of costal scutes. This reduced pair is
probably the second.

Stage III. - - Eight neurals and seven pairs of costals, the
original second costals suppressed, original third becoming
second, etc.

Stage TV. — Seven neurals and seven pairs of costals, but fifth
neural and fourth pair of costals (original fifth), in a state of
reduction.

Stage V. — Six neurals and six pairs of costals, owing to
complete suppression of fifth neural and fourth (original fifth)
pair of costals.

Stage VI. --Six neurals and five pairs of costals, brought
about by fusion of last two pairs of costals into one or, perhaps,
by suppression of one pair. This is the normal condition in
Thalassochelys.

Stage VII. — Six neurals and four pairs of costals. Normal
condition in the majority of tortoises to-day, brought about by
suppression of first pair of costals.

Stage VIII. — Six neurals and four pairs of costals, first neural
(nuchal) greatly reduced.

"ABNORMALITIES" IN CHELONIA. 83

Stage IX. — Five neurals and four pairs of costals, first neural
(nuchal) suppressed as seen in pleuroderous tortoises.

Beyond this last stage chelonians have not ventured yet, at
least normally.

The order of loss in scutes is according to Gadow : (i) No. 2
costals, (2) no. 7 neural, (3) no. 5 neural and no. 4 (original
no. 5) costals, (4) no. 7 or 8 costals (by fusion or suppression),
(5) no. i costals, (6) no. i neural.

Gadow's paper, while most suggestive, must be criticised in
several particulars, but before proceeding to the criticism it will
be necessary for me to produce the data that to a large extent
form the basis of the criticism. The data are derived from a col-
lection of a large number of abnormal specimens, principally of
two species, Graptemys geographica and Chrysemys marginata.
Gadow worked on a species that is normally abnormal — if such
an expression be permissible. He selected the commonest con-
dition and arbitrarily called it normal. As a matter of fact, there
is no normal or fixed condition. The species TJialassochelys
cwetta is evidently in a highly variable state as to scute number
and arrangement, and no stability has as yet been attained. The
species I have studied have, on the contrary, reached an advanced
state of stability. Yet a sufficiently large number of abnormali-
ties occur to give one nearly as many examples as Gadow had.
Out of 476 specimens of Graptemys, varying from embryos to
adults and taken at random, there occurred 48 specimens with
supernumerary carapace scutes, while iSS CJirysemys yielded 8
such abnormal specimens. Four other species belonging to
widely diverse groups yielded one abnormality apiece. It seems
probable that abnormalities of exactly the kind that I have
found so plentifully in the case of Graptemys and Cliryscmys are
to be found in any species if enough specimens be examined.

In order to economize space in the tabulation of these abnor-
malities brevity in the nomenclature of these vestigial scutes must
be attained by numbering them. Combining Gadow's figures
with my own results, I have good reason to believe that vestigial
scutes occur between every two surviving normal scutes and that
the first, second and last costals are also found in a vestigial con-
dition. On this basis, then, there were eleven neurals and ten

84 H. H. NEWMAN.

pairs of costals. These, if numbered from anterior to posterior,
would give the numbers I, 3, 5, 7, 9 and 1 1, to surviving neurals,
and numbers 2, 4, 6, 8 and 10 to vestigial or lost neurals ; the
numbers 3, 5, 7 and 9 to surviving costals, and numbers i, 2, 4,
6, 8 and 10 to vestigial or lost costals. In the tabulation these
numbers will be used without further explanation. Furthermore,
the sex, length and breadth of carapace, brief descriptions of both
scutes and bony plates, will be given in separate columns. The
significance of the tabulation of conditions of bony plates will be
seen later when the subject of correlation between scute and plate
abnormalities is discussed. The specimens are numbered and
arranged in the order of abnormality, the specimens with largest
number of extra scutes coming first, and those with less than the
normal number of scutes last. Extra neurals will be listed be-
fore extra costals and the latter before extra marginals.

Two kinds of abnormality may be distinguished : symmetrical
and asymmetrical. The former are less common and are impor-
tant in that they furnish clearer cases and thus throw light on
the latter. Under the head of symmetrical abnormalities may be
mentioned extra neurals in the median line or nearly so ; extra
costals in pairs symmetrically placed ; extra paired marginals.
The great majority of abnormalities are asymmetrical, consisting
of: extra neurals crowded to one side or the other but usually
showing clearly enough the position they would normally oc-
cupy ; unpaired costals or marginals. In the case of asymmet-
rical neurals it is sometimes difficult to distinguish the supernu-
merary scute from the normal scute, on account of the large size
of the former and the fact that crowding has forced the two
scutes to lie approximately side by side. There are usually cor-
related points of asymmetry that may be of assistance in deciding
the point, but occasionally I have been compelled to trust to my
judgment and may possibly have erred. Gadow would probably
consider the type in which the normal and supernumerary scutes
lie side by side as evidence of the original paired character of the
neural row. Were it not for transitional conditions this view
might be tenable.

Occasionally it becomes difficult to determine which of five
costals is the supernumerary scute, but a reference to the mar-

ABNORMALITIES

IN CHELON1A.

GRAPTEMYS GEOGRAPHICA.

No.

Fig.

Sex.

Length

Breadth

Scute Abnormalities

Plate Abnormalities.

in mm.

in mm.

I

2O

?

60

54

Complete transverse

Plates not fully

row of 6 large scutes in formed.

middle region. Difi&cult

to diagnose. Double 6

neural (large). L 6

costal (large). R 4 cos-

tal (large). Paired

marginals.

2

6

F

188

144

Median 2 neural

No. I procaudal

(large). Paired i cos-

fused with 8 neu-

tals (large). Extra R

ral. Extra R mar-

marginal (large).

ginal.

3

35

?

44

42

R 8 and 10 neurals

Plates not fully

(large). R 10 costal

formed.

( medium ) . L 10 costal

(small).

4

4

F

122

97

3 neural partially di-

R Marginal.

vided. Probably indi-

cates fusion of 3 and 4

neural. Paired I costals

(large). R marginal

,

(large).

5

5

F

2OO

H5

R 8 and 10 neurals

9 neural. Dou-

(large). R 8 costals

ble extra procau-

(large).

dal. R costal

( medium ). L

costal (small).

6

3

F

189

139

R 10 neural (large).

9 neural. Dou-

R to costal (large). R

ble extra procau-

marginal (medium).

dal. Paired 9 cos-

tals (large). R

marginal.

7

13

M

9S

7i

R 8 and 10 neurals

Normal.

(large). R 8 costal

(large).

8

H '

M

84

70

L 10 neural (large).

Normal.

9 neural, partly divided,

and probably represents

8 and 9 neurals fused.

L 10 costal (medium).

9 9

F

98

81

L 8 neural (large).

Normal.

10 and 1 1 neurals fused.

L costal (medium).

10

36

p

52

47

R 8 and 10 neurals

Bones not

(large). Lacks a R

formed.

marginal.

n

34

(embryo)

21

18

L 8 and -10 neurals

B o.n e s not

(large). formed.

12

37

a

25

21

R 10 neural (large). Bones not

R 10 costal (medium), formed.

13

8

Y

174

136

R 10 neural (large). Extra procaudal.

R ro costal (large).

14

10

M

U3

76

L 10 neural (medium ). \ Normal.

L 10 costal (medium).

15 12 M

98

70

R lOneuralf .nedium). 9 neural. Paired

R 10 costal (medium). 9 costals.

86

II. II. NEWMAN.

\ITEMYS GEOGKAI'HICA. — Coutillllt'i/.

No.

Fig.

Sex.

Length

Breadth

Scute Abnormalities.

Plate Abnormalities.

in mm.

in mm.

16

40

?

57

51

L 10 neural ( medium ) .

Bones not

I, 10 costal (medium).

formed.

17

Same as

M

87

73

L I o neural ( medium ) .

Normal.

10

L 10 costal (medium).

1 8

7

F

170

138

6 and 7 neurals par-

Only p b o t o-

tially fused. 9 and II

graphic record re-

neurals completely

tained.

fused. L 6 costal (me-

dium).

*9

56

(embryo)

12

10

R 8 neural (large).

Rirs very ab-

R marginal.

normal.

20

17

M

89

69

Paired lo costals

Normal.

(small).

21

See 17

F

160 118

Paired IO costals (me-

Normal.

dium).

22

See 17

?

61

54

Paired IO costals (me-

Bones not

dium).

formed.

23

See 17

F

209

169

Paired lo costals (me- Normal.

dium).

24

26

?

5i

45

Paired I costal (me- Hones not

dium). formed.

25

38 (embryo)

23 19 Median 8 neural Hones not

(large). formed.

26

39 (embryo)

19 16

L 8 neural (large).

Bones not

formed.

27

23

F

192

154

Median IO neural Normal.

(medium).

28

15

M

83

67

L I o neural (medium). Normal.

29

it

M

109

82

L 6 neural (large). Normal.

30

22

M

75

63

Paired extra marginals Paired extra

(small). marginals (small).

31

1 6

M

78

67

R 10 costal (large).

Normal.

32

See 1 6

M

82

70

R 10 costal (large).

Normal.

33

See 1 6

F

72

62

R to costal (medium).

Bones not

formed.

34

See 1 6

?

52

47

R IO costal (medium).

P> i) n e s not

formed.

35

See 1 6

i

63

56

R IO costal (medium ).

Bones not

formed.

36

21

F

66

57

L IO costal (medium).

1! o n e s not

formed.

37

See 21

?

57

5i

L lo costal (medium).

Bones not

1 formed.

38

See 21

?

29

26

L 10 costal (medium).1 Bones not

formed.

39

See 21

?

60

54

L I o costal (medium).

Bones not

formed.

40

See 21

F

73 63

L IO costal (medium).

Bones not

formed.

4i

See 21

F

99 82 L TO costal (small).

Normal

42

See 27

p

47 42 L marginal lacking.

Bones not

formed.

43

27

?

58 52 L marginal lacking. Hones not

formed.

44

19

F

72

64

R marginal lacking.

R marginal

lacking.

ABNORMALITIES IN CHFIONIA.

i ;K \rrrM\s CKUIKAI-IIICX.—

No.

Fig.

Sex.

Length

in nun.

Krendth

in nun.

Seine Abnormalities.

Plate Almormnlities

4S

See 19

?

58

53

R marginal lacking.

1 '. 11 n e s not

formed.

46

18 F

77 65

Paired marginals luck-

Paired marginals

ing.

lacking.

47

( embryo)

1'aircil marginals lack-
ing and costals not fully

Hones not
formed.

1

differentiated.

CllKVSK.MVS MARC1NATA.

No

Fig.

Sex.

Length

in nun.

Breadth
in mm.

Scute Abnormalities.

Plate Abnormalities

4S

31

F

130

93

R 8 and IO costals

Kxtra 1'iocau-

(large).

dals.

R 10 costal (large).

49

32

F

73

61 I, 8 and 10 neurals

Normal.

(large).

50

30 M

104 So Paired IO costals

Normal.

(large).

51

25 52 47 Paired 4 costals (me-

Bones not fully

dium ).

formed.

52

24 57 50 L 4 or R 6 neural.

Hones not fully

formed.

53

20 F 84 67

R 10 neural (large ).

Normal.

54

28 M

97

(,o

L I costal (medium).

Normal,

55

41 F

73

63

L 8 costal (medium).

Normal.

ginals will usually settle the point, as the normal condition has a
very characteristic arrangement of these t\vo sets of elements.

Another source of difficulty arises from the complete or in-
complete fusion of adjacent scutes. Fusion is due to the inhibi-
tion of the process of division into epidermal areas at a rather
late embryonic stage. In some cases the fused scutes show their
separate identity, after a year or two of growth, by a separation
of their growth rings. Coker has called attention to several such
cases in connection with Malaclcunnys ccntrata and I have
observed the same phenomenon in the marginals of Graptcinys
on several occasions. Usually, however, the indications are clear
enough to enable one to recognize the individual elements in a
fused scute. It seems reasonable in the present discussion to
consider the number of scute primordia involved in a fusion and
to give them the full rank of independent scutes.

As in the previous tabulation, the arbitrary terms, large-, small
and medium, are used. Lor R in connection with neurals will

88 H. H. NEWMAN.

indicate that the extra scute is crowded to left or right. The
same letters indicate the side on which extra costals and mar-
ginals occur.

The following isolated abnormalities have come to hand and
may be listed :

Large specimen of Terrapene Carolina has a R 10 costal
(medium).

Medium-sized shell of Cyclcmys doitata has paired 6 costals
(medium). See Fig. 42.

Two medium-sized specimens of Chelydra scrpcntina have L
10 costal (large).

Large AronwcJielys odorata has R 10 costal (medium).

A reference to the literature enables me to list a considerable
number of similiar abnormalities. The names used in the refer-
ences will be retained.

1. Ptychemys clcgans, Agassiz, L., Contributions to the Nat-
ural History of the U. S., Plate I, Fig. 13, showing: L 4 neu-
ral, paired 4 costals, paired extra marginals. (Figured in this
paper as Fig. 33.)

2. Chelopas insailptns, Parker, G. H., paired marginals lacking.

3. Same : R 8 and 10 neurals, R 10 costal, R marginals lacking.

IN CATALOG OF SHIELD REPTILES IN THE BRITISH MUSEUM.

4. Emysvermiculata^zb. XIII.), 24 neural, R 8 and 10 neural.

5. Emys singuinulenta (Tab. XV.), R 6 neural, R 6 costal, L
4, 6 and 10 costals.

6. .Cyclcmys dentata (Tab. XIX.), median 4 neural, R 6 neural,
L 8 neural.

7. Chclodina oblonga (Tab. XXIV.), R 8 neural.

IN HISTORIA TESTUDINUM, SCHOEPFF, J. B., 1792.

8. Testudo cincra (Tab. III.), Fig. 2, paired I costals.

9. Tcstndo arcolata (Thunberg), Tab. XXIII., median 8 neural,
L 6 costal.

10. Tcstndo planiccps, XXVII. , L 4 costal.

Discussion.

A scute, whether normal or supernumerary, is a separate and
definite entity, resulting from a definite embryonic primordium.

ABNORMALITIES IN CHELON1A.

89

The fact that supernumerary scutes have been found between all
of the normal scutes as well as at both ends of the costal series
must have some significance. If one assumes that these super-
numerary scutes represent the atavistic recurrence of scutes that
have been lost in the course of phylogeny, it is possible that the
following tabulation will throw some light on the sequence of
loss.

Neural 's.

Costah.

No. of Scute.

Numbers of Recurrences.

No of Scute.

Numbers of Recurrences

2

I

I

6

4
6
8

10

2 (one doubtful)
5 (one doubtful )
II (one doubtful)
1 6 (one doubtful)

2

4
6
8

o

3
4 (one doubtful)
4 (one doubtful)

10

35

It will be seen that the most frequent recurrences are at the
posterior end of the carapace, and that, with the exception of the
first costal, the frequency of recurrence diminishes as we proceed
anteriorly. What significance attaches to this fact ? It seems
quite probable that the most frequent recurrences represent the
most recent losses and the rarest recurrences the most ancient
losses. This rule held good for the suppression of rows of
scutes and should apply here also.

On this basis then we can at least say that the succession of
suppression was in general antero-posterior, that the earliest
losses occurred at the anterior end of the carapace and the most
recent losses at the posterior end. One might go further and say
that in the neural series the order of suppression was probably
2, 4, 6, 8 and 10. The antero-posterior order of loss is not so
clear in the case of the costals, as no. I costal recurs more fre-
quently than any other except no. 10. This means a modifica-
tion of the regular mode of progression. In the costal series it
is probable that the antero-posterior succession of losses was
interfered with by the rounding-in of the marginals both ante-
riorly and posteriorly. This rounding-in would necessarily
begin about medially and proceed in two directions, hence the
second supernumerary scute would be put under pressure before
the first and the eighth before the tenth. The antero-posterior

H. H. NEWMAN.

tendency, however, would bring about the suppression of anterior
scutes, as a whole, before posterior scutes.

Evidences are not wanting that scutes may be suppressed and
the method of suppression seems clear. In a specimen of Cycle-
mys dentata, listed as no. 58 and figured on Plate III., Fig. 42,
the paired sixth costals are being encroached upon by the seventh
costals. The anterior growing margins of the latter have pushed
in under the posterior edges of the former in such a way as to
severely cut into their growth centers. The dotted line shows
the amount of encroachment. Several specimens of Graptemys
show the same phenomena, and the scutes encroached upon are
always the supernumerary ones. This may be looked upon as a
recurrence of an ancestral condition and we may infer that the
loss of certain scutes has been brought about through the
encroachment, more and more severe with succeeding genera-
tions, of more vigorous upon less vigorous scutes, resulting in
the final complete suppression of the latter. We must also
suppose that the rudiments of the lost scutes lie dormant in the
embryonic tissues and occasionally for some reason reappear
more or less completely. Those that have been suppressed for
the longest time would naturally reappear least often and vice
versa. On this basis, then, we may safely say that the order of
loss is orthogenetic if by this we simply mean onward develop-
ment.

Applying the same methods to Gadow's figures I find a very
general agreement, although 1 am unable to agree with the
author's interpretations. The vestigial scutes that occur in
Gadow's figures are : neurals 2, 8 and 10; costals I, 2, 4, 6, 8
and 10. No. 2 costal was not found in my specimens, but is so
clearly seen in Gadow's Fig. I that I have introduced it into my
system. It is possible that No. 2 costal was the most ancient
loss and hardly likely to recur in specialized types such as
Graptemys and C/iiysernys, since it occurred only once in Gadow's
specimens of TJialassoclielys. It will be seen that Gadow finds
no vestiges of neurals 4 and 6. An examination of his figures
will show that TJialassochclys has attained a high degree of fixity
in the anterior portions of the mid-neural series, while all other
regions are still in a decidedly variable condition. Hence we are

"ABNORMALITIES IN CHELONIA. 91

unlikely to find vestigial scutes in this region unless a much
larger number of specimens is examined.

Carapace abnormalities have been pictured by authors for over
a century and I have on my lists fourteen species, belonging to
widely diverse groups, that show the same general abnormalities.

These scattering cases could scarcely be used in determining
the order of loss of scutes, but are of importance in that they
show that certain abnormalities that are comparatively rare in
Graptemys and CJiryseitiys occur with a fair degree of frequency
in other forms. For example: neurals 4 and 6, and costals i,
4 and 6 occur from 2 to 4 times in these specimens. The prev-
alence of abnormalities of this sort over such a wide range of
forms strengthens my idea of the universality of the process of
scute reduction in Chclonia. I have no doubt that such ab-
normalities will be found in any species if enough forms are
examined.

In Gadow's diagrams illustrating the progressive reduction of
epidermal scutes (p. 217) it will be seen that the order of reduc-
tion differs from the one I have proposed in two points ; in the
first place he indicates that no. 10 neural is suppressed before
no. 8, but this is not borne out by his own figures. Figs. 4, 6,
7, 14, 20, 26, show no. 10 persisting after the total suppression
of no. 8, Fig. 26 being especially convincing. Figs. 8, 9, 10,
on the other hand, show no. 8 persisting after the suppression
of no. 10. The balance is decidedly in favor of the earlier sup-
pression of no. 8, yet there must have been some individual vari-
ation in this matter. My own figures show that no. 8 recurs
twelve times as compared with seventeen times for no. 10. In
my own specimens there are eight cases in which nos. 8 and 10
neurals recur together, nine cases of no. 10 recurring alone,
and only four of no. 8 recurring alone. It would seem then
that these two scutes were undergoing a process of suppression
at about the same time, but that no. 8 was in most cases the first
to disappear.

In the second place it seems clear that no. I costal persisted
longer than no. 10 in Thalassochelys, but that the opposite was
the case in all the forms in my collection can scarcely be doubted,
no. 10 recurs thirty-six times and in many species, while no. i
recurs only six times.

92 H. H. NEWMAN.

Some rather remarkable conclusions are expressed in Gadow's
paper and should be discussed in this place.

1. He makes the following statement : "Abnormalities are 4 to
7 times as common in new-born as in mature specimens, hence
scute reduction must take place during the lifetime of the indi-
vidual." I have not had the opportunity of putting this matter
to a test in the case of TJialassoclielys, but the examination of
several complete nests of Graptemys has brought to light the
following facts. Two nests containing respectively thirteen and
fourteen embryos showed no abnormalities. One nest contain-
ing fourteen just-hatched young showed one slight abnormality,
a vestigial no. 10 costal. A fourth nest in which twelve eggs
came to maturity contained five decidedly abnormal specimens,
listed as nos. 11, 19, 25, 26 and 48. This means that barely
10 per cent, of the embryos of four broods are abnormal, while
out of 476 specimens of Graptemys 48 were abnormal in the
carapace scutes, a little over 10 per cent. A large proportion
of Gadow's new-born specimens came from one nest, the whole
brood of which was abnormal. The others were taken in small
sets from various collections, and I believe that such specimens
had been preserved because of their abnormalities. A survey of
my tabulations will show that abnormalities are no more common
in one size than in another. Finally, Coker, in a very recent
preliminary paper, delivered before the American Society of Zool-
ogists in Philadelphia, December, 1904, claims that observations
on embryos of TJialassochelys gave no support to the theory of
Gadow.

2. Gadow considers that certain specimens (Figs. 6 and 24)
show evidences that the neural row was originally a double one.
That this was the case seems very unlikely from an examination
of such primitive conditions as are seen in the tail of CJielydra
and in the neural keel of DermocJielys. The appearances seen in
Figs. 6 and 24 may be due to the crowding of linear members of
the row until they come to lie side by side. Indications of an
approximation to this condition are not uncommon in the speci-
mens which I have had to deal with.

"ABNORMALITIES IN CHELONIA. 93

4. Supernumerary Scutes on tlie Plastron.

The plastron has, as a rule, reached a higher degree of fixity in
the matter of numbers and arrangement of scutes than has the
carapace, but that this portion of the chelonian armor has not al-
ways had so fixed a character may be seen in the high state of
variability of Aromochclys odorata, which is almost as marked as
that seen on the carapace of Thalassochelys. In Aromochclys the
number of plastron scutes varies from 14 to 9 and all intermediate
conditions are readily found. Fig. 5 2 shows the largest number of
scutes seen in the specimens of my collection. In this case there
is a well-developed extra pair of scutes between the abdominals and
femorals. Fig. 53 shows the commonest condition in which
there are the usual five pairs of plates and the gulars are par-
tially fused. Fig. 54 shows the extreme of reduction in which
the pectorals have been lost either through crowding or fusion,
and the gulars have fused into a single median element.

As in the case of the carapace, we find in several species that
have attained a high degree of fixity in the plastron, marked
traces of supernumerary scutes. Fig. 49 shows the plastron of
a small specimen of Chelydra that has an extra pair of scutes be-
tween femorals and anals. Fig. 50 shows another specimen of
Chelydra with a vestigial scute on the right side between humeral
and abdominal. Fig. 51 shows a specimen of CJirysemys with a
well-marked supernumerary scute on the left side between ab-
dominal and femoral. In Figs. 49, 50, 51 and 52 we have super-
numerary scutes in four places out of a possible five. As yet I
have been unable to find supernumerary scutes between gulars
and numerals.

Losses seem to have taken place in two ways : by fusion and
by crowding out. Some curious examples of the latter might be
mentioned. In Chelydra the abdominals have been forced to the
sides, but have been retained to bridge the gap between the small
plastron and the margin of the much larger carapace. In other
cases the pectorals have played a similar role. Van Lidth de
Jeude describes a specimen of Tcstudo ephippium (Gthr.), in
which the pectorals have been crowded to the two sides lika the
abdominals of Chelydra. Other specimens of the same species,
according to Rothschild, have the same abnormality to a greater

94 H. H. NEWMAN.

or less degree. The Catalogue of Shield Reptiles in the British
Museum shows a specimen of Mononria fnsca (Tab. III.), in
which the pectorals are crowded to the margin of the plastron
and have become small and triangular. The same volume shows
a specimen of Sternotherus Derbianns (Tab. XXII.), in which the
pectorals seem to have a tendency to be suppressed or crowded
to one side.

On the whole it seems evident that an orderly suppression of
alternate scutes has taken place in the plastron as well as in the
carapace.

5. Correlated Abnormalities in tJie Scutes and Bony Plates.

The next question that comes up for discussion is whether or
not there is any correlation between scutes and bony plates. It
has long been noticed by morphologists that there is a certain
definiteness about the relative positions and sizes of scutes and
plates. This may be described in brief as a definite overlapping of
bony sutures by scutes. In the marginal series (see Fig. i) this
is seen in its simplest form — every bony suture being covered
by a scute. In the neural and costal series one scute as a rule
covers one whole plate and half of two adjoining plates. This
arrangement is modified in the anterior and posterior regions. In
the former the nuchal plate is partially overlapped by six scutes,
viz. : nuchal, first pair of marginals, first pair of costals (normally
involving only small corners of the plate), and the first neural.
The first costal scutes cover the first and half of the second cos-
tal plates as well as the inner edges of first, second, third and
fourth marginal plates. The last neural covers normally parts
of eight plates, viz. : the two procaudals and the anterior margin
of the pygal, about half of the eighth neural and eighth pair of
costals, and the anterior margins of the eleventh marginals.
Only in the middle portions of the carapace is any definiteness
of arrangement seen, yet there is a marked fixity of relations even
in the most specialized regions. Gadow bases his reduction series
upon an arbitrary connection between these structures, according
to which there was originally a scute for each vertebra and rib.
He gives no reason for assuming a vital connection between these
structures, but simply implies one. In an earlier portion of the

"ABNORMALITIES IN CHELONIA. 95

present paper it has been shown that there is no ontogenetic con-
nection between the scutes and plates, the former being laid down
before the latter have begun to form, while the latter appear com-
paratively late in development as mere outgrowths of the ribs
and neural spines.

If, however, there be any essential connection between these
scutes and plates, we would expect to find scute irregularities and
abnormalities associated with plate irregularities and abnormalties
and vice versa.

G. H. Parker ('99) expresses himself at some length on this
point in a paper entitied " Correlated Abnormalities in the Scutes
and Bony Plates of the Sculptured Tortoise." He describes in
detail two abnormal specimens and on this slender basis reaches
some rather general conclusions.

Specimen no. i has extra eighth and tenth neurals and a small
right tenth costal. No plate abnormalities are found in the neu-
rals or costals, but one right marginal plate and a correspond-
ing scute are lacking. Parker designates these conditions as :
(a) Scute abnormalities unassociated with plate abnormalities, ($)
scute abnormalities associated with plate abnormalities.

Specimen no. 2 has normal neural and costal scutes, but lacks
an entire horizontal row of plates consisting of a neural, a pair of
costals and a pair of marginals. The lack of marginal plates is
associated with the lack of a corresponding pair of scutes. These
conditions are designated as : (rt) Plate abnormalities unassociated
with scute abnormalities, and (//) scute abnormalities correlated
with plate abnormalities. Parker finds the second specimen
shorter than the average normal specimen of the same sex in the
proportion of 1.298 to 1.313. This is due to the loss of a verte-
bra and pair of ribs.

In both cases it will be seen that the only real correlation
occurs in the marginals and that here the correlation is perfect.
Another correlation not mentioned by the author is seen in the
second specimen, where there is a reduction of marginals in sym-
pathy with the reduction in neurals and costals. This correla-
tion is, I believe, of a different sort from that seen in the marginal
plates and scutes, in that a common cause has brought out the
same general effect in both sets of structures. In both cases of

g6 H. H. NEWMAN.

correlation in the marginals Parker concludes that the abnor-
mality is in the anterior portion of the carapace, but an examination
of his figures fails to convince me that he has any criterion for
thus locating the point of suppression or recurrence of scutes,
for a loss anywhere in the marginals would necessitate a general
readjustment with reference to the costals. Yet the author con-
cludes that correlated abnormalities are likely to occur only in
the anterior portion of the carapace.

To aid this hypothesis he adds a second one, based on a paper
by Harrison ('98), in which it is shown that in the frog larva the
ectoderm is proliferated chiefly at the anterior end and the meso-
derm chiefly at the posterior end. This would cause the ecto-
derm to slide back over the mesoderm. Parker concludes that
the ectodermal structures in the carapace have migrated away
from their mesodermal connections so that the posterior scutes
are far from their original plates, while in the anterior part of the
carapace the scutes are over the same plates that they originally
covered.

This ingenious hypothesis loses its force when, after the ex-
amination of large numbers of abnormalities, it is found that nearly
all true correlations of plates and scutes occur at the posterior
end of the carapace.

A reference to the tabulation of abnormalities will bring to
light the following facts :

1. Specimens 2, 4, 6, 30, 45, 47, 61 and 62 show very precise
correlation of abnormalities in the marginals. All abnormalities
in the marginal scutes are correlated with similar abnormalities
in the plates.

2. Specimens 5, 6 and 15 have abnormalities of both plates
and scutes of the costal series, but in no case are these strictly
correlated. In all three specimens the extra costal plates are
paired while the extra costal scutes occur on the right side only.
These specimens are abnormally long and the undue length may
be the common cause of both extra plates and scutes.

3. No. 9 neural plate recurs in the same three specimens (5,
6 and I 5) and is associated in each case with one or more extra
neural scutes (Nos. 8 or 10). Examination shows that these
extra neural plates are irregular structures, are not associated
with the neural processes of vertebrae, and hence are to be con-

ABNORMALITIES IN CHELONIA.

97

sidered in the same category as the procaudals. On this account
I am strongly of the opinion that they have had a dermal origin
like the procaudals.

4. Specimens 5, 6, 9, 13 and 49 have extra procaudal, either
single or double, and in each case this plate abnormality is asso-
ciated with the recurrence of an extra scute in the posterior part
of the neural row. The procaudals are certainly of dermal origin
and were probably the bony cores of the last neural scutes before
the crowding-out process began. Consequently the reappearance
of an extra plate and extra scute in this region may with justice
be considered as a case of correlation.

5. Specimens 7, 8, 14, 17, 20, 21, 23, 27, 29, 31, 33, 41, 50,
51 and 58 have various kinds of scute abnormalities and per-
fectly normal plates.

Summarizing, we find that the only invariable correlations are
in the marginals which represent scutes with their bony supports
(the plates) only slightly separated from one another ; that corre-
lations between extra procaudals (also true dermal plates) occur
in five cases and no case of plate abnormality in this region is
without a scute abnormality ; that correlations between an extra
neural of small size and irregular shape occur in every case ; that
extra paired costal plates are associated with asymmetrical extra
costal scutes ; and that very frequently scute abnormalities appear
without any corresponding plate abnormality.

It would appear then that abnormalities are never truly corre-
lated except in regions where dermal plates persist and never in
the anterior part of the carapace, because the plates of that region,
except the nuchal, have given way to the periosteal plates of the
neural and costal series. The nuchal plate still possesses its
scute in a reduced condition, so we would not expect to find any
abnormality in that region.

Only in the regions of dermal plates do we find any interde-
pendence of plates and scutes, because only here, according to
our theory, is there any genetic connection between dermal and
epidermal structures.

That we find no true correlations between the plates and scutes
of the neural and costal series is just what we would expect from
our knowledge of their ontogenetic independence. In the most
irregular specimens, having the most grotesque scute displace-

98 H. H. NEWMAN.

ments and additions in the mid-neural and costal regions, there is
not a single case of sympathetic plate abnormality.

Another strong case in favor of the entire independence of
these structures is furnished by examples of vertebral distortions.
A paper by Wandolleck (1904) entitled "A Hump-Backed
Tortoise," describes a specimen of Tcstudo Grceca in which the
neural and costal plates were in utter confusion, due to lateral
curvature of the spine. Yet the number and arrangement of the
scutes was perfectly normal. Another case in point is that of a
deformed embryo in my possession that has ten ribs on one side
and the number of costal scutes is perfectly normal (Fig. 56).

Further evidence for the independence of neural and costal
scutes and plates is derived from the fact that certain land tor-
toises undergo striking modifications of plates in order to form a
dome-shaped shell of great strength. The costal plates become,
dovetailed one into the other and are decidedly wedge-shaped.
This condition is not followed in any respect by the scutes, which
retain their typical form and arrangement. A. Bienz ('94) de-
scribes this condition in Dennatcmys inavit, and shows that the
form and the arrangement of the plates conform with the most
approved architectural principles.

The origin of the bony plates of the plastron has been thor-
oughly worked out and in each case a plate is to be interpreted
as a modification of preexisting bony structures — clavicle, ster-
num or abdominal ribs. If such structures as these have been
transformed into large, flat, bony plates without any fusion of
fascia bones or any outside factors of any sort, why is it neces-
sary to explain the costal and neural plates as other than mere
modifications of the preexisting ribs and neural spines ?

6. Correlation in Scute Abnormalities.

Two tendencies may be noted in this connection :

1. Scutes of the same horizontal row have a tendency to recur
together (see Figs. 3, 5, 6, 7, 8, 9, 10, 12, 13, 14, 20, 31, 33
35,40). A i o neural most frequently is accompanied by a 10
costal, etc.

2. In asymmetrical abnormalities the tendency is for super-
numerary neurals, costals and marginals to occur on the same
side (see Figs. 3, 5, 7, 8, 9, 10, 12, 13, 14, 31, 35, 40).

BIOLOGICAL BULLHIN, VO^. X

Fig. i

Fig. 3

Fig. 4

Fig. 2

BIOLOGICAL BULLETIN VOL. X

Fig. 8

Fig. 6

BIOLOGICAL BULLETIN, VOL. X

PLATE III

Fig.i 9

Fig. 20

Fig.2i

Fig. 2 2

BIOLOGICAL BULLETIN, VCL. X

PLATE IV

Pig- 3

Fig. 4 o

Fig. 4 1

ilOLOGICAL BULLETIN. VOL. X

PLATE V

Fig. 4 8

Fig. 5 4

Fig- 5 2

F.ig.53

Vol. X. February, 1906. Aro.

BIOLOGICAL BULLETIN

THE SIGNIFICANCE OF SCUTE AND PLATE
"ABNORMALITIES" IN CHELONIA.

A CONTRIBUTION TO THE EVOLUTIONARY HISTORY OF THE
CHELONIAN CARAPACE AND PLASTRON.

H. H. NEWMAN.
CONTENTS.

PART II.

PAGE.
IV. The "Tail-trunk" of Chelydra serpentina, as a Close Approximation to

Ancestral Conditions of Carapace and Plastron 99

V. Further Evidence of the Former Existence of a Dermal Carapace in Che-
Ionia, as Derived from Specimens of Graptemys... 103

VI. The Color Pattern of Chelonia as Confirmatory Evidence of the Former

Existence of Dermal Armor 106

PART II.

IV. THE TAIL-TRUNK OF CHELYDRA SERPENTINA AS A CLOSE
APPROXIMATION TO ANCESTRAL CONDITIONS OF CARA-
PACE AND PLASTRON.

Baur, Hay and others have used as the hypothetical ances-
tral form an aberrant and perhaps highly specialized Chelonian,
DcrmacJielys coriacea, but in the light of the various atavistic re-
currences discussed in this paper and in view of the fact that
certain definite rows of scutes invariably predominate over others,
I have been led to seek elsewhere for primitive conditions.

A closer approximation to the true ancestral conditions is, I
believe, to be seen in that portion of the trunk of Chelydra that
is commonly called the base of the tail. Here if anywhere one
would expect to find primitive conditions. The Chelydridae are
generally acknowledged to be our most generalized chelonians,

99

IOO H. H. NEWMAN.

and it is natural to look for primitive characters in this family,
especially in the less specialized regions of the body.

A careful study of the portion of the body posterior to the
carapace, which for convenience may be called the "tail-trunk,"
is fruitful of suggestions. A preparation of the bony structures
of this region was made from a very large specimen (Fig. 58)
measuring nearly three feet from snout to tip of tail, having a
carapace sixteen inches long and a tail-trunk fifteen inches long.
The preparation shows the following structures :

There are 33 vertebrae ranging in size from very large bones
of nearly a cubic inch displacement to very minute ossicles at the
tip of the tail. The first five vertebrae "are beneath the carapace
and have their dorsal processes in close contact with a longitu-
dinal bony ridge that traverses the last three plates of the neural
row of the carapace. These five vertebrae have definite flattened
ribs that project laterally about at right angles to the axis of the
vertebral column and remind one of the flattened ribs in the cara-
pace of Dernwclielys. The first and second of these ribs are very
large and articulate by means of enlarged heads with the proxi-
mal ends of the ileum. These specialized vertebrae and ribs form
the sacrum (Fig. 58, I and 2).

Surmounting the dorsal processes of the eighth to the fifteenth
vertebra is a row of six large bony tubercles ( /. I to 6) that
have no reference to individual vertebras. The dorsal processes
of the latter are not the centers of ossification for the tubercles
and there is no articulation or fusion between the two series of
structures. Posterior to the sixth bony tubercle are fourteen
tubercles (j. 1 — 14), with either membraneous cores or no cores
at all, ranging in size from structures almost as large as the sixth
bony tubercle to extremely small scales with dorsal ridges. An-
terior to the first bony tubercle are three small soft tubercles (x)
occurring at intervals of about half an inch, and anterior to these
we find the two procaudal (//•. i and 2) and the single pygal
plate (/) of the carapace, overlying the first four vertebrae. We
have then the following heterogeneous series of structures : two
procaudals, one pygal, three small soft tubercles, six large tuber-
cles with bony cores and finally a graduated series of 14 tuber-
cles merging into ridged scales. It seems reasonable to suppose

"ABNORMALITIES IN CHELONIA. IOI

that all these structures of the dorsal row, whether scutes with
bony cores, scutes without bony cores, or bony plates that have
been separated from their original scute coverings, are essen-
tially homogeneous, and that the differences seen in the differ-
ent regions are secondary or perhaps tertiary modifications. It
seems probable that the scutes at the posterior end of the row
represent the most primitive condition, which, through a process
of continuous variation have become more tubercular in form
and have acquired bony cores by the gradual ossification of
membraneous tissue. The three soft tubercles surmounting the
fifth to seventh vertebrae probably represent a reduced condi-
tion in adaptation to the fact that the base of the tail requires
flexibility and must be swung from side to side or partially with-
drawn under the carapace. The presence of such large promi-
nences as the bony tubercles would seriously interfere with the
mobility of the tail. The procaudal plates seem to have been
the last of the neural tubercles to have been flattened out to
form the dermal carapace. The last neural scute is probably the
original chitinous sheath of one of the procaudals, doubtless the
second one. The scute of the first procaudal I believe to have
been crowded out in the process of scute reduction that will be
discussed later.

Over the 33 vertebrae there occur 27 structures of homogene-
ous origin, that may be designated scutes. This number is suffi-
ciently at variance with the number of vertebrae to preclude the
possibility that they have had a segmental arrangement. More-
over, their irregular arrangement and the dermal origin of the
bony cores make it certain that they are independent structures,
in opposition to views expressed by Gadow and others.

An examination of the entire tail-trunk of another large speci-
men revealed the rather striking fact that the number of principal
rows of tubercles and large scutes in this region is identical with
that of the carapace and plastron, and that smaller, less regular
rows of tubercles and scales represent the principal lost rows. A
section of the tail-trunk was slit down the median ventral line and
flattened out for convenience in drawing, and from this the some-
what diagrammatic Fig. 5 5 was constructed. I was able with
certainty to homologize seven principal rows, three dorsal, two

IO2 H. H. NEWMAN.

lateral and two ventral, and have named them after their homo-
logues in the carapace and plastron, neurals (TV), costals (C),
marginals (M), and plastrals (P). Smaller and less regular rows
of tubercles and scales are homologized with the secondary and
lost rows as follows : inframarginals (JM\ supramarginals (SAT),
interplastrals (IP], and neuro-costals (NC\ a row not occurring
in any modern species. In the following more detailed descrip-
tion the above names will be used for the rows concerned.

The costals are large scutes with a marked tendency toward
dorsi-ventral flattening. Their apices are directed posteriorly
and the growing point is thus situated near the posterior margin
of the scute, a fact that causes the scute to grow anteriorly and
laterally and very little posteriorly. The marginals are as large
as the costals and are flattened so as to expose two surfaces, a
dorsal and a ventral. The apex is directed away from the axis
of the body and the growth of the scute is principally inward and
slightly forward. The plastrals are paired, large, flat and nearly
rectangular. They suggest by their appearance the plastron
scutes. Their growth is inward and forward, as is that of their
homologues in the plastron.

The two secondary dorsal rows, supramarginals and neuro-
costals, consist anteriorly of small tubercles that fade out pos-
teriorly into small irregular flat scales. The inframarginals are
not tubercular but are rather large and fairly regular flat scutes.
The interplastrals are diamond-shaped scutes forming a discon-
tinuous row and occupying the angles made by four plastrals.
This row is sometimes entirely absent on the tail-trunk of
Chelydra.

It is clear then that in the tail and tail-trunk of Chelydra we
have seven principal rows of scutes and between each pair of
principal rows a secondary row, less regular and far less promi-
nent. This would give a total of fourteen rows, of which seven
are primary and seven secondary. Supposing that the carapace
was originally continuous with the tail-trunk, we must imagine a
gradual suppression of the secondary rows by the primary ones,
until in the most highly specialized condition, seen in the terres-
trial Emydidae, only the seven primary rows have survived.
The order of loss of the seven secondary rows may be con-

ABNORMALITIES IN CHELONIA. IO3

jectured from an examination of the directions of encroachment
of the principal rows upon the secondary rows. On this basis
one might give the order of loss as follows : first, neuro-costals,
which on account of inward encroachment of the costals and the
outward encroachment of the neurals, would receive the severest
pressure ; second, the supramarginals located between the in-
wardly-encroaching marginals and the outwardly-encroaching
costals ; third, the interplastrals, between the two inwardly-
encroaching plastrals, but occasionally able to escape pressure by
occupying angles between four plastrals ; fourth and last, the
inframarginals, which were subject to pressure only on one side —
that of the marginals — and were in addition required in more
primitive forms with small plastrons to help bridge the space
between the carapace and plastron. With the great increase in
the size of the plastron which has taken place in higher forms,
the space allotted to inframarginals became more and more con-
tracted until they were crowded out entirely.

Does this order of loss correspond to any facts in nature ? The
same conclusion as to order of loss, I believe would be reached
if we based our results on the prevalence, scarcity or absence of
these rows as normal structures in existing species. No trace of
neuro-costal scutes are found. Only one species, Macrochelys
tennnincki, possesses supramarginals. Interplastrals occur as iso-
lated median ventral scutes in several families. Inframarginals
occur normally in all the more primitive families and as axillaries
and inguinals in all but the most specialized terrestrial forms.

This condition as derived from the tail-trunk of CJielydra differs
rather radically from the ancestral condition derived from Der-
mochelys, in which there are twelve rows of equal rank. I am
inclined to believe that DernwcJielys is an extremely aberrant type
with only a most distant connection with the phylogenetic line of
Chelonia. The twelve keels of Dermochelys are comparable, I
believe, to the seven keels of modern forms, and the irregular rows
of plates and scutes between the keels are comparable to the
secondary rows of scutes seen in the tail-trunk of CJiclydra and
represented in the carapace and plastron by inframarginals, supra-
marginals, interplastrals and neuro-caudals.

IO4 H. H. NEWMAN.

V. FURTHER EVIDENCE OF THE FORMER EXISTENCE OF A

DERMAL CARAPACE IN CHELONIA, AS DERIVED FROM

SPECIMENS OF GRAPTEMYS.

The position just taken rests on the assumption that the cara-
pace and plastron were at one time continuous with that portion
of the trunk just posterior to them and that the carapace and
plastron have undergone a gradual process of specialization that
has caused them to depart widely from ancestral conditions. The
tail-trunk would then preserve to a greater or less extent its
original character.

On this assumption, then, there once existed a complete row
of dermal tubercular ossicles overlying the vertebrae. That
certain ancient forms did actually possess these ossicles was shown
by Hay in the case of Toxochelys serrifer, but the question arises
whether or not we have sufficient evidence that Toxoclielys repre-
sents the ancestral condition of our modern forms.

For a long time I looked for definite traces, other than the
keels, of such ossicles as are seen in Toxochelys, but met with
no success until I had nearly completed the present paper. Then
by merest chance I stumbled on the evidence needed to clinch
the argument.

I had kept alive a few specimens of Graptemys in a small
aquarium, but one by one they sickened and died, with one ex-
ception. Their death was doubtless due to the fact that this
species is highly specialized in its diet, feeding exclusively on a
species of viviparous gastropod that is abundant in Lake Max-
inkuckee. They never learn to use other food, and, in lack of
their special diet, starve themselves to death.

The surviving specimen was a nearly adult female that had
been kept on account of its many peculiarities. After eleven
months of captivity it was killed and examined for plate abnor-
malities. This examination revealed the presence of several
small, loose, ossicles that were inlaid, as it were, in the bone of
the neural plates and were situated beneath the keels of the sec-
ond, third, fourth and fifth neural scales, /. e., exactly in the
positions occupied by the ossicles found in Toxochelys. The
largest of these ossicles (see Fig. 5) was situated beneath the
keel of the third scute and extended partially under the anterior

"ABNORMALITIES IN CHELONIA. IO5

margin of the fourth scute. The fifth scute is a supernumerary
neural (No. 8) and hence the occurrence of an ossicle at its keel
shows that it belongs in the neural row in spite of the fact that
it is crowded to one side. All of the ossicles are imbedded or
inlaid in the centers of certain of the neural plates. If they
were merely the loosened centers of these plates, we would
surely expect to find them on all the plates instead of just
those which lie beneath the keels of the scutes. In what respect
do these ossicles, then, differ from those seen in Toxochelys
serrifcr by Hay ? Merely in this, that they are much reduced in
size, so that each is confined to one neural plate, extending back
so as to overlap the anterior portion of another plate.

The specimen is an oddity in many respects. It is unusually
long in the carapace ; possesses fifteen plates in the neural row
instead of the normal number,, twelve (three of these probably
representing supernumerary procaudals) ; two extra costal plates
or ribs, one quite vestigial ; two supernumerary scutes of large
size ; one supernumerary costal scute of large size on the right
side ; and two well-developed supernumerary inframarginals on
each side. Other minor peculiarities might be noted, but they
do not concern the carapace or plastron. All of the anomalies
mentioned may be viewed as of atavistic character, and it should
not be surprising to find that the curious specimen shows an even
more significant atavistic recurrence than any other specimen thus
far examined, namely, a reversion to the condition seen in Tox-
ochelys. That the genus Grapteinys originally possessed a me-
dian dorsal keel composed of prominent bony tubercles covered
with chitinous sheaths (scutes) is rendered extremely probable
when we examine the young, especially that of G. pseudogco-
grapJiica, a specimen of which is pictured in Agassiz' Contribu-
tions to the Natural History of the United States, Vol. II., Plate
II., Figs, ii and 12. This specimen, which may perhaps be an
extreme type, although Agassiz does not suggest that such is
the case, shows a series of three very remarkable dorsal tubercles
on the second, third and fourth scutes. These tubercles furnish
a close approximation in general form to those seen on the tail of
Chclydra. It will be noted that these tubercles occur exactly in
the places where I have found the vestigial ossicles in a specimen

IO6 H. H. NEWMAN.

of Graptemys geograpJiica, and that the most prominent tubercle
is that on the third scute, a fact that is of interest in view of the
larger size of the vestigial ossicle on that scute.

The adult of Graptemys psctidogcographica retains decided
traces of these tubercles throughout life and their location is
marked by dark blots of pigment that in later life form the only
prominent color-markings of the animal. The discovery of ves-
tigial ossicles in Graptemys geographica led me to investigate a
very large female specimen of Graptemys psendogeographica, that
has been in confinement, and, like its relative, starving for many
months. Immediately beneath the dark blots of pigment on the
neural scutes there are on this specimen thin, scale-like discs of
bone with a looser and less dense texture than the underlying
bony plates. I have been able to examine but one adult spec-
imen of Graptemys psciidogeograpliica, but this one appears to be
perfectly normal in every other respect. It is to be noted that
the vestigial plates in the last-mentioned specimen occupy exactly
similar positions, with respect to scutes and plates, as do the ves-
tigial ossicles in the anomalous specimen of Graptemys geo-
grapliica, described and pictured above (Fig. 5). The probable
explanation of both these conditions is that long-continued star-
vation has brought about a resorption of the portions that
united these ossicles with the underlying neurals. That bone is
resorbed either by normal processes or as the result of patholog-
ical conditions I have observed in several cases where holes have
been eaten entirely through the bone of the carapace, as the
result of starvation. Only a thin cap covering the top of the
tubercle is of dermal origin, the main portion of the prominence
being merely an outgrowth of the periosteum of the neural
process. Examinations of developmental stages have revealed
no discontinuity between the cap and the rest of the keel, but the
microscope reveals the difference in histological structure be-
tween the cap and the underlying bony plate.

That tubercular ossicles existed over the neural processes of
other ancient reptiles is shown by the fossil Stegosaurus. Here
the dermal processes are very large and prominent and are much
fewer in number than the neural processes that underlie them. This
points to the entire independence of dermal bones and the vertebrae,
and hence to the non-segmental character of the dermal plates.

"ABNORMALITIES" IN CHELONIA.

VI. THE COLOR PATTERN OF CHELONIA AS CONFIRMATORY EVI-
DENCE FOR THE FORMER EXISTENCE OF A DERMAL ARMOR.

Evidence is not lacking that points to an original striped con-
dition of the chelonian carapace. The neck and tail of most
tortoises show characteristic stripes which on careful examination
may be analyzed into rows of scales with similar coloration.
When the scales are large enough, it will be seen that each has a
center of pigmentation coinciding with its center of growth. Now
the coloration of the carapace and plastron is nothing more than
a series of scales or scutes, each with its pigmental center. The
striped effect is lost through the great increase in the size of
the scute and the consequent separation of the centers of pig-
mentation.

The pigmentation of scutes is typically concentric in character,
whether the pattern consist, as in CJielydra and Aroinoc/ielys, of
radiating bands of pigment having their center located at the
center of scute growth, or concentric rings as in Graptcmys, or
lastly, of a light area occupying the center of growth, and all
the rest solidly colored, as in Clcmnys guttatits.

Frequently a great complexity of marking arises through the
secondary complications of primary markings, but these conditions
are seen in a simplified condition in the developmental stages.

It strikes one very forcibly that there is an intimate relation
between scute growth and pigment distribution. The two
processes have a common center and go hand in hand. The
areola, or egg-plate (see Fig. 57, dotted lines) forms a convenient
locus for measuring both processes. This location corresponds,
curiously enough, with the keel of the scute and hence with the
center of dermal ossification.

An examination of embryos of CJielydra shows that the colora-
tion consists of dark patches of melanin pigment at the tip of the
tubercular processes of the keels. The marginals are marked
with small black spots at the posterior edge of each scute, some-
times running back over the anterior margin of the next scute.
Specimens of CJielydra ?L year or two old have a radiating pattern
with the center of pigment proliferation' at the keel. In older
specimens a solid coloration obscures everything.

Older embryos of Graptemys have, as the first indication of pig-

IO8 H. H. NEWMAN.

ment, one dark spot at the median posterior margin of each scute,
i. c., at the keel (see Fig. 56). Later on two or more other spots
of a similar appearance are produced in very definite positions on
the various scutes. These secondary spots never become quite
so prominent as the primary ones, but continue to develop like
the latter into concentric ocellated markings (Fig. 57). The rings
that constitute these ocellated spots are formed by repeated split-
ting of the innermost ring of pigment, in brief are proliferated
from a center of pigment deposit. After three or four cencentric
rings have been laid down, a well-marked unpigmented band ap-
pears on the periphery of the whole spot. It is this light band
that produces the most pronounced secondary complexity, for it
sends out processes to neighboring spots and forms the charac-
teristic reticulated pattern that has given the name " map-turtle "
to the species. In the marginals the pattern is not so complex,
since no secondary spots appear. The concentric rings of the
primary spot, however, often take on fantastic shapes that all but
obscure the original unity of the coloration. It is very interest-
ing to note that the spots in the marginal series are found as a
rule half on one scute and half on the next, so that the light per-
ipheral band that separates adjacent spots frequently coincides
with the sutures of the dermal plates beneath. The spots then
occupy the growth centers of the plates and no longer hold any
close relationship with the scutes. Pigmentation scans to be inti-
mately connected with dermal ossification. The scutes must have
grown away from their original positions, thus overlapping the
sutures of the plates, and lending strength to the general struc-
ture. The direction of movement has been forward, as is found
by examination of the first pair of marginal scutes that run over
upon the nuchal plate. The original cause of the forward move-
ment was, I believe, the increase in size of the nuchal plate which
must have pushed back the marginal plates. The scutes would
of course occupy their original positions or would continue to
crowd the nuchal scute into a still smaller space.

That the primary ocellated spots denote centers of dermal ossi-
fication will, I believe, be admitted. What then is the signifi-
cance of secondary ocellated spots, which have an exactly simi-
lar appearance and method of development ? They must, I

"ABNORMALITIES" IN CHELONIA. 109

believe, represent centers of dermal ossification and hence the
location of scutes that formerly occupied the position now occu-
pied by the spots. Evidences that serve to increase the proba-
bility of this conclusion may be adduced :

1. Whenever supernumerary scutes recur each has at its
growth center an ocellatecl spot.

2. These supernumerary scutes come in at places where normal
specimens have definitely placed secondary ocellated spots.

3. There are never any secondary ocellated markings on the
marginal scutes, which agrees with our idea that the marginal
rows contain nearly their original number of scutes and hence we
would not expect to see traces of lost scutes in these rows.

4. The light bands that form the reticulated pattern of adult
Graptemys have often the exact shapes of existing dermal plates.
This is particularly the case in the procaudal region.

5. Considering the light bands as original scute boundaries,
we can count ten costal scute areas in Grapteinys or TracJiemys.

6. On the neural scutes of Graptemys there are several much
smaller ocellated spots that lie near the outer edges of the scute.
These spots, I believe, are the vestiges of the small scutes that I
have earlier designated as neuro-costals, and that were the earliest
rows of scutes to be crowded out (Fig. 57). Four or more
well-developed spots occur close to the marginals on the costals
and occupy positions similar to the supramarginal scutes of
Macroclielys tcmmincki.

At the angles made by adjacent neurals and costals occur
ocellated spots that have the appearance of having been squeezed
out between two scutes. These I believe to be the vestiges of
the lost neurals and costals (see Fig. 57). When the lost scutes
recur they have these spots at their growth centers.

In the bridge region of Graptemys and Clirysemys a confused
series of dark colorations appears that seems to have no reference
to any existing structures. But when the inframarginals recur,
we find that these apparently meaningless markings fall into place
as the spots of this lost row of scutes.

On the plastron scutes we find in most species a spot of pig-
ment for each member, but confusing secondary complexity
often obscures the real pattern. In Graplemys some specimens

BIOLOGICAL BULLETIN, VOL. X

PLATE VI

Fig. 55

GO

IV

\r

\i

I

v

Fig. 5 6

Fig. 5 7

Fig. S3

"ABNORMALITIES IN CHELONIA. Ill

show a fairly complete set of ocellated spots, but others show
almost colorless scutes. Adult specimens seldom show any sign
of pigmentation on the plastron. The markings then are largely
juvenile in character and are subject to very great individual vari-
ation that tends toward the total obliteration of the original color
pattern. An examination of a number of specimens, however,
shows that each of the plastron scutes may have its ocellated
spot.

The most confusing markings of all are those that occupy the
median portion of the plastron in juvenile specimens. A remark-
able secondary complexity of pattern has arisen in this region
that would be almost impossible of solution were it not for the
close series of stages leading up from the simplest conditions.
The simplest form of marking consists of small diamond-shaped
patches of pigment at the inner angles of the plastron scutes.
These spread along the margins of the scutes, form bands by
splitting, and finally produce the complex lyriform pattern that
one finds quite frequently. The position of the simplest marking
is identical with that of the hypothetical interplastral scutes and
probably once constituted the color-marking of this row of scutes.

All of the carapace and plastron markings have thus been
accounted for as the growth centers of existing or lost scutes.
This has been done in a species with a higftly intricate color pat-
tern and could be applied successfully, I believe, to any other
species.

It should be mentioned that the color pattern of Graptemys
reaches its highest development in specimens of the first year.
This is the time when protective coloration is a necessity, as the
carapace is not sufficiently Ossified to furnish a protection. Old
specimens retain scarcely a trace of the original pattern, only a
very faint reticulation being visible.

It might be suggested that the ocellated spots of the Triony-
chidse are vestiges of scutes long since lost. The general number
and arrangement of these spots tends to bear out this suggestion.

SUMMARY.

Palaeontological and embryological evidence is at variance as
to the origin and character of the neural and costal plates, but

112 H. H. NEWMAN.

observations point strongly to a periosteal origin of these struc-
tures, which means that^ they are in no sense dermal or the de-
scendants of the original dermal carapace.

The testimony of comparative anatomy leads to the belief that
the nuchal, procaudal, pygal and marginal plates are the remnants
of a once more or less complete dermal carapace and that these
plates formed the cores of scutes that must have had a more or
less tubercular form. The keels of existing scutes represent these
tubercules. The testimony of the tail-trunk of Chelydra indi-
cates that there were originally seven primary rows of such scutes
and that less prominent rows of scutes occupied the interspaces.
These less prominent rows were gradually suppressed, first on the
carapace and then on the plastron, beginning in the middle and
proceeding laterally. Thus the first loss was the neuro-costal
rows, second the supra-marginals, third the interplastrals, and
fourth the inframarginals, \vhich to-day persists normally in many
primitive forms.

Accompanying the suppression of rows occurred a reduction
in the number of scutes in the primary rows, and this reduction
took a general antero-posterior direction. At the same time the
rapid secondary growth of the neural spines and ribs caused the
suppression of the corresponding dermal plates, leaving only the
nuchal, procaudals, pygal and marginals in places where the in-
ternal skeletal portions failed to extend. Traces of the dermal
armor in the mid-neural region have been found, however, in
ToxocJiclys and Graptcmys.

No correlation of abnormalities is to be expected in the neural
and costal regions, since the scutes and plates of this region are
entirely independent in origin, but in the marginal series, where
the plates and scutes retain nearly their original connections, the
correlation is perfect. In the procaudal region we find frequent
correlations, but, that the correlation is not a necessary one, is
shown by numerous uncorrelated abnormalities.

A study of the color-markings of Graptentys and CJielydra
lends confirmation to all the above theories of the chelonian car-
apace and plastron, and at the same time serves to rationalize
the patterns themselves.

"ABNORMALITIES IN CHELONIA. 113

BIBLIOGRAPHY.
Agassiz, L.

'57 Contributions to the Natural History of the United States, Vol. II., 1857.
Baur, G.

'96 Bemerkungen uber die Phylogenie der Schildkroten. Anat. Anz., Bd. 12,

Xo. 24/25, 1896.
Baur, G.

'89 On the Morphogeny of the Carapace of the Testudinata. Amer. Nat., XXI.,

1889.
Baur, G.

Die systematische Stellung von Dermochelys. Biol. Centralbl., Bd. IX.
Bemmelen, J. F. van

'96 Bemerkunger iiber den Schadelbau von Dermochelys coriacea. Gegenbauer

Festschrift II., 1896.
Bienz, A.

'95 Dermatemys mavii, Grey. Eine osteologische Studie mit Beitragen zur Kent-

niss vom Bau der Schildkroten. Rev. Suisse Zool., III., 1895.
Coker, R. E.

'05 Gadow's Hypothesis of " Orthogenetic Variation in Chelonia." Johns

Hopkins Univ. circular, No. 178, May, 1905.
Cope, Edw. D.

'96 The Ancestry of the Testudinata. Amer. Nat., Vol. 30, 1896.
Dollo, M. L.

'86 Premiere Note sur le Cheloniens du Bruxellien (Eocene Moyen) de la Bel-

gique. Pamphlet. 1886.
Gadow, H.

'99 Orthogenetic Variation in the Shells of Chelonia. Zool. Results. Willey,

A., 1899.
Gadow, H.

'01 Amphibia and Reptiles. The Cambridge Nat. Hist., Vol. VIII., 1901.
Gegenbauer, C.

Vergleichende Anatomic der^Wirbeltheire. Erster Band, Leipzig, 1898.
Goette, A.

'99 Uber die Entwicklung des knockeren Riickenschilds der Schildkroten.

Zeit. f. wiss. Zool., Bd. LXVL, 1899.
Gray, J. E.

'55 Catalogue of Shield Reptiles in the Collection of the British Museum. " Part

I. London, 1855.
Harrison, R. G.

The Growth and Regeneration of the Tail of the Frog Larva. Arch. f.
Entwicklungsmechanic d. Organism. Bd. VII., pp. 431-485.
Hay, 0. P.

'91 The Batrachians and Reptiles of Indiana. I7th Report of the State

Geologist, 1891.
Hay, 0. P.

'97 On Protostega, the systematic position of Dermochelys, and the Morphogeny

of the Chelonian Carapace and Plastron. Amer. Nat., XXXII., 1897.
Hay, 0. P.

'oo The Composition of the Shell of Turtles. Science, Vol. 13, p. 624. 1900.

114 H. H. NEWMAN.

Haycraft, J. B.

'90 The Development of the Carapace of Chelonia. Trans. R. Soc Edinh.,

XXXVI, 1890.
Jordan, D S

'99 A Manual of the Vertebrate Animals of the Northern U. S. 1899.
Owen, R.

'49 On the Development and Homologies of the Carapace and Plastron of Che-
Ionian Reptiles. Phil. Trans., Lond., 1849.
Parker, G. H.

'01 Correlated Abnormalities in the Scutes and Bony Plates of the Sculptured

Tortoise. Amer. Nat., Vol. 35, 1901.
Rathke, H.

'48 Ueber die Entwicklung der Schildkroten. 1848.
Schoepff, J D.

'92 Historia Testudinum. 1792.
Stoffert, A. T.

'89 Bau und Entwicklung der Schaale von Emyda ceylonensis. Basel, 1889.
Van Lidth de Jeude, Th. W.

'98 On Abnormal Pectoral Shields in Testudo ephippium. Notes Leyden Mus.,

Vol. 20, 1898.
Wandolleck, B.

'04 An Abnormal Tortoise. Zool. Jahrb., XX., pp. 151-166, 1904.
Werner, F.

'95 Bemerkunger uber Schildkroten-zeichnung. Biol. Centralbl., Bd. 14, 1895.

NOTE ON THE INFLUENCE OF SURFACE-
EVAPORATION UPON THE DISTRI-
BUTION OF INFUSORIA.

T. BRAILSFORD ROBERTSON.

(From the Rudolph Spreckels Physiological Laboratory of the University of Cali-
fornia. )

During the course of my experiments on the chemotaxis of
Paramccdum and Colpodiinn L I was struck by the fact that in
certain media the infusoria showed a remarkable attraction to the
edge of the fluid under the cover-glass — these media were aV/25
methyl alcohol, TV/so CaCl2, ti/ 5,000 NaOH and N/ 5,000 KOH.
This is contrary to the usual behavior of the organisms, in other
saline media, in sugar solutions and in the culture medium they
avoid the edges of the film of water under the cover-glass and,
if undisturbed, form a cluster in the middle of the film, or, if the
cover-glass be supported at one end, a little away from the center,
towards the supported end.2

Jensen attributed this to the increase in concentration at the
edges of the film due to evaporation. Since infusoria appear to
be very generally attracted by solutions of lower concentration
they tend to congregate in the more dilute part of the film, that
is, near the center. In considering the apparent contradiction to
this rule displayed by the infusoria when immersed in the above-
mentioned media it struck me that in all these cases we had the
converse of Jensen's experiment — for these media tend to
become more dilute at the edge of the film. Thus methyl
alcohol evaporates more rapidly than water, so that a solution of
methyl alcohol becomes more dilute as it stands exposed to the
atmosphere — while solutions of highly hygroscopic substances
CaCl2, NaOH and KOH absorb water-vapor from the atmos-
phere and so also tend to become more dilute at the surface. In
order to see whether this was the case I tried the effect of adding

^Journal of Biological Chemistry, January, 1906.
2 Jensen, Pftiiger1 s Archiv. /urges, physiol., Vol. 53 (1893), p. 428.

Il6 T. BRAILSFORD ROJ5ERTSON.

different reagents to the culture-medium, in which the infusoria
were suspended, upon their distribution under the cover-glass.

When a substance which evaporates more rapidly than water
is dissolved in a watery medium containing salts in solution two
contrary effects take place on evaporation — the solution becomes
more dilute in respect to the more rapidly evaporating substance
owing to its evaporation and more concentrated in respect to the
salts owing to the evaporation of water. When the decrease in
osmotic pressure due to the evaporation of the more rapidly
evaporating substance is greater than the increase in osmotic
pressure due to the increased concentration of the salts — then
the total result is a decrease in the osmotic pressure. But if the
initial concentration of the more rapidly evaporating substance be
less than that necessary to bring about the above result then the
increase in concentration of the salts will proceed more rapidly
than the decrease in concentration of the volatile substance and
the total result will be an increase in osmotic pressure at the sur-
face of the fluid. Thus it was to be expected that while under
the ordinary conditions of my experiments on Chemotaxis —
when to a very small amount of culture a comparatively large
amount of solution was added --methyl alcohol at a concentra-
tion of Ay25 caused a dispersion of the infusoria to the edge of
the cover-glass, when I added I c.c. of a solution of methyl alco-
hol to 5 c.c. of culture it was found necessary to bring the final
concentration of the methyl alcohol up to |A7 in order to get
the first indication of a tendency to seek the edge — since in this
latter case the salts were correspondingly less dilute. The solu-
tion of the volatile or of the water-absorbing substance in the
culture medium having been made, drops of the mixture were
placed on a slide under a cover-glass which was slightly raised
at one end.

The following are the experimental results :

Methyl Alcohol.- -\ c.c. of a ^N solution of CH3OH was
added to 5 c.c. of the culture-medium in which numerous colpodia
and paramcecia were suspended. Result at first uniform distribu-
tion of the infusoria under the cover-glass, in a few minutes, how-
ever, they began to congregate at the edges --especially at the
edges farthest away from the supported end. In a short time all

THE INFLUENCE OF SURFACE EVAPORATION. I I/

parts of the film, except the edges, were free from infusoria, while
dense clusters had been formed along the edges farthest from
the supported end. When a 6N solution of methyl alcohol was
used instead of a 5 A7" solution the effect was more marked and
appeared more quickly.

Ethyl Alcohol. — i c.c. of a 5 A" solution of C2H5OH was added
to 5 c.c. of culture. The same effect was obtained as with methyl
alcohol, but it was slightly less marked and took longer to appear.

Propyl AlcoJiol. — jj-A7" C3H,OH was too toxic — the organisms
being killed too soon to obtain any result — but in -y^A7, which
is about the strongest solution they will stand, there was no
attraction to the edges of the film.

Ethyl Acetate. — i c.c. of a saturated aqueous solution of ethyl
acetate was added to 5 c.c. culture. The organisms were at
first uniformly distributed — they then showed marked attraction
for the edges of the film, particularly for those edges most remote
from the supported end — and in a few minutes all the infusoria
were congregated at the edges at the shallower end of the film

Calcium Chloride. — i c.c. of a fA7" solution of CaCL was added

O 4

to 5 c.c. of the culture — attraction to the edges was observable
for a short time. As in the other cases the edges farthest from
the supported end became the most densely populated.

Potassium and Sodium Hydroxides. — In my experiments on
Chemotaxis and under the conditions of those experiments I
obtained a marked, though more or less transient, attraction to
the edges of the film at a concentration of A7"/ 5,000.

All the above results hold equally for Colpodinm and for Para-
maecium.

In all cases the attraction to the edges finally disappeared -
with ethyl acetate the effect passes off in about half an hour.

All these substances which I have found to cause aggregation
of the infusoria at the edges of the film are substances the solu-
tions of which tend to become more dilute on exposure to the
atmosphere. Methyl alcohol, ethyl alcohol, and ethyl acetate
evaporate more rapidly than water while CaCl2, KOH and NaOH
are well known absorbents of water-vapor. There is also a gen-
eral correspondence between the rate at which the volatile sub-
stances evaporate and the intensity of the effect which they pro-

Il8 T. BRAILSFORD ROBERTSON.

duce — thus methyl alcohol has a lower boiling-point than ethyl
alcohol and it also causes a more marked attraction to the edges
of the film.

The fact that in all these cases the attraction to the edges of
the film disappears after a certain time may perhaps be due to the
volatile substance having become so dilute in the film that the
increase in concentration of the salts due to the evaporation of
water now takes place more rapidly than the decrease in concen-
tration of the volatile substance. While in the case of water-ab-
sorbing substances — they may have become so dilute that the
water-absorption due to their presence in the solution now goes on
less rapidly than the loss of water due to evaporation. It seems
probable therefore that these phenomena of attraction to or re-
pulsion from the edges of the film are in reality special cases of
osmotaxis.

I have alluded to the fact, mentioned by Jensen in the paper
to which I have referred, that under ordinary circumstances Para-
nicccia under a cover-glass supported at one end tend to collect
in the middle of the film but towards the supported end. But when
the substances which cause the infusoria to congregate at the
edges of the film are added to the culture we obtain precisely the
converse effect — the infusoria collect at the edges, especially at
the edges farthest from the supported end. It thus appears prob-
able that this is also an osmotactic phenomenon connected with
surface dilution. One would be inclined to fancy that it was due
to surface evaporation having more effect upon the concentration
of the small bulk of liquid at the shallower end of the film than
upon the greater bulk of liquid at the supported end — but the
ratio of surface to volume is not greater at the shallower end
than at the supported end. Dr. Loeb has suggested to me that
it may be due to the fact that diffusion and therefore equalization
of concentration is less rapid in capillary spaces than in the bulk
of the liquid — that, in fact, the shallow end of the film may act
in a manner analogous to Liebreich's " dead space." *

1 Zeitschrift Jiir Physikalische C/iewie, Vol. V. (1890), p. 529 and Vol. VIII.
(1891), p. 83.

THE INFLUENCE OF SURFACE EVAPORATION. I 19

CONCLUSIONS.

1. That whereas, under ordinary circumstances, infusoria tend
to collect near the center of a film under a cover-glass, when
methyl alcohol, ethyl alcohol, ethyl acetate, CaCl.,, KOH or
NaOH are added in sufficient concentration to the culture-
medium, this tendency is reversed and the organisms now gather
at the edges of the film.

2. That this lends support to Jensen's explanation of the nor-
mal tendency of the infusoria to gather near the center of the
film, namely, that it is a special case of osmotaxis brought about
by surface- evaporation.

HISTOGENESIS IN INSECT DEVELOPMENT, AND

CELL SPECIFICITY.

VERNON L. KELLOGG.

In the development (ontogeny) of insects with complete meta-
morphosis the imaginal antennae, mouth-parts, legs and wings
are produced from small buds, or histoblasts, in the larval derm.
These histoblasts or imaginal buds arise by the shallow or deep
invagination (one for each histoblast) of small regions of the
larval cellular skin layer including originally, in each case, only
comparatively few derm cells. The position of these invagina-
tions, and therefore the participation of derm cells in the future
leg or wing development, seems to be determined wholly with
reference to the future imaginal organ, and not at all with refer-
ence to any difference in degree of differentiation among the cells
of the larval derm. The wing-buds arise from the latero-dorsal
regions of the meso- and meta-thoracic segments, the leg-buds
from the latero-ventral regions of each of the three thoracic
segments.

The larval derm is certainly not to be looked on as composed
of wholly undifferentiated embryonic cells. These derm cells
make up a definite organ, or part, of the larval body, with defini-
tive position and particular functions. All these cells, and no
others in the body, secrete1 chitin ; some of them secrete noxious
fluids, ill-smelling, acrid, poisonous. Many of them, perhaps all,
secrete moulting fluid at the times of the regular larval moults ;
many are specially sensitive, many bear sense-hairs or papillae.

The invagination and beginning development of small parts
of this derm, in a wingless and legless larva of a fly or honeybee
or of some other specialized insect, may occur in a very early lar-
val stage (in some cases, indeed, indications of the future histo-
blasts are apparent in just-hatched larvae), but in most cases the
invagination does not appear until a certain part of the free lar-
val life has been lived. That is, the larval derm has for awhile

1 The chitin secreting capacity of the anterior and posterior thirds of the insect ali-
mentary canal is due to their deeply invaginated derm.

120

HISTOGENESIS IN INSECT DEVELOPMENT. 121

subserved only, and has subserved fully, the functions of skin, and
its cells have certainly attained whatever differentiation they need
for the performance of these functions. It is far and away a long
cry back to the embryonic, undifferentiated, the non-specific con-
dition. But apparently any part or region of this derm which
may, by its position, be the region or part needed to develop an
antenna, a wing, a leg, male clasper, female ovipositor, or a sting,
can respond to the need, and by invagination (for protection's
sake), rapid growth and proliferation of cells, quick differentiation
and arrangement, and final evagination (at the time of the last lar-
val moulting, /. e., pupation) produce the needed organ. This
organ may be tubular and segmented, and the segments may be
similar (antennae) ; or dissimilar (leg) ; or it may be a great flat-
tened sac, supported by tubular skeletal ribs and covered by a
million and more tiny other striated, pigment-bearing, flattened
sacs (the butterfly's wing with its scales) ; or it may be the ex-
quisite mechanism of the bee's sting.

This histogenesis of the imaginal parts of the fly, the bee, and
the moth is, to my mind, an extremely suggestive phenomenon
when considered in the light of its relation to the theories of cell-
specificity or cell-non-specificity. Quite as positively as the
more familiar cases of restorative regeneration (legs, eye-lenses,
tails and what not of various vertebrates), does this radical histo-
genesis, common to the ontogeny of all insects with complete
metamorphosis, make it impossible to limit the germ-plasm to the
germ-cells. It stands strongly opposed to any theory of abso-
lute cell-differentiation or cell-specificity.

STANFORD UNIVERSITY, CAI.IF.

ONTOGENY OF THE ANNULUS VENTRALIS.

E. A. ANDREWS.

Ill crayfish of the genus Cambarns there is a sperm receptacle
in the female which has been known as the Annuhts ventralis
and made use of as a specific character in systematic works.

Nothing has been known of its mode of development beyond
the brief mention by Mary Steele L who figured the external views
of the annulus in specimens of C. gracilis, 20, 22, 27.5, 30, 35,
36, 50 and 60 mm. in length. Of these the latter four were
probably adults and the former four immature young. These
figures show an increasing complexity of external sculpturing
and an increase in relative longitudinal diameter ; however, as
they were made incidentally and without reference to the use or
internal structure of the organ they are necessarily insufficient.
While the annulus is a necessary reproductive organ in Cam-

KlG. I.

barus it is also a new organ in the history of crayfish since it is
found only in the most specialized form, Cambarns. It seemed,
therefore, of interest to find out just how and when the organ
develops in the ontogeny of Cambarns.

The following account of the origin and growth of the annulus
refers almost exclusively to Cambarns affinis reared in the labo-

l(J>iii'. of Cin. Bulletin, X., 1902.

122

ONTOGENY OF THE ANNULUS VENTRALIS. 123

ratory. In large adults, 100 mm. long, the annulus in this
species has the appearance outlined in Fig. I, which is enlarged
about twelve diameters. It is a transversely elongated plate,
part of the shell, with a central depressed area bounded behind
by a cross ridge and in front by two high tubercles or tuberosities.
Across the depressed area runs a zigzag line which is in reality a
closed suture whence a slit leads inward to a curved tube repre-
sented by the thick shaded line. The suture and the curved
tube both open out on one side into the depressed area by an
orifice partly under one of the tuberosities. The walls of the tube
are thick chitinous continuations of the shell, as indicated by
the broken lines. Underlying this chitinous mass is the epi-
dermis which forms it and which was found to be folded in as a
bent groove. A comparative study of annuli in several species
showed that while the external sculpturing is various, the presence
of a curved epidermal groove is constant and that, morpho-
logically, this sperm receptacle is a bent epidermal pocket lined
by chitin and opening to the exterior by a more or less closed slit.
The position of the annulus, as seen in Fig. 2 which is enlarged
one and a half diameters, is on the ventral surface of the thorax

FIG. 2. FIG. 3.

between the bases of the fourth pair of legs. The sternal surface
behind it is elevated as a rounded knob that may be of importance
in discharging the receptacle. Projecting forward towards this
are the short pleopods of the first abdominal somite. Anteriorly,
on the bases of the third pair of legs, are the large elliptical open-
ings of the oviducts whence the eggs when laid pass back over the
annulus.

124 E. A. ANDREWS.

In using the annulus as a sperm receptacle the male passes the
sperm into the orifice and thence into the posterior part of the
tube. The anterior part of the organ, the orifice itself and the
following transverse part, which may be called the vestibule, is
filled not with sperm but with a cement that protects the sperm
from the water. Eventually the sperm issues out through the
more posterior part of the suture at the right time to meet the eggs.
When the young C. affinis hatches from the egg there is no
annulus present. The ventral surface (Fig. 3), multiplied fifty
diameters, shows no specialization of the wide level area between
the bases of the fourth legs. The sexes are not yet distinguish-
able and the first abdominal appendages are in all cases but very
faintly indicated by slight elevations. Here under a higher power
the epidermis was seen to be specialized as a group of nuclei, over

which the cuticle was elevated
as a slight protuberance in the
region that is later to grow out
as the first abdominal appendage.
After this larva sheds its shell
and passes into a second stage
the sternal surface is larger (Fig.
4), and the first abdominal ap-
pendages are somewhat more
protuberant. There is, however,
^ s_x_ no annulus and no external signs

of sexual differences. Between
FIG. 4.

the bases of the fifth legs there

is now a transverse elevation of the sternum, which will become
the prominent tubercle posterior to the annulus in the adult.

After a second moult the larva in the third stage has a much
larger sternal area and for the first time sexual openings and the
beginning of the annulus. In the female (Fig. 5), the sternum
between the fourth legs is divided by a cross-line into an anterior
part articulating with the legs, and a broad posterior region that
is, however, markedly short. On the middle of this plate and at
its posterior edge is a slight depression or groove destined to
become the receptacle for sperm in the adult. In the male the
same differentiation of a posterior sternal area is found, but the

ONTOGENY OF THE ANNULUS VENTRALIS.

125

plate lacks any groove and henceforth remains without any spe-
cial development. The female is also recognizable by the ap-
pearance of the short, curved
cuticular ridge on the base of
the third leg, each side of the
body, which is to be the me-
dian edge of the future orifice
of the oviduct. From this
superficial ridge a cylindrical
epithelial tube, the oviduct,
leads into the interior of the
body. The first abdominal

FIG. 5.

appendages, or pleopods, are

blunt papillae and apparently

the same in both sexes, though sometimes in the male they seem

more pointed and possibly longer.

The annulus is thus a narrow transverse part of the sternum
and since its posterior edge projects and overhangs somewhat it
might be regarded as a sort of transverse fold. Its surface is en-
tirely flat and simple as compared with the complex adult surface,
Fig. i, except for the slight median depression. Looking at the
shell only, Fig. 6 enlarged two hundred diameters, this depres-
sion is a wide shallow groove indicated by converging wrinkles
in the cuticle on each side the middle line of the animal and ex-

FIG. 6.

FIG. 7.

tending forward but a slight distance from the posterior edge of
the annulus. The posterior edge of the annulus is rounded and
protuberant and the groove to some extent extends over this pos-
terior face of the annulus and cannot well be all seen at once from
a ventral point of view. While this groove in the shell differs in
different specimens it is in no way an artificial result of methods

126

E. A. ANDREWS.

of preservation but is based upon a special arrangement of the epi-
dermis, which, Fig. 7 enlarged two hundred diameters, is a single
layer of polygonal cells, close under the cuticle. This layer dips
down under the groove and on the side of the groove, where seen
in profile, the cells are elongated and continued as fine fibers that
connect with connective tissue cells and with a membrane sepa-
rating the epidermis from the large blood sinus just above it.
In the figure these cells are represented in black as seen in opti-
cal section and also represented as seen in surface view upon
another focus. Posteriorly the epidermis had shrunken away from
the cuticle, in preparation, and left a wide artificial space. In an
actual cross section of this region the groove is found to be just
below the posterior part of the ganglion, Fig. 8, that supplies the
fifth pair of legs, so that one might argue that the annulus be-
longed to the last thoracic somite while lying in the penulti-
mate one. As seen in Fig. 8, which is enlarged two hundred

FIG. 8.

diameters, the epidermis cells fit together as polygons only at their
outer ends next the shell while their inner ends taper off as fibers
which diverge right and left of the central part of the groove.
The epidermal cells are thus stretched out laterally and arranged
on each side with reference to the superficial median groove that
is to become the sperm receptacle.

Enclosing the epidermal cells is a coagulum of blood partly
separated by a membrane from the blood space beneath the nerve
cord. In this space lies the large median ventral artery that
anterior to this section connects through the nerve cord with the

ONTOGENY OF THE ANNULUS VENTRALIS. I 2/

descending artery from the heart, between the fourth and fifth
ventral thoracic ganglia.

Young crayfish in the above third stage are about 7—8 mm.
long. In the fourth stage they are usually 1 1 mm. long. By
this time the males and females differ noticeably in the lengths of
the first pleopods. In the female, Fig. 9, they are still very small
but larger than in the previous stage, this figure being enlarged

FIG. 10.

but thirteen, and Fig. 5, fifty diameters. In the male, Fig. 10,
on the same scale, the pleopods are long, slender, but simple
cylinders pointing toward one another. These two figures also
show the increasing longitudinal diameter of the anmulus in the
female and its simple form in the male, as well as the reproductive
openings upon the fifth legs in the male and the third legs in the
female.

In these females 1 1 mm. long the median groove of the annulus
is much more evident than at first, but is still a simple groove as
shown in Fig. 1 1 enlarged 200 diameters. The walls of the folded
in and thickening shell that bound the deep and narrow groove
are indicated by the broken line. In this preparation the epider-
mis had shrunken far away from the shell and is indicated in
optical section to show its invaginated state where it came under
the cuticular groove. The groove seems to grow forward from
the posterior face of the annulus and becomes more narrow and
closed in anteriorly. At the anterior tip the groove was partly
overhung by a cross fold, or tended to burrow under the surface
as a short citl de sac.

In different females, however, the groove had different lengths
and more or less of this covering-over of the front end. In a
specimen 10 mm. long the groove was longer and more narrow
than the one above figured. In others of 1 1 mm. the anterior

128 E. A. ANDREWS.

end was crossed by a sharp fold and the cut de sac was more pro-
nounced. In some the posterior face of the annulus, which pro-
jected and overhung considerably, bore a curved transverse ridge
strongly suggesting the posterior rim of the adult annulus, Fig. I.
In a large female of 13 mm. the groove was as in Fig. 1 1, though
a little longer.

The single layer of epidermal cells under the shell was still
visible as a series of polygons as in the third stage. While the
annulus itself showed no setae upon it in any stage the transverse
ridge of the sternum behind it bore scattered over it some four
or five sharp seta in the third stage, about ten in the fourth and
twenty in the fifth.

The fifth stage includes individuals 15-18 mm. long, and in
them marked changes had taken place in the annulus, but before
describing these we will refer to a few observations made upon the
developing annulus in an early stage of another species, C.
Clarkii, from New Orleans. Here also the young reared in the
laboratory had no annulus in the first and second stages, which
were about five and six millimeters long, and they also showed
no external sexual differences. Young of this species were seen
by Hagen.1 when .3 inch long and still on the abdomen of the
female ; in which the females had no pleopods upon the first ab-
dominal somite, but showed the openings of the oviducts, while
the males had the first pleopods as little knobs longer than
broad and turning inward. Later Faxon - stated that the young
of this species from under the abdomen of the parent had the
general form of the adult when 7 mm. long. Evidently both
authors saw the third stage only.

Eggs hatched April 21 reached the third stage June I, but no
observations were made upon the annulus till they had passed
into a later stage, June 19, and were 1 1 mm. long. In this condi-
tion the male and female were much like specimens of C. affinis
of that length, but the first pleopods in the female were much
smaller and in the male noticeably shorter than in C. affinis.

The groove on the annulus of the female now formed a deep,
closed in cavity, Fig. 12, which is enlarged to the same extent as

1 "Monograph N. A. Astacidte," Harvard, 1870.

2 " Revision of the Astacidje," Harvard, 1885.

ONTOGENY OF THE ANNULUS VENTRALIS. I 29

Fig. ii. Iii C. Clarkii the groove has closed, leaving a suture,
represented by the black line, above an internal cavity, repre-
sented in white and this is bounded on the sides and bottom by
the infolded cuticle, represented in dotted lines. At the posterior
end the groove passes around the edge of the annulus and seems

I:;

ft V>
ifti i. ,,.

1%

JJ

'' //x7/

mi

FIG. 12. FIG. 13.

to be still open there. At the front end the groove pushed for-
ward as a blind growth beneath the surface. Thus in C. Clarkii
the receptacle though probably formed at the same stage and in
the same way as in C. affinis, yet advances more rapidly, so that
in the young 1 1 mm. long it is comparable, in its closed up con-
dition, to C. affinis when 21 mm. long and in the sixth stage.

Returning to C. affinis in the fifth stage, 15-18 mm. long, the
general appearance, Fig. 13, enlarged 13 diameters, shows an
increase in size of the animal and a great growth of the first
pleopods of the female which are now much longer than in the
female of the fourth stage and nearly as long as the male pleopods
of that stage, Fig. 10.

The receptacle itself is quite diverse in different individuals. In
all there is added onto the median groove two elevations or
folds which tend to cover over the anterior end of the groove.
The groove itself is bent on one side more or less and the over-
hanging folds are more or less developed. In the specimen
shown in Fig. 14 as enlarged 200 diameters the groove bends to
one side and passes in under a marked flap or fold that grows
over the tip of the groove. This lateral fold may be called the
" hood " to distinguish it from the longer fold which passes along
the opposite side of the groove and tends to overhang it.

130

E. A. ANDREWS.

This second fold is slightly developed in this case. In another
specimen, Fig. 15, the groove bent very far to one side and the
anterior part was concealed under the second fold. We will call
this second fold the "transverse" fold, since it ultimately lies

FIG. 14.

FIG. 15.

more nearly across the median plane. Both folds are oblique,
the "hood" and the "transverse" fold being about at right
angles where the latter passes under the former. In Fig. 1 5
not only is much of the groove overhung by the transverse fold,
but the entrance to the groove in under that fold is cut off, near
the hood, by a short posterior fold that runs parallel to the
transverse fold as it emerges from under the hood.

The cases of unusual bending of the groove to one side sug-
gest the state of things found in the adults of *C. immunis and
C. Bartoni where the receptacle is more transversely placed than
in C. affinis.

The specimen last figured was about to shed and the delicate
new shell within the one here figured was more like that figured
below for the next stage.

In another case the shell cast off by a larva going into the
sixth stage had the groove but slightly bent, Fig. 16, and the

FIG. 16.

FIG. 17.

transverse fold was less oblique. As there was no fold opposing
the transverse fold an object could have passed under its pos-

ONTOGENY OF THE ANNULUS VENTRALIS. 131

terior edge and thence to the anterior part of the groove. The
receptacle was now being formed of three different parts : the
original median groove, the two oblique folds. The former re-
mains as the posterior part of the adult sperm tube and the latter
help to make the orifice and the vestibule, or anterior part of the
adult receptacle.

In all these cases the groove bends more or less to one side,
and in only one case observed was the bending to the observer's
right, which is the left of the animal. In this exceptional animal,
Fig. 17, not only does the groove bend to the right, but the hood
is on the right and the transverse fold and the posterior opposing
fold run at right angles to their usual course. Comparison of
Fig. 17 with Fig. 15, shows that they are, in the main, mirror-
images of one another. Each reversed, as seen through the
paper, would have the symmetry of the other.

This reversal of symmetry in the receptacles of some young is
the first visible expression of the peculiar dimorphism of the an-
nulus in the adults of this and other species. While many of the
adults have the orifice upon the right side of the median plane
others have it upon the left and in all respects these two forms
of annulus are mirror-images of one another. Both forms are
used as sperm receptacles. Those with the orifice upon the right,
Fig. i, are more common ; in one lot of 41 females, 38 had
these right-handed annuli.

The characteristic tuberosities of the adult grow out in later
larval life and they also are right- and left-handed in the following
way : while, usually, as in the above figure, the tuberosity upon
the left of the animal sends a ridge under the right tuberosity,
in left-handed adult annuli the behavior of the tuberosities is the
reverse. In the production of the two symmetrical adult forms
there is thus the harmonious development of the groove, folds
and tuberosities at different periods and from several areas of the
epidermis.

In July the larva may pass into a sixth stage 2 1 mm. long.
On the ventral side, Fig. 18, enlarged thirteen diameters, there is
a noticeable increase in the longitudinal diameter of the annulus
and in the length of the first pleopods which are now longer than
those of the male of stage four, Fig. 10, though closely applied

132

E. A. ANDREWS.

along the sternum and, apparently, of no use as yet. The annulus,

however, has acquired a rather ma-
ture appearance. When enlarged
fifty diameters, Fig. 19, it shows
I i toward the center two gentle eleva-

\\( ^* tions to represent the future tuberosi-
" ties and posteriorly a transverse curved
rim like that of the adult. The whole
plate is still narrow from before back
but much less so than in previous
stages. The suture of the receptacle
is indicated by the zig-zag line and its
lateral walls by the broken lines. The

former represents : anteriorly on the left the former edge of the
hood ; posteriorly the closed up groove ; and in its middle course
the transverse fold.

Enlarged 200 diameters, Figs. 20 and 21, the receptacle
presents varieties of development in different individuals. The
groove itself may be open posteriorly as in Fig. 20 or quite closed

FIG. 18.

FIG. 19.

up as in Fig. 2 1 . The anterior part of the groove bends far to
one side to end under the hood at the right end of the transverse
fold. This part of the groove may be still accessible by passing
in under the transverse fold, Fig. 20, or it may be quite shut off
by a posterior fold growing along against the posterior edge of
the transverse fold as in Fig. 21. Fig. 20 is not much advanced
beyond the preceding stage, Figs. I 5 and 16, the open groove, the
hood and the transverse fold are larger and the invaginated shell
is more extensive, as indicated by the broken lines. But in other
cases, Fig. 21, the deeply buried groove connects with the sur-
face only by oblique planes that come to the surface as the three

ONTOGENY OF THE ANNULUS VENTRALIS.

133

successive parts of a zig-zag suture. The inner course of the
groove is more sinuous than before in correspondence with the
increased bending of the invagi-
nated shell that forms its walls. /

/

The chief features of the adult ',

\

sperm pocket are thus present \
except that the anterior orifice
seems more closed in some indi-
viduals.

Though in some cases itseemcd
as if the groove itself did not
bend to one side to end under the
hood, but rather that the trans- -___
verse fold left a passage from "
the groove to the hood, the true
state of the epidermal groove upon and in which the shell groove
is formed was seen when the shell was pulled off. Then, in Fig.
22, enlarged 200 diameters, there is a deep, bent furrow with high
sides formed of long epidermal cells, as somewhat diagrammati-
cally represented in this camera lucida sketch of an optical section

FIG. 20.

FIG. 21.

FIG. 22.

of the epidermal part of the receptacle. In this specimen the shell
was easily removed since a new shell was in process of formation to

134

E. A. ANDREWS.

line the groove. This made the groove more distinct than other-
wise is the case. The groove passes forward from the posterior
edge of the annulus and then bends to one side much as does the
bottom of the shell groove seen as a shaded line in Fig. 21. In
each individual, however, the amount of bending seems to be
different.

Such a bent groove is very like the simple bent epidermal
groove of C. Clarkii, which is one of the less specialized species.
While previous experience had shown that larvae after living
eleven days in the above sixth stage might turn into a seventh
stage 29 mm. long in the middle of July, no observations were
made upon the above larvae 21 mm. long till October 3, 1904,
when they had turned into individuals 25—53 mm- long. Those

25-35 mm- agreed with one an-
other in the development of the
annulus and probably represented
larvae that from lack of enough
food had remained in the seventh
stage while the female, 53 mm.
long, belonged to some later
stage. In this laboratory it has
been found that crayfish of that
latter size have their sexual in-
stincts developed in the fall and
females 52 and 53 mm. long, being
x^_ then supplied with sperm by
males of the same or even smaller

FIG. 23.

sizes, laid good eggs the follow-
ing spring when not quite one
year old. Thus the above female was probably mature in instincts
and in external sexual organs when examined.

First, however, taking up the specimens 25-35 mm- long which
were all essentially alike except in size, one 35 mm. long enlarged
13 diameters, Fig. 23, presents a noticeable growth of the first
pleopods which are now turned forward and sparsely set with
hairs as in the adult, and also much increase in the relative im-
portance of the curved ridge bounding the mouth of the oviduct
opening on the third leg. As an opaque object the annulus now

ONTOGENY OF THE ANNULUS VENTRALIS.

135

showed very definite tuberosities separate from the hood and the
transverse fold and the tuberosity upon one side of the animal
extended over the median line, in carrying out the asymmetry of
the adult. The general proportions of the various folds and
thickenings of the shell now approximated the adult condition.

The receptacle had not materially changed from the previous
stage, Fig. 21, but its invaginated shell walls were greatly thick-
ened and laminated, Fig. 24, enlarged like Fig. 21. It would
appear that with the dropping of the bottom of the invaginated

— cx^

" >^: ^vr. r _-L~ —
<J*^~: •*•- - -Ib>~ -=• •-

^^^S^^x

v x\\x >t->-i «:-->5>$ov v

fV, X^ -'- - X ° -"»>>>

' V<^3S1^S^X\^^ N^

'^-x\"xA;sV'^\

FIG. 24.

groove away from the surface its sides have closed in to form a
narrow crevice which comes to the surface as a suture line. Then
by the bending of the bottom of the invagination more than its
surface suture line these crevices are made into curved oblique
planes. The hood and the transverse fold and the posterior
opposing fold are intimately associated with the bending of the
original groove to one side and then the sinking of that lateral
bend away from the surface and diagonally forward.

The last young to be considered, the larva, 53 mm. in length

136 - E. A. ANDREWS.

and sexually mature, has an annulus very much like that of the
full-grown adults, but it is still very small, as is seen on com-
paring Fig. 25 with Fig. i, both enlarged 12 diameters. Still
the tuberosities do not yet overhang the transverse depression
enough to conceal the hood and the transverse fold, nor are the

transverse depression and the pos-
terior rim as well developed as later.
The receptacle itself is much like
that in Fig. 24, but its orifice is more
perfect, though not yet as patent as
in the large specimens, Fig. i.
This young annulus still lacks the

posterior enlargement of the sperm tube, indicated in Fig. I, and
the complexity of bending of the tube is less, nevertheless it func-
tions as a sperm receptacle.

SUMMARY.

The specialized sternal plate of the shell of Cainbarns, which
in the adult female bears the sperm-receptacle, is first differen-
tiated in the third larval stage, in C. affinis.

The epidermis under this plate grows inward and outward in
special areas to form the receptacle. First there is formed an
open median groove that then bends to one side, sinks away from
the surface and becomes closed. In C. Clarkii also one stage
shows a like development.

Next definite folds overgrow the anterior end of the groove as
it sinks from the surface.

Later elevations and depressions complete the external sculp-
turing of the annulus.

The anterior part of the groove, with the accompanying folds,
form the orifice and vestibule of the sperm receptacle, while the
posterior part of the groove forms that part of the receptacle in
which the sperm is stored.

A functionally complete annulus is made within five months,
but subsequently it becomes more complex. Comparative study
of the adults of several species has shown that in Canibarus the
essential part of the annulus, as a sperm receptacle, is a curved
pocket ; the above facts indicate that in Canibanis in general this
pocket arises as an open epidermal groove.

ONTOGENY OF THE ANNULUS VENTRALIS. 137

The right or left handed symmetry of different adult annuli is
first visible in the fifth larval stage when the median groove has
bent to the right or to the left and the accompanying folds have
a reversed position in the two cases. Later other outgrowths
may also harmonize with the groove and folds to complete the
two adult forms ; mirror-images of one another.

While the sperm receptacle is a necessary organ in Cainbants
it is phylogenetically a new one since Canibarns is the most
specialized genus of crayfish and other genera have no receptacle.
In C. affinis this new organ is not seen till the individual has
reached a third stage after leaving the egg.

Besides beginning late in ontogeny this new sperm receptacle
is variable in all its stages of growth. In its early and simple
state in the larva it resembles the adult receptacle of a less spec-
ialized species, C. Clarkii.

In some of its variations of lateral bendings in early larval
stages it suggests adult conditions in other species.

BALTIMORE,

December 12, 1905.

Vol. X. MarcJi, 1906. No. ./

BIOLOGICAL BULLETIN

CORRELATION IN THE DEVELOPMENT OF
FASCIOLARIA.1

O. C. GLASER.

The following pages contain in abstract an account of the
embryology of the prosobranch Fascio!aria tnlipa (var. distans),
and a discussion of such occurrences in its development as seem
to me to be of general interest.

I. THE CANNIBALISM OF FASCIOLARIA EMBRYOS.

The breeding season of Fasciolaria at Beaufort, N. C., lasts
from about the first of May until the first of July, although egg-
cases containing various stages of development were occasionally
found in August. The capsules, held together by a basement
membrane fastened to oyster shells, conchs or wharf-piling, occur
in bunches of I 5-30 varying in size from one half to three inches
across.

When fresh they are soft and so translucent that the pink or
white eggs suspended in their albuminous contents give their color
to the whole mass ; later, however, the egg-cases become firm and
elastic, and obscured by the algae, polyzoons, and other organ-
isms which grow over them. Often isolated capsules, or bunches
containing only a few are found. These are produced either by
females interrupted at laying, or by young females, which usually
deposit fewer and smaller capsules than the old ones. The last
capsule to be laid in a bunch, whether deposited by an old or

1 For the privilege of collecting the material on which this work was done I am
indebted to the Hon. George M. Bowers, U. S. Commissioner of Fisheries. The
preparation of this paper was begun during my tenure of the Adam T. Bruce fellow-
ship in the Johns Hopkins University, and was finished for the press in the Zoological
Laboratory of the University of Michigan.

139

140

O. C. GLASEK.

young female, often lacks either the upper or lateral flanges, or
both.

The number of eggs in each capsule is much greater than has
been supposed, and has an important effect on development. By
actual count I found that one capsule contained 2,308 eggs.
The highest estimate which I remember to have seen in the lit-

o

erature is from 600 to 800. The ova, densely crowded with
pink, brown, or white yolk spherules, which are separated from
an unusually large and eccentrically placed germinal vesicle by

FIG. I. Egg-cases of Fasciolaria tulipa (var. distant} one half natural size.

a zone of clear cytoplasm, vary in diameter from .17 to .25 mm.
Even eggs of this size are minute enough to produce the same
optical effects that much smaller spheres do, for in making a
numerical estimate in a watch crystal one almost invariably sup-
poses fewer to be present than the dish actally contains.

Of all the eggs in fresh capsules only a few are fertilized and
develop. The remainder with the germinal vesicle intact are in-
gested, usually within a week after deposition by the developing
embryos. After the eggs have been swallowed several days
their germinal vesicles fragment, their pellicles disappear, and
their yolk finally is digested.

The fertilized eggs are so few in number and so difficult to find

THE DEVELOPMENT OF FASCIOLAKIA. 141

that the only method of determining the average number per
capsule was by finding the average number of larvae in the later
stages. The number of fertilized eggs in each capsule cannot
have been less than the number of larva; that come from it.
There are reasons moreover, such as the frequency of accidents
and the occurrence of dwarfed embryos, easily overlooked,
for believing that the number of fertilized eggs exceeds the
number of young found in later stages. The contents of 145
advanced capsules showed 6.2 larvae in each as an average
between a maximum of I 5 and a minimum of 2. Between these
extremes all degrees of fertility occurred.

The stages of development found in a single capsule are as
variable as the number of larvae. Two cell stages may occur in
the same case with advanced embryos. It is difficult therefore
to form an idea of the rate at which an individual egg develops,
particularly as development ceases shortly after the embryos are
removed from the capsules. Judging from the great variety of
stages found in a single egg-case one may conclude either that
not all the eggs are fertilized in the oviduct, but that some are
impregnated after the capsules have been deposited, or that for
some reason certain ones undergo a longer resting period than
others. The importance of these discrepancies at the beginning
of development becomes apparent in later stages.

After the early developmental processes have been passed
through a larva results so irregular that no two individuals of

o *— *

this age are alike. Fig. 2 omits two eggs which this larva had
swallowed, but will serve very well without these to give an
idea of the general external appearance of the embryo before the
crisis of its larval life has occurred. The larva is represented as
viewed from the ventral surface. Anteriorly is the head vesicle
(/i.i'.] and posteriorly the body vesicle (£.?'.). Between the two,
under the right external kidney (ex.k.\ two of the yolk spheres,
derived from the four macromeres of the segmentation period,
can be seen. The ectoderm of the body and head vesicles lacks
definite cell boundaries, though the indefiniteness is much greater
in the anterior than in the posterior region. Each nucleus of the
ectoderm stains deeply and is surrounded by vacuoles which
decrease in size inversely as their distance from it. Those

142

O. C. GLASEK.

furthest away from the nucleus are the smallest, and finally only
minute scattered granules can be seen where one cell abuts upon
another.

The mouth (;«///.) is nearly perpendicular to the antero-poste-
rior axis of the body. On each side of it are two pear-shaped
patches of highly vacuolated ectoderm [which I described ('04)
as early stages of the external kidneys, two organs which in

hv

m t k

i
bv

I' ic,. 2. Pre-cannibal larva of Fasciolaria ; b.v., body vesicle; ejc./c., external
kidney; h.v., head vesicle; mac., macromere ; /nt/i., mouth. Zeiss D obj. 2 oc.
Drawn by Mr. Carl Kellner.

Fasciolaria reach surprising proportions and probably take a most
active part in the excretion of waste products.

The cells of the excretory organs differ from the ordinary
ectoderm cells chiefly in [the great size of their vacuoles and
nuclei. They resemble at this stage the ordinary ectoderm in
in the comparative indefiniteness of their boundaries, which
curiously enough are perfectly distinct at earlier as well as at
later stages of development. This disappearance and subsequent
reappearance of cell boundaries is due I believe to changes in

THE DEVELOPMENT OF FASCIOLARIA. 143

the intra-cellular pressure, which varies with the size and number
of the vacuoles ('05).

When the larvae have attained the external condition described,
the process of cannibalism begins. The eggs which up to that
time are quite uniformly scattered throughout the albumen of
the capsules are collected in more or less dense groups by the
action of the cilia around the mouths of the embryos. Some
hours after the eggs have been collected in the center of each
capsule smaller groups, still more densely packed, can be seen,
and at the end of two or three days all have been gathered into
from two to fifteen spheres which are the cannibals, stuffed almost
beyond comprehension. All larvae which secure sufficient eggs
finally have the appearance of the one shown in Fig. 3.

FIG. 3. Fully gorged cannibal larva of Fasciolaria, //.?'., head vesicle. En-
larged 20 diameters. Drawn by Mr. Carl Kellner.

This embryo is a fully gorged cannibal, so distended with eggs
that the body wall is scarcely visible, and the mouth and excre-
tory bodies are quite obscured by the dark background of yolk.
All that can be definitely made out is the folded and slightly
irregular head vesicle (Ii.v.') which marks the anterior end.
Larvae in this condition are comparatively regular, being in fact
nothing more than balls of eggs held together by exceedingly
thin transparent membranes. The diameters of fully gorged
cannibals vary from 1.16 mm. to 1.90 mm. depending on the
actual number of eggs that have been taken in. The average
diameter of larvae in this condition is about 1.48 mm. and one

144 O. C. GLASEK.

near this average contained by actual count over 300 eggs.
Some larvae of larger size undoubtedly contain more eggs, and
some are entirely devoid of them.

What regulates the number of eggs swallowed? As already
noted, eggs in all stages of segmentation may occur in the same
capsule. Similarly in older capsules larvae of very different de-
grees of development may occur together. Of these the most
advanced, the first to be ready to swallow, gets most of the eggs,
while its tardy mess-mates must take what remains. Those
that are very late in reaching the point where they can ingest
eggs often get none at all. They remain dwarfed and subsist
on the jelly as best they can. Most of these dwarfed larvae de-
generate or are ingested by the cannibals, but occasionally some
hatch as " runts " which are normal in all respects except size.

Of the larvae which develop at a uniform rate those which have
the most distensible mouths and the most violent ciliary action
in the adoral field get more eggs than those with less distensible
mouths and a less active ciliary mechanism. Thus in general,
the number of eggs which a given larva secures depends on how
early it enters the period of cannibalism and how rapidly and
easily it swallows.

Repeated experiments were tried to find out the effect of arti-
ficially increasing the food supply of a given larva. One of these
experiments gave a very remarkable result.

I wras not able to keep the larvae in a healthy condition outside
of the egg cases, nor to satisfactorily reseal a capsule once opened.
The problem of artificially increasing the food supply of a given
embryo offered some difficulties until I tried the plan of injuring
some of the larva? inside of a capsule. This can be done easily
by allowing them to drift into a corner of the egg case, holding
them there by pressure on the rest of the capsule and then com-
pressing them with a pair of strong forceps. With care one or
two larvae can be kept out of the corner, and all those which have
been coralled can be injured in the manner described.

An injury to the excessively thin body-wall has the same effect
as a hole in a bag of grain. The eggs which have been swal-
lowed roll out and leave an empty membrane behind. In this
way a thousand eggs, which have been swallowed once, may be

THE DEVELOPMENT OF FASCIOLARJA. 145

set free. In the experiments which I made all but two or three
larva: in each of a large number of capsules were forcibly in-
jured, and their undigested contents offered to the uninjured,
larva:. These, although fully distended, in every case began to
devour the additional eggs offered them, and in three or four
days had mastered twice as man}' as had before fallen to their
respective lots. One experiment, which was particularly suc-
cessful, consisted in an attempt on my part to ascertain whether
one larva would eat all the eggs which had been taken in by its
competitors. This was accomplished by first injuring two of
these and waiting until the extra number of eggs had been dis-
posed of. Then two of the remaining larvae were injured and the
eggs which they contained --some of which had been swallowed
once and others twice -- were offered to the three uninjured in-
habitants of the capsule. After this second offering of eggs was
disposed of, two of the remaining three larvae, which had grown
to more than double their original size, were injured, and their
contents offered to the sole survivor. This individual, already
excessively gorged, began to ingest the eggs which had been
swallowed from one to three times ; unfortunately the elasticity
of its body was not equal to the undertaking and the larva burst
from over-eating before it had finished.

Over-feeding comparable to what happened in these experi-
ments may take place without human interference. It is easy to
see how the natural disturbances to which the capsules are sub-
ject might be sufficiently violent to burst some of the larvae whose
contents then would be devoured by the survivors. I am con-
vinced that this actually takes place, since the larvae in those
capsules which contain only two or three are always much larger
than the average. From these observations and experiments I
believe that the number of eggs secured depends on promptness
and structural aptitude for seizing and swallowing.

II. THE EXTERNAL AND THE ACCESSORY EXTERNAL

KIDNEYS OF FASCIOLARIA.

Develop incut.— - The external kidneys, also known as "excre-
tory cells" (Conklin) or " subvelar masses" (Osbore) and very
generally called " Urnieren " by German writers, originate early

146

O. C. GLASKK.

in Fasciolaria, and are remarkable not for their size and posi-
tion on the embryo, but also because they seem to take a very
important part in the economy of the larva. Both the presence
and the activity of these remarkable bodies are of interest when
considered in the light of the feeding habits just described, for
these, as might be expected, tax to the utmost the powers of
assimilation and excretion.

Without considering the cytological changes which occur in the
kidney cells, and which have been described fully in another
place ('05) I will mention some of the more important alterations
which involve the entire organs. What Conklin ('97) has said of
the excretory cells of Crepidula applies equally well to those of
Fasciolaria : "In the early stages these cells form a part of the

FIG. 4. Vertical section through external kidney of F,isciolnria showing connec-
tion with unmodified ectoderm and the plug (f>lg.) of undifferentiated cells which

later fill the lumen.

y obj. 8 oc.
Leitz " — . Drawn by Mr. Larl Kellner.

ectodermic layer, but as the embryo grows older they grow more
prominent, and the whole mass is constricted at the base, so that
it becomes pear-shaped, the narrow end being attached to the
embryo, and the larger end being distal. The surrounding ecto-
derm cells crowd in at the neck of this constriction and work
their way entirely beneath these excretory cells."

In F'asciolaria three very important changes take place in addi-
tion to those already mentioned. The ectoderm cells which

THE DEVELOPMENT OF FASCIOLAKIA.

H7

" crowd in at the neck of the constriction " do'more than this, for
some of them coming from opposite sides join to form a plug
which projects into the hollow of the rounded excretory organ.
(Fig. 4, />/£•.) As the excretory organs at later stages have no
lumens whatever, I conclude that these cells become modified
secondarily into excretory cells like the primary ones.

The second important change which takes place in the external
kidneys involves their position and results in the adoption of that
curious relation with the velum which led Osborn ('85) to speak

ExK

ExR

Fi<;. 5. Four stages in the migration of the external kidneys (ex.k. ) from the
lateral surface of the cannibal to their final subvelar position on the veliger.

of them as " subvelar masses." When this change beeins the

o o

embryo as a whole has enlarged greatly (Fig. 5), and has both
by growth and by the stretching, due to the ingestion of eggs,
become much wider than it was, especially at the base of the
head vesicle. This increase in size brings about the removal of
the external kidneys from the anterior suiface of the larva to the
lateral surfaces. Later the great activity in the velar ridge, by
which this becomes prominent, as well as the growth of the whole
anterior end of the larva, result in lifting the external kidneys
upward and carrying them laterally away from the central mass
of yolk on which at first they lie directly. Further lateral growth

148

O. C. GLASER.

of the velum carries the external kidneys outward and still
farther away from the embryo, until finally they are far off at the
sides of the body and hang down from the under side of the
velum (Fig. 5). In this position the organs, which are white,
viscid masses of ovate shape, having at the height of functional
activity an average width of .7 mm., a length of i.o mm., and a
depth of .9 mm., remain throughout the life of the veliger until
this begins to assume the adult form. When that stage is reached
the external kidneys begin to decrease in size, to degenerate, and
the nuclei of the cells to disappear. Finally, in most cases, the
kidneys are dropped off before the velum is resorbed. There
are exceptions to this however, for often a much more intimate

si^s^iPtei^s

FIG. 6. Vertical section through external kidney showing the intimate relation
which it mav have with the velum (vel. }. Leitz J-

2

relation exists between the velum and the excretory organs hang-
ing down from its under surface. This greater intimacy is well
illustrated by Fig. 6, which is a typical section of these cases.
In this organ, after the original lumen had been filled by the plug
cells, the fully developed velum itself grew into the small pit
which remained at the proximal (final anterior) end of the
kidney. In this instance the velum was so crowded in the
remaining hollow of the external kidney that the velar cavity
itself was entirely obliterated by the cloj-e approximation of its
upper and lower walls. In cases of this kind the velum drops
off with the external kidney when the veliger assumes the adult
form.

THE DEVELOPMENT OF FASC1OI.ARI A.

149

This intimate connection between the velum and the kidneys
recalls the " ans.ne " of BitJiynia tcntaculata, described by Sarasin
('83). In this form the relation between the external kidneys and
the velum is still more intimate than in these exceptional cases
in Fasdolaria. Indeed the velum of BitJiynia seems to be
modified into a functional, excretory apparatus, without having
renounced its original duties as an organ of locomotion.

. \rccssory l-.xtcrnal Kidneys. - - One of the most unexpected

FIG. 7. Surface view of a portion of a functionally active external kidney, show-
ing the thickened cell boundaries, the vacuolated cell contents, the amitotically
dividing nuclei, and the nucleoli with their surrounding clear areas. Zeiss A obj. 4 oc.

facts which I have encountered in my study of Fasdolaria
embryos is that in addition to the large excretory organs no
small proportion of larvae possesses, sometimes in surprising
places, accessory multicellular or unicellular external kidneys.
In many cases these are found on the underside of the velum just
behind the great kidneys ; here they may become a third as large
as the primary organs. In other instances single cells of the
postoral or preoral velar row, and even of the head vesicle
become modified into secondary unicellular kidneys. In some

I5O O. C. GLASER.

of these accessory organs amitosis was observed, and in all the
same extra-nuclear signs of activity which characterize the cells
of the primary kidneys.

Function. - - Between the two extremes of development which
I have described the nuclei of the external kidneys divide ami-
totically and the cells become polynucleated. The nuclei are
large, granular, very irregular in outline, and each one has at
least one nucleolus surrounded by a clear area. In addition the
cells are characterized by their thick boundaries and the highly
vacuolated condition of their contents.

On a former occasion ('04) I showed how the vacuoles in the
cells might be traced to the halos surrounding the nucleoli, but
the figures intended to illustrate this are unsatisfactory since
certain features which I hoped would not appear so unnaturally
prominent when reduced to one half are over emphasized. More
satisfactory figures will be found in my later paper ('05). The
vacuoles which in some cases can be traced directly to the extra-
nucleolar halo, in others to the nucleus from which I have seen
them escaping, certainly suggest a high state of metabolic activity.

In the summer of 1903, with a view to determining the nature
of the activities in the external kidneys, I removed several hun-
dred, extracted them in chloroform water and asked my father to
make a careful analysis • of the extract. The details of this
analysis have already been published ('05) but as all the decimal
points were omitted by a careless printer, I republish the follow-
ing corrected summary. One liter of the aqueous extract of the
external kidneys contained

N as albumen, .1242

N as free ammonia, .0099

N as urea or homologues, -2163

Total N, -35°4

My interpretation of the above analysis was that the external
kidneys excreted waste products, but a reconsideration of the
evidence, shows that it can become decisive only after corre-
sponding analyses of extracts of other tissues have been made.
Regardless of the value of the chemical evidence, which is cer-
tainly not negative, I still think that all the morphological results
point to the probability that the external kidneys are excretory

organs.

THE DEVELOPMENT OF FASCIOLAKIA. 15!

With the aid of such evidence as I have been able to obtain,
and Osborn's interesting discovery ('04) of amitosis in the endo-
derm, I think I may formulate a scheme to account for the
transfer of materials which probably takes place. Osborn ('04)
announced that he had discovered amitosis in the endoderm of
the gastrula^ of Fasciolnria and associated, very justly, I think,
these nuclear phenomena with digestion. I have been able to
verify these results, not, however, until I had been thoroughly
led astray by the assertion that the nuclear divisions in question
occur in the gastrula stage. All my earlier larvae show no signs
either of the " cuboidal endoderm" or of the amitoses of which
Osborn speaks. The latter phenomenon I finally discovered in a
larva that had a velum, and was therefore well past the period of
cannibalism. Fig. 8 is compounded from several sections ob-

vei

op cp.

FIG. 8. Composite figure made up from several adjacent sections oblique to the
antero-posterior axis. The external kidneys, not yet subvelar in position, are cut
horizontally. Opsp., optic cup. vel. , velum. The arrows are intended to indicate

the probable paths of waste products reaching the exterior from the digestive tract

? obj. 3 oc.
seen in transverse section, via the external kidneys. Leitz

lique to the antero-posterior axis of this larva. The oesophagus
is cut transversely and the cells which compose it are dorsally,
large, polynucleated, and highly granular, with vacuoles at their
distal ends. The nuclei appear to have arisen amitotically.
Some of these cells have burst and their contents may be seen
oozing out. They bear a striking resemblance to the cells of the
external kidneys.

The cells of the lateral and ventral walls of the cesophagus
have a very different appearance. These, instead of being long

152 O. C. GLASER.

and granular, are very irregular in shape, the cytoplasm being
almost invisible, in irregular strands, or crowded closely against
the cell boundaries. In the center some of the cells have a large
clear space in which the nucleus lies. The nucleoli sometimes
occur in the clear vacuoles which are scattered about in the
nuclei. Many of the latter are in process of amitotic division.

Around the outer margin of the lateral and ventral walls of the
oesophagus many of its bounding cells seem to have broken
down, whereas others are clearly engaged in this process. By
this means the outline is made extremely irregular, being frayed,
and fringed with granules and the fragments of cells. This
breaking down of the cells, their general vacuolated appearance,
the occurrence of amitosis, and the finely divided state of the
chromatin in the nuclei, all indicate that these cells are actively
engaged in metabolism, and the simplest supposition is that they
are engaged, as Osborn has suggested, in the process of digestion.

That something is leaving these cells is demonstrated by their
appearance. I believe therefore that in the course of their activity
as digestive cells, they excrete waste products, and that these
pass across the "body cavity " into the cavity of the velum,
transude through its basal membrane into the external kidneys
and through these ultimately to the exterior. The arrows are
intended to give a graphic representation of this process.

Homologies.— - The earlier writers who considered the functions
of the external kindeys of gastropods fall into two general groups,
those who affirm and those who deny that these organs are renal
organs. The same writers however have been at much greater
discord with regard to the homologies. Thus Salensky ('72)
homologized the external kidneys of Calpytrea sincnsis with the
primitive kidneys of pulmonates. Fol ('75) did the exact reverse
by attempting to homologize the " Urnieren " of pulmonates with
the primitive kidneys of prosobranchs. Bobretsky ('77) who de-
scribes the "Urnieren" of Nassa nnitabilis, Natica, and Fusns,
seems to have accepted the homologies advocated by Salensky.
Biitschli ('77) in the same year, however, in a paper on Pahidina
invipara took strong exception to Fol's homology. He pointed
out that this supposed homology was between totally different
structures, and that while the primitive kidneys of Palndina

THE OKYELOPMENT OF FASC1OLAKIA. 153

might be homologous with the primitive kidneys of Lyiu incus and
Planorbis, the external excretory organs " Urnierenzellen " of
these forms could not be homologized with the primitive kidneys
of Pahtdina.

Thus the significant fact was brought out that there are two
kinds of " Urnieren," the one, since proved to be of mesodermal
origin (Palndina, Planorbis], opening to the exterior by a pore,
the others, collections of modified ectoderm cells, situated either
on the postoral row of the velum (I.yjnnceiis, Planorbis], or be-
hind the velum (Orpidnla, Fn/gitr]. The confusion due to the
attempts to homologize the external kidneys of prosobranchs
with the primitive kidneys of pulmonates, and the converse prop-
osition of tracing an homology between the external kidneys of
pulmonates and the primitive or head kidneys of prosobranchs,
was probably greatly increased because all these structures were
called by some German writers "Urnieren." In this connection
it is interesting that Korschelt and Heider ('93) in copying Bob-
retsky's figures for their text-book changed the labelling so that
" Un " is replaced by " Ex."

Rabl ('79) came to the conclusion that the external kidneys of
Planorbis were a part of the velum, and that they have nothing to
do with the " head kidneys." Sarasin ('82) describes a most in-
timate connection between the velum and the external kidneys
in Bithynia tentaculata. In fact he described the two under one
name, the " Ansae." In speaking of their homologies he says :
" Nach den Erfahrungen von Bobretsky, Biitschli und mir, liegt
auf jeder seite der Prosobranchier Embryonen, ein Haufchen
grosser Ectodermzellen, das bei Paludina und Bithynia mit Wim-
feroffnung nach aussen miindet. Nach Biitschli und Fol finden
sich diesselben bei Planorbis. 1st dies richtig, so haben die Siiss-
wasser Pulmonaten zwei Organ paare, die als Urniere zu deuten
sind ein vorderes und ein hinteres Paar. Hat Rabl recht dass
die von Biitschli zuerst gefundengrossen Zellen bei Planorbis und
Lymncens zum Velum gehoren, so sind wahrsheinlich die von
Biitschli und mir bei Pahtdina und Bithynia gefundenen Organe
den hinternen Urnieren der Siisswasser Pulmonaten homolog."
Sarasin's position, except that it admits the ectodermal origin of the
primitive kidneys in Paludina, is in the present state of our knowl-

154 O. C. GLASER.

edge tenable. Erlanger ('92), though denying the excretory
function of the " Ansae " of Bitliyuia, conceded that on mor-
hological grounds these cells might be considered the equivalents
of the peculiar velar cells of the pulmonates and of the marine
prosobranchs.

Heymons ('93) introduced another source of confusion by
attempting to homologize the excretory cells of opisthobranchs
with the external kidneys of prosobranchs. This homology
seems far-fetched, since the excretory cells of Umbrella are situ-
ated near the anus of the larva, though they originate much
further forward. This difference in position seemed of no im-
portance to Heymons, since, as he says, McMurrich had already
shown that the excretory cells may lie at various distances
behind the velum. Conklin ('97) says in reply : "this difference
in position seems to me, however, to be a very considerable one.
In all prosobranchs these cells lie close behind the velum, while
in Umbrella they are removed from that structure by almost the
whole diameter of the embryo. Further, the fact that they sink
into the interior in Umbrella would indicate that they are differ-
ent from the excretory cells of prosobranchs."

Conklin ('97) finds that the "external excretory cells " of
Crepiditla have no connection with the velum. This is also true,
in a sense, of Fasciolaria. In this form the external kidneys
originate long before the velum, so that their ultimate connection
with that organ is not primitive but secondary, and due to their
origin near the place where the velum originates. In this respect
they are not fundamentally a part of the velum, any more than
the external kidneys of Crepidula.

It is obvious from the literature considered so far that the dis-
cussion of homologies has involved at least three different kinds
of excretory organs. In order to emphasize the differences I
shall henceforth call those mesodermal, or possibly ectodermal,
structures of pulmonates and prosobranchs which open to the
exterior through a pore, primitive kidneys ; those modified ecto-
dermal excretory cells which may occur in addition to primitive
kidneys, external kidneys ; and finally such external organs as
Heymons and others have described in opisthobranchs, excretory
cells.

THE DEVELOPMENT OF FASCIOLAKIA. 155

Anyone who has seen Stauffacher's ('98) beautiful figures of
the primitive kidney of the trochophore of Cyclas cornea can never
compare similar structures with external kidneys in prosobranchs
and pulmonates, or excretory cells in opisthobranchs. Meisen-
heimer ('98) introduced a complication however, for his very
complete study of the development of Liinax ina.viunts leaves
almost no doubt that the primitive kidney of this gastropod is of
purely ectodermal origin. The reasons which Stauffacher has
advanced that the primitive kidney of Cyclas is mesodermal, are
as convincing as Meisenheimer's that in Liinax it is ectodermal,
so that it is necessary for the present to subdivide the group of
primitive kidneys into ectodermal primitive kidneys and meso-
dermal primitive kidneys. If the conclusions of Stauffacher and
Meisenheimer be indeed correct, then two sets of homologies can
be granted ; .the mesodermal primitive kidneys of prosobranchs,
pulmonates and lamellibranchs may be homologous ; and the
ectodermal primitive kidneys in the same groups may be homolo-
gous. No homology however can be granted between a primi-
tive kidney of mesodermal and one of ectodermal origin without
doing violence to the whole conception of homologies. Whether
a conception which separates organs as much alike in structure
and probably in function as the ectodermal and mesodermal
primitive kidneys of larval molluscs should be violated, is a ques-"
tion which at present I do not feel able to discuss.

Are the external kidneys homologous with ectodermal primi-
tive kidneys ? I believe that there is no more reason to homol-
ogize external kidneys with ectodermal primitive than with meso-
dermal primitive kidneys, for the differences between external
kidneys and primitive kidneys, of whatever layer, are the same.
That in the one case both organs should originate from the same
germ layer is no criterion on which to base homologies, for if it
were we should be logically driven to homologize not only all
ectodermal structures, but all structures of whatever origin.

In considering the relations of the larval excretory systems of
molluscs it seems to me to be of great importance to keep the
differences which I have tried to emphasize constantly in mind,
but the distinctions once made, the task of recognizing true
homologies is by no means a simple one. Possibly all meso-

I 56 O. C. GLASER.

dermal primitive kidneys are homologous, and possibly all ecto-
dermal primitive kidneys are, but certainly not all external
kidneys. The occurrence of unicellular or multicellular accessory
external kidneys in different regions of embryos already so well
endowed with excretory organs as Fasciolaria shows that the
embryological measuring rod which has been so carefully applied
to these larval structures of molluscs is less accurate than some
of the investigators who have used it for the discovery of alleged
detailed relationships.

III. THE ORIGIN OF THE HABIT OF CANNIBALISM.

Even though there are differences in the early extra-ovarian
histories of the food products, the consumption of eggs and em-
bryos by the developing larvae of Fasciolaria is fundamentally
similar to those other cases among gastropods in which certain
young are used as food after being broken down, or are preyed
upon directly by their competitors. It seems to me justifiable,
therefore, to include all these methods of nutrition, based on the
consumption of materials derived from the ovary of the mother,
but not contained within the eggs from which the consumers
come, under one term, cannibalism. The various degrees to
which cannibalism is developed in different gastropods have been
arranged in series by earlier writers, and though this series is both
interes'ting and instructive it has no phylogenetic significance,
and I shall try to show that the phenomenon in Fasciolaria can
be explained independently.

McMurrich ('85) noticed that some of the " ova" in the capsules
of Crcpidula fornicata, C. plana and C. convexa break down and
are used as food by the survivors, although this process is not so
pronounced as in Purpura floridana. Koren and Danielssen
('57) described the case of Bnccinnni nndatwn which is very sim-
ilar to that of Fasciolaria in the disproportion between the num-
ber of embryos developing in the capsules and the number of
eggs furnished by the female, Buccinum, however, differs from
Fasciolaria because many of the ova, which do not form embryos,
divide. Carpenter in the same year ('57) corrected the view of
Koren and Danielssen that the embryos originate by conglomer-
ation, by describing the development of Purpura lapillus in

THE DEVELOPMENT OF FASCIOLARIA. 157

which from 12 to 30 eggs develop into embryos, whereas each of
the remaining 500 or 600 divides without regularity into from
14 to 20 fragments.

Selenka ('72), who confirmed Carpenter's results, did not con-
sider the division of the " sterile " eggs equivalent to the regular
segmentation of the fertile ones, not only because of the irregu-
larity of the former process, both as to form and occurrence, but
also because he found no nucleus. Neritina fluviatiiis, according
to Blochmann ('Si), has capsules in which all the eggs are pro-
vided with nuclei that take the usual part of the formation of
polar bodies and female pro-nuclei. After these processes no
regularity can be detected in the divisions of the sterile ova, the
later behavior of which led Blochmann to agree with Biitschli
('77) that these eggs are unfertilized.

Brooks ('77) observed that of the 6 to 20 eggs in the capsules
of Urosalpinx cincrca, all undergo development normally, though
exceptionally some may break down and serve as food for the
survivors. This case of exceptional cannibalism furnished McMur-
rich (loc. cit., p. 408) with "a clue to the manner in which the
phenomena seen in Fasciolaria, Pitrpura lapillits, etc., have been
brought about. An occasional egg in a capsule has from some
cause or other broken down, and has been drawn into the diges-
tive cavity of the developing embryos. This process having
proved useful is continued, and an arrangement such as I have
described above for Purpnra floridana obtained. From this it is
but a step to what occurs in Bnccimnn, Pitrpitra lapilhis and Ner-
itina. In Fasciolaria the process is, as far as we know at present,
at its culmination."

This series was not conceived of as genetic, for this would have
been justified only if it had been previously shown that each of
the forms mentioned was evolved from one next lower in the scale
of cannibalism. That such has been the history of cannibalistic
prosobranchs is not supported by any evidence known to me,
and the series is therefore probably a collection of graded par-
allelisms.

The real clue seems to me to lie in another factor which, as
McMurrich himself pointed out, is invariably present where can-
nibalism has advanced to any great extent. " This is (loc. cit.,

158 O. C. GLASER.

p. 409) the non-fertilization of the majority of the ova, whereby
it is impossible for them to develop to any great extent, and
whereby they naturally break down when they have endeavored
to segment. We see this in Neritina, Buccinuui and Pnrpnra
lapillns. In Fasciolaria, as stated above, the process reaches its
climax, and in this case the sterile, nutritive ova do not show the
least trace of segmentation, nor do they ever show signs of ma-
turation."

This second factor seems to me most important, and in view
of recent results on the development of the germ cells of gas-
tropods, places the origin of cannibalism in a new light. In
the first place the ingested materials may consist of undivided
eggs, or of embryos or of a combination of these classes of
constituents. Thus in Ncritina and Fasciolaria the ingested
eggs do not divide. In Buccimtin most of the ingested ova
divide, though some do not, and in Crepidula, Pnrpnra and
Urosalpin.v, the ingested materials are derivatives of either early
or late stages of development. The habit of Fasciolaria, though
coming in the class with that of Neritina, also resembles that of
Crcpidula, Pnrpnra and Urosalpin.v, for the abnormal larvae of
the pre-cannibal period, as well as many of the abnormal late
ones, including some dwarf larvae, are ingested together with the
unfertilized eggs. That these materials are taken in by the
embryos requires no explanation in this connection, for not
only are they obviously useful as food, but the structures in
virtue of which the larvae cannot help ingesting all the available
substances in their environment were evolved, for locomotion and
for swallowing other things, in all probability long before Fascio-
laria exchanged its pelagic larval life for the safer one within
capsules. To explain the origin of the habit of cannibalism it
seems to me necessary to answer only two questions : (i) Why
do some of the embryos break down ? (2) Why are many
of the eggs infertile ? By answering these questions materials
to be swallowed are accounted for : the consumers, and the
mechanism by which the consumption is brought about, are the
results of that portion of the phylogenetic history of Fascio'.aria
which preceded the origin of cannibalism. Given therefore
ingestors, and materials to ingest, cannibalism follows as a natural
consequence, needing no further explanation.

THE DEVELOPMENT OF FASCIOLARIA. 159

That some of the embryos break down is due, in Fasciolaria at
least, to the fact that they have been outstripped in development
by their competitors, whose superior strength, traceable to indi-
vidual differences in the sizes of the fertile eggs, in the time and
places of fertilization, in the ability at various stages to withstand
the effects of exposure at low tide, or the jostling and even in-
juries to which the capsules are naturally exposed, probably
enables these more fortunate ones to set up such lively currents
by means of their cilia that their weak relatives are either whirled
about until seriously harmed, or are so crowded into the mass ot
collected food ova that injury is certain and death probable. I
do not believe that it is necessary, therefore, to trace the sterile
eggs historically to degenerating embryos, for even if it were easy
to see how selection could occur in this particular case, the phys-
ical conditions which account for the breaking down of some of
the embryos at the present time, were probably always operative
and seem to me of themselves to explain the facts sufficiently well.

That many of the eggs are unfertilized is more difficult to
understand. This is true of Fasciolaria, of Neritina (Biitschli,
Blochmann), and of Purpura (Selenka). In this last form
Selenka was not able to discover a nucleus. The infertile eggs
of Fasciolaria do not react to the influence of chemicals, though
in some cases when a few drops of ammonia were added to the
water Professor Conklin found that a number of eggs formed pro-
trusions at one end. I have repeatedly tried artificial fertiliza-
tion, with perfectly fresh eggs, in pure sea water, in a mixture
of sea water and the albuminous contents of the capsules, or in
the pure albumen. In no case did I succeed in finding more
than the usual number of segmenting eggs in such collections,
and in no case did the undivided ones seem to have attracted the
spermatozoa. Since fertilization normally precedes maturation in
the gastropod ovum, I conclude that these eggs are imperfect.

In what does their imperfection consist ? This is a question
which I cannot answer. It seems to me, however, that the di-
morphism of the spermatozoa may point the way to a solution.

Meves ('02) described most remarkable differences between
the development of the oligopyrene or worm-shaped sperma-
tozoa and the eupyrene or ordinary hair-shaped spermatozoa of

l6o O. C. GLASER.

Pahtdina. These differences in development first become man-
ifest by the occurrence of two sizes of primary spermatocytes, the
larger of which give rise to the oligopyrene spermatozoa, whereas
the smaller ones give rise to the eupyrene forms. When the
larger primary spermatocytes undergo maturation no reduction
takes place, but the normal number of chromosomes (fourteen
for Paludina) appears at the nuclear plate. These chromosomes
are so distributed in the ensuing division that one of the resulting
daughter cells (secondary spermatocytes) has four and the other
has ten. When the secondary spermatocytes divide to form the
spermatids, only one of the chromosomes of the secondary sper-
matocytes undergoes division, all the others (three for one class
of spermatocytes and nine for the other class) degenerate. In
this way it happens that the nucleus of the oligopyrene sperma-
tozoa is composed of a single chromosome, whereas that of the
functional eupyrene sperms has seven.

Fasciolaria, in company with a large number of other proso-
branchs, has two kinds of spermatozoa, and - there can be little
doubt that these correspond respectively to the oligopyrene and
the eupyrene sperms of Paludina. What the reason for this
dimorphism may be is not clear, but so far as I can see, there is
no evidence to show why it might not also occur among eggs,
particularly of a form presenting such well marked differentiation
of its male sexual elements as Fasciolaria does.

That Fasciolaria has two kinds of primary oocytes which differ
most remarkably in their reactions with spermatozoa, and con-
sequently in their ultimate fate, is beyond dispute. This difference
may possibly be due to an homology between the infertile oocytes
and those primary spermatocytes which give rise in the manner
described by Meves to the oligopyrene spermatozoa. Whether
further investigations establish this homology or not, the presence
of the infertile eggs is the keystone of the conditions that deter-
mine cannibalism. The origin of this process therefore is to be
sought in those circumstances that determine the formation of
the sterile ova. That these ova should be ingested follows from
the automatism of the larvae, based on structures much older than
the habit of cannibalism, and not to be explained by it. The
persistence of those processes which give rise to the nutritive ova

THE DEVELOPMENT OF FASCIOLARIA. l6l

however can be explained by the theory of selection, since this
method of feeding the young is useful to the species.

The sterile ova --to whatever cause due — have an important
influence on development ; indeed, all the facts which I have dis-
cussed or mentioned in the preceding pages can be united into a
system of correlations, each part of which has antecedents or
consequents, or both, traceable to the nutritive ova as the first
link in a long chain of events. It is clear that without the sterile
eggs cannibalism could not occur. It is equally clear that the
larvae prepare for this process, and that they are profoundly
modified as the result of it. Thus, to consider first external
characters, the frothy, irregular ectoderm of the precannibal
period is well fitted for the stretching caused by the ingestion of
the eggs, for these produce a distension so great that unless
provision were made for it in advance a far larger number of
embryos would be destroyed by it than as a matter of fact
succumb.

The immediate results of the ingestion of the eggs are an en-
tire change in the shape of the embryo, and a great increase not
only in its size, but also in the size and organs of the young.
The external kidneys are most clearly correlated with the canni-
balism. In this case also we meet with provision, for the excre-
tory organs appear, and are more highly developed than in any
other gastropod embryo known to me, long before their chief
need can be felt, and long before they have reached the highest
development which they ultimately attain. The early appear-
ance of the external kidneys, which a comparison with other
prosobranchs snows to be secondary, brings about a change in
their position, for if they developed at the time the velum ap-
pears, as they do in Fulgnr (McMurrich), or after this appears
as they do in Crcpidula (Conklin), they would not be carried out-
wards by this organ, ultimately to hang down from its underside.
The early development of the external kidneys is thus a case of
ccenogenesis, and their final location on the embryo an excellent
example of a conspicuous result due to a remote influence, for
although this ultimate position and activity are connected with the
egg-swallowing habit, this connection is indirect, since fully-de-
veloped external kidneys occur in dwarfed larvae devoid of

1 62 O. C. GLASER.

ingested ova. It might be urged from this that there is not even
an indirect connection between the cannibalistic habit and the
excretory organs, but this is by no means true. It is an acci-
dent in the lives of the dwarfed embryos that they fail to secure
any eggs, for they prepare for them as much as their more suc-
cessful competitors do. The fact that the preparations for an
event which never comes to pass are elaborate cannot show that
this event had no influence on the lives of the anticipators.
What it does show is that this influence is not direct, for the
habit of preparing for cannibalism has become fixed through
selection.

Another correlation of importance is the amitosis in the exter-
nal kidneys and in the oesophageal endoderm. Here the need
for the rapid digestion of great quantities of food material and the
excretion of waste products has called forth a process unusual in
embryonic cells, but, as I shall try to show in another paper, not
pathological. There can be little doubt that through these ami-
toses other correlative changes are brought about, particularly in
the development of the oesophagus, where the gaps made by the
degenerating embryonic digestive cells are certainly repaired
before the adult stage has been reached. Thus the whole devel-
opment, from early stages to late, the structure, shape and size
of the larvae and the size and hardiness of the young, and what-
ever these stand for in their further lives, are affected by cannibal-
ism, the origin of which is traceable, I believe, not to the advan-
tage which accounts for its persistence, but to some as yet
unknown cause which determines the existence of the sterile
nutritive ova.

ZOOLOGICAL LABORATORY, UNIVERSITY OF MICHIGAN,
ANN ARBOR, January 4, 1906.

LITERATURE CITED.
Blochmann, F.

'82' Uber die Entwicklung der Neritina fluviatilis. Zeit. f. w. Zoo]., Bd.
XXXVI., 1882.

Bobretzky, N.

'77 Studien iiber die embryonale Entwicklung der Gastropoden. Arch. f. mikr.
Anat., Bd. XIII., 1877.

Brooks, W. K.

'78 Preliminary observations upon the development of the marine prosobranchiate
Gastropods. Chesapeake Zool. Laboratory, Johns Hopkins Univ. Scient.
Results, 1878.

THE DEVELOPMENT OF FASC1OLARIA. 163

Biitscbli, 0.

'77' Uber Paludina vivipara. Zeit. f. w. Zoo!., Bd. XXIX., 1877.

Carpenter, W.

'57 On the development of Purpura. Ann. Mag. Nat. Hist., 2 ser., Vol. XX.,
1857.

Conklin, E. G.

'97 The Embryology of Crepidula. Journal of Morphology, Vol. XIII., 1897.

Erlanger, R. v.

'92 Beitrage zur Emvicklungsgeschichte der Gastropoden. Erster Theil. Zur
Entwicklung von Bithynia tentaculata. Mitth. Zool. Station Neapel., Bd.,
X., 1892.

Fol, H.

'75 Sur le developpement des Gasteropodes pulmones. Compt. Rend. Acad.
Sc., T. LXXXI., 1875.

Fol, H.

'792 Etudes sur le developpement des Gasteropodes pulmones. Arch. Zool. exp.
gen., T. VIII., 1879.

Glaser, 0. C.

'c>42 Excretory Activities in the Nuclei of Gastropod Embryos. American Nat-
uralist, Vol. XXXVIII., 1904.

Glaser, 0. C.

'05 Uber den Kannibalismus bei Fasciolaria tulipa, etc. Zeit. f. w. Zool., Bd.
LXXX., 1905.

Heymons, R.

J932 Zur Entwicklungsgeschichte von Umbrella mediterranea Lam. Zeit.
f. w. Zool., Bd. LVI., 1893.

Koren u. Danielssen.

'56 Fauna littoralis Norvegiae. Bergen, 1856.
Korschelt u. Heiser.

'93' Lehrbuch der vergleichenden Entwicklungsgeschichte. Jena, 1893.

McMurrich, J. P.

'861 A contribution to the embryology of the Prosobranch Gastropods. Stud. Biol.
Lab. Johns Hopkins University, Baltimore, Vol. III., 1886.

Meisenheimer, J.

'98' Entwicklungsgeschichte von Limax maximus, II. Zeit. f. w. Zool., Bd.
LXIII., 1898.

Meves, F.

'02' Uber oligopyrene und apyrene Spermien und iiber ihre Entstehung nach Beo-
bachtungenan Paludina und Pygaera. Zeit. f. w. Zool., Bd. LXXIL, 1902.

Osborn, H. L.

'862 Development of the Gill in Fasciolaria. Stud. Biol. Lab. J. H. U., Balto.
Vol. III., 1884-1886.

Osborn, H. L.

'04 Amitosis in the Embryo of Fasciolaria. Science, Vol. XIX., 1904.

Rabl, G.

'791 Uber die Entwicklung der Tellerschnecke. Morphologisches Jahrbuch,
Bd. V., 1879.

164 O. C. GLASER.

Salensky, W.

'jz2 Beitraga zur EnUvicklmigsgaschichte der Prosohranchier. Zeit. f. w. ZooL,
Bd. XXII., 1872.

Sarasin, P.

'Sz2 Entwicklungsgeschichte der Bithynia tentaculata. Arb. Zool. Inst. Wurz-
burg, Bd. VI., 1882.

Selenka, E.

'72' Die Anlage der Keimhlfitter bei Purpura lapillns. Niederland. Arch, fur
Zoologie, I. Bd., 1871-1873.

Stauffacher, Hch.

'g82 Die Urniere von Cyclas cornea. Zeit. f. w. Zool., Bd. LXIII., 1898.

A CHAMELEON-LIKE CHANGE IN DIEMYCTYLUS.1

(PRELIMINARY REPORT.)
CHARLES G. ROGERS.

In connection with some work done upon the heliotropic re-
sponse in the salamander Diemyctylus viridescens, it was noticed
that under certain conditions the animals changed color within a
few hours in a most remarkable manner. The following paper
is a report of the results obtained in a study of some of the con-
ditions affecting this color change.

That light and heat are active stimuli in the case of many ani-
mals is a well-known fact. It has also been shown repeatedly
that not only do these stimuli affect the movements and orienta-
tion of the animals, but may, in some instances at least, bring
about changes of color as well.

The literature upon this part of the subject is not extensive, so
it is impossible at present to formulate any general rule with
respect to which these changes may take place. It may be said,
however, that there is a remarkable uniformity in the results
obtained by different investigators.

In attempting to explain the pigmentation of the salamander,
Salamandra maculata, Fischel2 found that the temperature to
which the animals were subjected was an important factor. He
observed that the larvae of the salamander which developed in
warm water were of a lighter color than those which developed
in cold water. If dark-colored larvae were placed in warm water
they became lighter in color, the degree of the change varying
with the age of the larvae at the beginning of the experiment. The
newly-hatched larvae were found, also, to be much more suscep-
tible to the changes in temperature than the older larvae. As the
larvae became older the effects of the temperature to which they had
been subjected tended to become fixed. If light-colored larvae
were placed in colder water, the converse change was noticed,

1 From the Zoological Laboratory of Syracuse University.

2 Fischel, Arch.f. Mikr. Anat., XLVI., pp. 719-748.

165

1 66 CHARLES G. ROGERS.

and this also tended to become permanent as the age of the ani-
mals became greater.

Flamming l observed that the conditions of light as well as of
temperature make their impress upon the coloration of the sala-
mander. Animals left in dark aquaria become and remain dark,
while those in the light, for example, in white porcelain dishes,
become light, the temperature of the water being the same in the
two cases.

The recent work of Carlton 2 upon the chameleon Anolis gives
us even more striking evidence of the part played by external
conditions in the coloration of animals. He found that the skin
of Anolis can be made to assume one of two colors, dark brown
or pea green. The brown state for animals in confinement is
taken on in daylight and is produced by the outward migration
of pigment granules from the bodies of the melanopores into the
processes and ultimate branches. This outward migration is
accomplished in about four minutes. It may be brought about
either by mechanical stimulation of the skin or by an act of
the nervous system. The brown state is ordinarily maintained
by a tonus established by the sympathetic nerves and dependent
upon the stimulation of the nervous end organs in the skin by
the light. The melanophores of Anolis are not directly stimu-
lated by the light. The green state is taken on in the dark, and
is produced by the inward migration of the pigment granules of
the melanophores whereby the reflecting ochrophore becomes ex-
posed to the light. This inward migration requires about twenty-
five minutes. It may be induced by any means which will bring
the melanophores into the unstimulated state.

EXPERIMENTS UPON DIEMYCTYLUS.

The material upon which these observations were made was
the salamander Dicmyctylus viridescens. Specimens were collected
»n the fall of the year and had been kept in the laboratory during
the winter in glass jars. The water in the jars was frequently
changed and the animals were fed upon raw beef at regular inter-
vals. In the spring observations were also made upon specimens
just taken from their usual environment with similar results.

1 Flemming, Arch. f. Mikr. Anat., XLVIII., pp. 690-692.

2 Carlton, Proc. Atner. Acad. Arts and Sci., XXXIX., No. IO, pp. 259-276.

A CHAMELEON-LIKE CHANGE IN DIEMVCTVLUS. 1 6/

When under their normal conditions the salamanders' are of a
dirty yellow brown color upon the dorsal surface and a lemon
yellow upon the ventral surface. Upon both surfaces there are
black pigment spots about one half millimeter in diameter-
There are also upon the dorsal surface a number of scattered red
pigment spots. This description may apply equally well to ani-
mals just taken from their native surroundings or which have
been kept in the laboratory for a considerable period at the room
temperature. Specimens taken at random from the ponds do not
vary much in color.

When the animals are exposed to a temperature which is much
below that to which they have been accustomed it was found
that there resulted a change in the appearance of the skin which
would continue as long as the low temperature was maintained.
Instead of a dirty yellow, the skin assumed a much darker color,
becoming dark brown, dark green or in extreme cases almost
black. This change took place in the course of a few hours and
remained so long as the animals were subjected to the artificial
condition. The amount of change in color which took place as
the result of the change in temperature was in a measure a func-
tion of the total amount of temperature change. If the change
were only a slight one the color response would be correspond-
ingly slight, while if the temperature change were greater the
color change would be also more pronounced.

If, on the contrary, the temperature of the water in which the
animals were living was raised, it was found that the animals
responded to the increased temperature by a lightening of the
skin. Under these conditions the yellow in the skin became more
pronounced and the darker colors less so. This change also was
found to continue so long as the water remained warm. When
the water became cold the color of the skin would return after a
given interval to what it was at the beginning of the experiment.

Another fact of interest in connection with this color change is
that if the temperature of the water be maintained at a constant
point and the intensity of the light be changed, the color of the
animals will be found to respond in a very definite way. It must
be stated, however, that the response in this latter case is less
pronounced than that obtained by a change in temperature, and

1 68 CHARLES G. ROGERS.

may indeed be obscured by the action on a change in tempera-
ture. If the animals were placed in the dark for several hours
they all became darker, if in the light for a number of hours they
all became lighter. It was further found that the change in
color due to a change in light conditions might be inhibited by
contrary changes in the temperature conditions. The following
table represents various combinations of light and temperature
conditions which have been tested and the results obtained in the
coloration of the animals.

TEMPERATURE. IN DARK. IN LIGHT.

High, Ordinary, Very Light.

Low, Very Dark, Ordinary.

By high temperatures are meant temperatures of from thirty-
five to forty degrees Celsius ; by low, temperatures below ten
degrees Celsius.

From the table it is readily seen that we are here dealing with
two sets of forces which tend to neutralize each other when
applied in opposite directions and to augment each other when
applied in the same direction.

The manner in which this change is brought about is the same
as that described by Carlton for Anolis. Small granules of pig-
ment migrate outward from the large pigment cells along the ray-
like processes of the cells when the animal is placed in the dark,
thus obscuring the yellow pigment which lies in the deeper
layers of the skin. When the animal is brought again into the
light the pigment granules again migrate into the bodies of the
pigment cells, and the lighter yellow pigment of the deeper layers
of the skin becomes visible. The operation is a rather slow one
and may occupy an hour or more of time.

In certain animals a condition of permanent darkness was
established through a section of the optic nerves. So long as
the animal had one or both eyes functional the action of the skin
was as stated above. When both eyes are made functionless a
marked change takes place. Within a short time the skin begins
to darken and within two hours has taken on a dark brown, dark
green or black color. The appearance of the animals is similar
to that produced by the action of darkness and low temperature.
At ordinary room temperatures this change is permanent.

A CHAMELEON-LIKE CHANGE IN DIEMVCTYLUS. 169

But it is a remarkable fact that in this blind condition the
animals are even more than normally sensitive to changes in
temperature. If the temperature on the water be raised by any
considerable amount the color of the animal changes correspond-
ingly, becoming much lighter. On the other hand if the tem-
perature of the water be lowered they again resume the intense
dark color which came on as the result of the operation. These
changes in color usually occupy a time of not far from two hours.

The question now arose as to how this color change was con-
trolled. It was at first thought that it must be controlled through
the central nervous system, but the following experiments seem
to throw doubt upon such an explanation.

Section of tlie Spinal Cord. — It has been shown above that
section of the optic nerves brings about a most remarkable result
in the coloration of the salamander. Now if the impulses effective
in bringing about this change pass to the skin through the brain
and spinal cord we should find upon section of the spinal cord
that the parts of the skin supplied by nerves arising from the
spinal cord below the cut should respond in a different way from
those parts of the skin supplied by nerves arising from the spinal
cord above the cut. In order to test this possibility a number
of salamanders were operated upon in the following way. The
spinal cord was completely sectioned at the level of the third or
fourth thoracic vertebra while under the influence of ether. At
the same time other individuals were subjected to the same
anesthetic for the same length of time to assure us that any
change in the appearance of the animal was not due merely to
the effect of the ether. After the animals had been given oppor-
tunity to recover from the shock effects of the operation, no
change being found in the color of the skin as the result of the

o o

first operation, they were again taken and subjected to section of
the optic nerves. Following this operation, usually within a
period of not more than two hours there was found that same
remarkable darkening of the skin which has already been
described. And this darkening of the skin involved the whole
dorsal surface of the animal and not merely that part of the skin
controlled by nerves arising either above or below the section of
the spinal cord. From this it is seen that the nervous impulses

I/O CHARLES G. ROGERS.

must have some other path of communication between the optic
centers and the pigment cells of the skin than that furnished by
the brain and spinal cord. It might be possible that the skin itself
could furnish such a channel of communication, but a more
reasonable means would seem to be through the the sympathetic
nerves. In this our results agree with those of other observers.
There is also in the vicinity another variety of Dieinyctylns
viridescens known as miniatus. This is a land form and is of a
vermilion color. It possesses few or none of the black pigment
spots and apparently very few pigment cells. A series of experi-
ments was also carried out upon this form but in no case was
there any change of color noticed after section of the optic nerves.

BIBLIOGRAPHY.
Carlton.

Proc. Amer. Acad. Arts and Sci., XXXIX., No. 10, pp. 259-276.
Fischel.

Arch. f. mikr. Anat., XLVI., pp. 719-748.
Flemming.

Arch. f. mikr. Anat., XLVIII., pp. 690-692.

THE GROWTH AND REGENERATION OF THE
GILLS IN THE YOUNG NECTURUS.

ALBERT C. EYCLESHYMER,

DEPARTMENT OF ANATOMY, ST. Louis UNIVERSITY, ST. Louis, Mo.

During the summer of 1904 the writer made a series of obser-
vations and experiments on the growth and regeneration of the
gills of the larval Necturus. The work was done with the hope
of obtaining more information as to the relation which the regen-
erated gill bears to the normal in its rate of growth, in its size,
and particularly in its pattern.

The larvae were kept in aquaria, from which they were taken at
intervals, placed in flat-bottomed watch-glasses above mirrors,
and sketches made of the ventral surfaces of the gills. The
period of observation on the growth of the normal gills was
from the time (12—13 nim) of their first appearance, to the time
(18-20 mm.) when four pairs of filaments were present. At this
time the gills were cut off and preserved.

These larvae were then placed in separate aquaria and like the
normal were sketched at successive intervals. The period of ob-
servation on the regenerating gills was from the time of excision
until about the same number of filaments were present as in the
normal at the time when they were cut off.

Ten series were started, but only eight were completed, owing
to the death of two larvae. That the normal processes might be
as little disturbed as possible but a single gill was removed from
each larva. This one was in each case the anterior on the right
side of the head. Since the eight experiments gave similar re-
sults, but three have been described and illustrated in detail.

The first appearance of the gill bars is to be seen in the 9-10
mm. larva as slight swellings on each of the three gill arches.
When the larva has reached a length of 12-13 mm., the first and
second gill bars measure about i.o mm., while the third measures
0.5 mm.

The first gill filament appears at this time as a slight swelling

171

172

ALBERT C. EYCLESHYMER.

on the postero-ventral surface of the first bar, midway between its
base and apex ; soon a second filament appears on the antero-

Q

8

e

9

H.

a ff

zzr

I 3

FIG. I. Each series is bracketed and designated by Roman numerals. The upper
row in each series represents the changes in the normal gill. The lower row of each
series represents the changes in the regenerating gill of the same larva. The nume-
rals indicate the days of the month : June 16-30 and July 1-17. All the figures are
magnified about ten times.

ventral surface opposite the first, as shown in Ser. II., Fig. 19.
Often the buds form at different levels, giving rise to unsymmet-

GILLS IN THE YOUNG NECTURUS. 1/3

rical patterns. While the lateral buds are elongating the tip of
the gill bar becomes drawn out into a median filament, as in Ser.
III., Fig. 21. Sometimes the tip of the gill bar forms one of the
first pair of filaments, in which case a median bud forms the me-
dian or apical filament.

A second pair of filaments forms some 48 to 60 hours later.
In most cases these are bilaterally symmetrical, as indicated in
Fig. 21. In two out of the eight series the buds did not arise at
opposite points, and in these cases more or less irregularity oc-
curred in the succeeding pairs.

The third pair of filaments shows considerable variation both
in time and place of origin. In five cases they appeared about
48 hours alter the second pair. In the remaining three cases they
were not formed until the end of the third day. The filaments
of this pair usually arise in such positions that they are bilaterally
symmetrical, as shown in Ser. III., Fig. 26. In two series con-
ditions were observed something like that represented in Ser. II.,
Fig. 26, where only one of the third pair arose in the usual posi-
tion. In place of the other there arose an intercalated bud on
the opposite side.

The fourth pair of filaments may develop as bilaterally sym-
metrical structures, as in Ser. I., Fig. 30, or may be represented
by a single intercalated filament, as in Ser. III., Fig. 30, or again
the filaments may show such great irregularity that it is impos-
sible to tell what filaments are to be considered as belonging to
the fourth pair, as in Ser. II., Fig. 30.

In the later growth of the gill the filaments not only continue
to form in pairs at the base of the gill bar but also to arise irregu-
larly from intercalated buds. In addition the filaments send off
lateral branches which form secondary filaments. In the older
larvae (30—40 mm.) the patterns become more and more
irregular.

As a general statement one might say that in the earliest fila-
ments much regularity prevails, but in the formation of the later
filaments the regularity decreases.

When the gills had reached the stages shown in Fig. 30 they
were cut off with spring scissors as near the head as possible and
as they regenerated their changes were carefully followed and

174 ALBERT C. EYCLESHYMER.

compared with the successive changes observed in the growth of
the normal gills.

o

In Ser. I. the first pair of buds formed at about the same rela-
tive time as in the normal, but instead of conforming to the rather
exceptional pattern of the normal they were symmetrically placed,
as shown in Fig. 6. The second pair of filaments while sym-
metrical were not present until 96 hours later. In the normal
the second pair arose 72 hours after the first pair. The third
pair appeared some 48 hours after the second and in both time
and place of origin conformed to the normal. The condition
about 132 hours later is shown by Fig. 17. Comparing this
figure with Fig. 30 it will be noted that but a single filament is
present to represent the fourth pair of the normal. Again there
are present two intercalated buds while the normal shows none.

In Ser. II. the filaments appear later than in the normal and
instead of arising as a pair there are three buds present (Fig. 6).
As the later stages show, the first pair is made up by the tip of
the median and the first lateral filament. About 96 hours later,
as shown in Fig. 11, three more buds are present. Two of these
form a new pair at the base of the gill bar, while the third arises
between the tip of the gill bar and the first lateral filament.
When compared with the normal of this stage it is obvious at a
glance that there is no conformity. No other changes occur
which merit detailed description. If the stage shown in Fig. 1 7
be compared with that of the normal shown in Fig. 30 it will be
readily observed that the regenerated gill is quite unlike the
normal, but on the whole it presents a more regular pattern than
the normal.

In Ser. III. the first pair of filaments appeared at about the
same relative time as in the normal but at widely separated points,
as shown in Fig. 6. The second pair formed some 96 hours
later and, alternating with those of the first, gave rise to the pat-
tern shown in Fig. 11. The third pair, as shown in Fig. 13,
appeared 72 hours after the second as symmetrically placed
structures. The fourth pair were likewise symmetrically placed.

It may be remarked that the remaining five series showed, with
slight variations, the same results as described above. In the

o

entire eight series not a single case was observed in which the

GILLS IN THE YOUNG NECTURUS. 175

pattern of the growth changes in the normal gill were repeated
in the regenerating gill.

If we compare the rate of growth in the regenerating gill with
the rate of the normal gill we find that the stage reached in nor-
mal growth in sixteen days is reached in regeneration in about
eighteen days.

That the pattern of the growing gill is not retraced by the
regenerating has been established beyond question in these ex-
periments. In the regeneration of the foot of the young Ncct/tnis
the course of normal development is pretty accurately repeated
and the same is known to be true of many other forms.

Previous experiments by others seem to indicate that if wide
variations occur in normal growth we should expect to find like
wide variations in regeneration.

One pattern of gill is as efficient for respiration as another pro-
vided it posesses the same number of filaments. The same is
true of the regenerated gill. The normal pattern of the foot is
duplicated in regeneration because the type evolved is that best
adapted to the needs of the animal.

If these considerations be well founded we are led to regard
physiological efficiency as the important factor in the regenera-
tion of the gills.

In other words we may conclude that functional and structural
regeneration may run parallel or they may follow widely diverg-
ing lines.

OBSERVATIONS ON THE CHANGES IN THE HYPO-

DERMIS AND CUTICULA OF COLEOPTERA

DURING ECDYSIS.

W. L. TOWER.

The internal changes which occur during the periodic removal
and redevelopment of the chitinous portion of the integument of
insects are little understood and but few observations have been
made thereon. In this paper are given observations and conclu-
sions concerning some of the changes found in the integument
of Leptinotarsa decimlineata and Clirysobotliris femorata. These
two beetle larvae are good examples of two types of larvae, the
first living freely exposed upon their food plants and are typical
examples of those insect larvae which pass their lives upon plants
in exposed places and must go through ecdysis exposed, and the
second of larvae that live in burrows or cells, protected from ex-
ternal interference to a great degree. Corresponding to the differ-
ence in their habitats they show differences in the internal changes
accompanying ecdysis.

The life of an insect larva is made up of a series of instars or
stages, each of which represents a very precise cycle of develop-
ment and physiological activities. These cycles of changes are
of interest and serve to give a good basis for the orientation of
the changes which I am about to describe as accompanying
ecdysis. The changes within a single cycle are given in the
tabulated form on page 177.

Each larval stage properly begins with the period of growth
following the reconstruction period of the preceding ecdysis and
ends with the end of the next reconstruction period. Looked at
from the exterior the stages extend from ecdysis to ecdysis, but
as shown in the above table the period of exuviation is the middle
one of the short and rapidly passed over periods in which the
process of ecdysis is begun, achieved and the animal recovers
from the effects thereof. The changes with which this paper
deals are confined largely to the last three periods in the cycle.

176

OBSERVATIONS ON COLEOPTERA.
CYCLE OF PERIODS IN ONE LARVAL STAGE.

177

Period of
Growth.

Period of
Maximum
Nutrition.

Period of
Differentia-
tion.

Period of
Preparation.

Period of
Exuviation.

Period of
Reconstruction.

Larva actively feeding.

Larva not feeding

.- ~J <u

1 1 • - -4-J 1

1 >, •- C T3

.255

<U 1> *-" >•* '-^ V *••* C/5

•ji e jj <u

<U C — C U

.a o 12 D •=

<-> •— <u .—

'H C3 -r- « P

eeding voraciously ; rapid increase in s
and in thickness of integument ; intens
cation of color.

eeding voraciously ; integument reac
its maximum thickness ; abdominal
gion distended to its full capacity; s
face smooth ; fat body growing rapid

eeding less actively ; abdomen much c
tended ; fat body greatly increased ; ma
ed growth of imaginal organs and rapid c
ferentiation of gland cells and leucocyl

ceding but little ; stops feeding ; restle
seeks sheltered place ; rapid developm
of exuvial glands ; of zymogens in
hypodemal cells ; hypodermis with b?
ment membrane drawn away.

esting quietly ; contractions begin,
come rhythmical ; . dissolving of sec
dary cuticula ; formation of exuvial flu
rupture of exuvial sutures ; detachm
of the muscles ; formation of new cu
ula ; removal of old cuticula.

arva weak ; soft ; extended ; quiet ; cu
ula hardens; color develops ; second
cuticula begins to form ; glands subsi
hypodermis becomes condensed into t
layer ; larva begins to move about ;
begins feeding.

to

to

to

to

K

i-l

The Exui'ial Glands. — The existence of an exuvial fluid in in-
sects was first clearly demonstrated by Newport, but its origin
remained obscure and was attributed to diverse sources. It was
first shown to be in part at least, due to glandular activity by
Gonin in Picris brassicte. Gonin found, especially upon the
pronotum, a considerable number of large unicellar glands which
were in his preparations definitely seen to be extruding their con-
tents between the old and new cuticula to form a part at least of
the exuvial fluid. Similar unicellular glands are found in other
insects, especially at pupation, in free living forms such as most
of the larvae of butterflies and moths and of many leaf feed-
ing beetle larvae. In L. dccimlincata there are found from 50 to
22$ of these glands upon the pronotum in the last larval stage,
and their sole function seems to be to develop an enormous
amount of the exuvial fluid. The glands are found all through
the life of the animal and upon all parts of the body, but in fewer
numbers than upon the pronotum at pupation. Some idea of the
abundance of these glands may be gained from the section shown
in Fig. i, a section from the pronotum of L. decimlineata at pupa-
tion when these unicellular glands are closely crowded and are
all ejecting their contents into the space between the old and new
cuticula (c1, c2, r3).

17° W. L. TOWER.

These glands arise in the embryo and in later larval stages
through the modification of the hypodermal cells. In Fig. 2 are
shown five stages in the development of these glands.

The first stage in the development of one of these glands in
larval life is usually found in the latter portion of the period of
maximum nutrition and shows the nucleus greatly enlarged, the
chromatin greatly increased in amount and scattered, with the

FIG. I. L. efecim/ineata. Section of pronotum of larva about to pupate, in early
stage of exuviation, showing large number of exuvial glands, g. C/, cuticula of larva
being removed ; C3, developing new cuticula ; H, hypodermis ; /, fat body ; »/,
muscles ; /, trachea.

cytoplasm seen (Fig. 2). In the period of differentiation thisd
cell grows very rapidly projecting inward below the hypodermis,
but retaining a delicate connection with the cuticula (Fig. 2, B}.
The cell remains in this condition through the following ecdysis
and appears after ecdysis unchanged, excepting that the gland
now has a delicate duct leading to the surface, this being devel-
oped as the result of the glands not forming chitin at this point
of attachment to the outer surface of the hypodermis (Fig. 2, C),
During the last larval stage the character of these glands, especi-
ally in the periods of differentiation and preparation, change rap-
idly (Fig. 2, D and E), becoming first larger, the nucleus and
chromatin growing immensely in size and quantity, the cyto-
plasm remaining dense and slightly vacuolated. At the begin-

OBSERVATIONS ON COLEOPTERA.

'79

ning of the period of preparation there is a rapid change, chroma-
tolosis of the nucleus, a marked shrivelling and decrease in size,
and the cytoplasm becoming entirely vacuolated (Fig. 2, E). As
ecdysis begins and the old cuticula separates from the hypodermis
the contents of these glands are forced out between the old cutic-

c

(9) £.

FIG. 2. /.. decinilineala. Five stages in the development of the exuvial glands.
A, young stage showing rapid growth of nucleus in hypodermal cell ; B, later
stage showing cell much enlarged and projecting below the hypodermis, taken just
before second ecdysis; C, third stage just after second ecdysis, showing the cell as a
fully developed unicellular gland with outlet to surface ; D and E show the great
increase in size of these glands preceding pupation (at E is shown glands extruding
their contents to form part of the exuvial fluid) ; <•', primary cuticula; r2, secondary
cuticula; h, hypodermis; »ii>, basement membrane.

i So

\V. L. TOWER.

ula and the new to help form the exuvial fluid. Eventually
nearly all of the contents of the glands are extruded, leaving
them small and shrunken as shown in Fig. 3, A.

After ecdysis and especially after pupation the degeneration of
these glands is rapid, as shown in Fig. 3, the cell speedily return-
ing to a normal hypodermal cell in size and ultimately it breaks
down completely.

As far as I can discover these glands are exactly like those
found by Gonin in Pieris brassiac, and I have observed the same
structure in Pieris rape? and frotodice, and Clisiocampa aincri-
cana among the Lepidoptera and they are widespread in the

B C

FIG. 3. L. dedmlineata. Three stages in the degeneration of the exuvial glands.
A, gland immediately after ecdysis ; B, stage of gland during the period of recon-
struction ; C, gland nearly reduced to normal hypodermal cell in period of growth.

Chrysomelidae and Coccinnellidae, especially in the tropical species.
Nowhere have I found these unicellular glands in larvae that live
in burrows, or in the soil or in cells, but only in larvae living
freely exposed upon plants where there exists the greatest liability
to rapid desiccation. In L. dedmlineata, innltitccniata and their
tropical allies these glands developed in the embryo in small
numbers over the entire body surface and are active at each ecdy-
sis. As the larva grows, however, the number of these glands
increase until at the time of pupation there are very many of them
scattered over the body, but they are most numerous upon the
pronotum. During the pupal period nearly all degenerate rap-
idly and but few are functional at the final transformation.

These glands and their increase in number is, I believe, an
adaptation in these freely exposed larvae, to enable them to pass
with the least mortality through such critical periods of their life
as ecdysis and pupation. I can see no reason why this adapta-

OBSERVATIONS ON COLEOPTERA. iSl

tion would not be one of direct selective value and be greatly
developed by selection, because those individuals with an abun-
dant supply of exuvial fluid would have a far better chance of
passing safely these critical periods than those with a lesser supply
of the exuvial fluid.

Changes in tJie Integument. - -In the integument the preparatory
changes preceding ecdysis begin before the larvae cease feeding
and consist largely in the withdrawal of the protoplasmic processes
of the hypodermal cells from the pore canals in the secondary
cuticula (Fig. 2, Z>) and the gradual change in the shape of the

FIG. 4. C. jetnorata. Section in latter end of period of differentiation of last
instar, from the raesothorax.

hypodermal cells whereby they become greatly elongated and
their outer ends and the basement membrane separated by two,
three or four times the usual distance. A section (Fig. 4) shows
this condition in which the hypodermis is in the form of a flat-
tened epithelium.

The changes from the condition shown in Fig. 4 go on slowly
until at the beginning of the first contractions all the proto-
plasmic processes have been withdrawn and the hypodermis is
much thickened, due to the drawing away of the basement mem-
brane. With the first contractions the old cuticula and hypo-
dermal cells separate over almost the entire body surface and
only the muscle attachments remain to hold the old cuticula to
the animal.

At the time when the contractions begin or slightly before, a

182

W. L. TOWER.

thin layer of exuvial fluid is found, especially in the anterior parts
of the body between the hypodermis and cuticula, and the inner
surface of the cuticula appears rough and corroded. This corro-
sion of the inner side of the secondary cuticula continues until it
is often almost entirely removed, as shown in Fig. 5.

FIG. 5. C. femorata. Section of the integument from same location as the sec-
tion from which Fig. 4 was taken, showing decrease in thickness of the secondary
cuticula and great extension of the basal ends of the hypoder.iial cells to form a rela-
tively thick but open layer of hypodermis.

This dissolving of the secondary cuticula is a most constant
phenomenon in ecdysis and has been foundin all the insects that
I have examined, but in varying degrees. The same disintegra-
tion of this layer is shown in Fig. 6 of L. decimlineata.

This origin of part of the exuvial fluid in L. decimlineata and
of all of it in C. femorata has not, I believe, been heretofore
suspected but we can at once see the great utility of this process
and especially the advantage gained in having the cuticula thinned

OBSERVATIONS ON COLEOPTERA.

"83

and softened by the dissolving action which is evidently going on.
As far as my experience goes this solution of the cuticula varies
greatly in different preparations, being almost but never entirely
absent in some and exceedingly active in others. I have not

C.

. x . .

/ ''..- ^*»*^

**•

':••••'
•

mb.

FIG. 6. L. decimlineata. Section of pronotum in early part of the period of
ecdysis showing greatly drawn out hypodermal cells and almost entirely dissolved
secondary cuticula. Fragments of cuticula are floating in the exuvial fluid (C2).

determined absolutely what brings about this dissolution of the
secondary cuticula, but it is probably due to enzyme action.
Attempts have been made to isolate these enzymes but thus far
without any marked success. I do find, however, in the hypo-
dermal cells in the late part of the period of differentiation and the
early part of the period of preparation, granules which react and
stain exactly like zymogen granules and which are derived from
the nucleus by chromatolosis and which disappear after the action
of dissolution of the old cuticula has begun. These may not,
however, have anything to do with the disintegration of the old
cuticula. It is clear, however, that the action upon the cuticula
is chemical as there is not the slightest indication of phagocytes
or other organic elements being present. The most logical sup-
position is that the hypodermal cells secrete a substance that
dissolves the old secondary cuticula and thus thins and softens
the integument as well as supplying a considerable part of the
exuvial fluid, thereby greatly facilitating ecdysis.

184 W. L. TOWER.

These changes which occur in the cuticula involve only the
secondary cuticula and are the same irrespective of what the
habitat of the larva may be, as shown in the figures given and
result in all in the thinning and weakening of the old integument.
All these changes take place in the latter part of the period of
preparation and the first third of the following stage.

After the separation of the cuticula and hypodermis the new
primary cuticula begins to form at once. It appears first as a
thin delicate lamella spread evenly over the entire outer surface
of the hypodermis and grows rapidly in thickness until finally
just before ecdysis takes place it reaches its final thickness. It is
developed as a delicate structureless membrane secreted by the
hypodermal cells and there appears at no time evidence in favor
of the oft-repeated statement that the cuticula is the hardened
outer ends of the hypodermal cells. After ecdysis this primary
cuticula hardens rapidly and develops its coloration through
enzyme action precisely as in the adult beetle, a process which I
have described elsewhere.

As soon as ecdysis is over the deposition of the secondary
cuticula begins. This layer is, as is well known (Vossler, Tower)
a carbohydrate allied to tunicin and is deposited in alternating
layers through the periods of reconstruction and growth when it
attains its maximum thickness. It is everywhere penetrated by
delicate pore canals which are the fine canals occupied by the
protoplasmic processes of the hypodermal cells which do not
become detached from the primary cuticula until just before
ecdysis when they are withdrawn. With poor killing and pres-
ervation, however, they are all withdrawn and the canals appear
empty, but are not so in life.

The hypodermis also goes through a regular cycle of changes
in the shape of the cells during each of the cycles. These
changes are first an increase in the number of cells in the growth

o *->

period, and second changes in shape and arrangement so as to
give the body wall the greatest rigidity and strength during the
period immediately following ecdysis before the new cuticula
hardens. In Fig. 7 I have given a series of stages in semi-dia-
grammatic form showing the change in the shape and arrange-
ment of the cells. In Fig. 7, A, in the growth and differentia-

OBSERVATIONS ON COLEOPTERA.

185

tion periods the cells are hexagonal
flattened epithelial and of them-
selves would be a relatively weak
layer if left alone. In the period
of preparation, however, the inner J
ends elongate, the basement mem- |
brane draws away and the cells i-
come to form a much thicker layer, -|
until during the period of ecdysis £
they present the condition shown
in Fig. 7, D, which arrangement,
even though the thick cuticle be
absent, gives a far greater rigidity
to the body wall than the arrange-
ment seen in Fig. 7, A, After |
ecdysis is over the cells gradually =

W '

assume their epithelial character. -3
This series of change represented |

<u

in Fig. 7 evidently is for no other
purpose than that of making the
integument as rigid and strong as
possible during ecdysis. The ar-
rangement developed is one which
in a mechanical way is the best
possible under the conditions. In
fact the arrangement of the base- «
ment membrane, the long drawn §
out hypodermal cells and the devel- ^
oping cuticle is exactly the system -g
used by engineers in large bridge
work or in the building of large
spans or girders.

The change in the integument
are all in the direction of strength-

—c.'

h.

B.

D.

F.

G.

FIG. 7. Diagrammatic representation of stages passed through by the hypodermis
before, during and after ecdysis, to show changes in the form of cells. cf , primary
cuticula ; c2, second cuticula; <-3, old larval cuticula being removed; //, hypodermis ;
tub, basement membrane. All figures are horn preparations of L. decimlineata.

:86

W. L. TOWER.

ening it to enable it to resist the strain and pressure it must
undergo during 'and after ecdysis, and by solvent secretions to
weaken the old cuticula and decrease the resistance offered by
it to its removal and to provide from the material to be cast away
a part of the lubricating fluid, to facilitate the withdrawal of the
body from its old cuticula covering.

The Exnvial Sutures. — It is commonly stated that the rupture
of the integument during ecdysis is caused by the pressure of the

.JOS.

c.

o

FIG. 8. L. decimlineata. Section of the head in the region of the occiput. Last
larval instar showing the line of weakness formed in the mid-dorsal line by the in-
flexed body wall, os, occipital suture ; /, crura of tentorium ; og, occipital fold.

haemolymph during the contractions. It is probably true that
the pressure of the haemolymph and the violence of the contrac-

OBSERVATIONS ON COLEOPTEKA.

i87

tion aid in this process, but that the rupture should occur so
regularly in the mid-dorsal line is remarkable, if this be the only
cause of the rupture. In the insects that I have examined there
always exists in the mid-dorsal line a special device to aid in the
splitting of the old cuticula at ecdysis.

This apparatus in L. decimlineata consists of a well-marked
laterally flattened invagination extending caudalward from the
region of the occipital suture, where it extends deeply into the
head (Fig. 8).

os.

FIG. 9. L. decimlineata. Section through mesothorax. (A) and second ab-
dominal segment; (B) to show continuation of the modifications of median dorsal
line of body to permit of rupture at ecdysis.

This line of weakness in the mid-dorsal line is also the last
point where the ectoderm closes in in development and is in dec-
imlineata turned inward to form a median dorsal fold from the
anterior edge of the occiput to the first abdominal segment and
in this median dorsal furrow the chitin is easily broken during
ecdysis. When the secondary cuticula represented in white in

iSS

W. L. TOWER.

the figures is nearly dissolved away the muscles are still attached
and a slight contraction of the muscles shown in Fig. 8 would
suffice to rupture the cuticula in the median line. And that is
exactly what occurs. That is, the rupture of the integument is
not due to pressure but to the pull of muscles after the cuticula
has become softened. The rupture occurs always in the same
place because there is there a weak spot in the cuticula and
special provision for the rupture thereof.

The same sort of structures are found in C. femarata, P. rapes,
C. aj/icricana and I suspect in most insects and it is perfectly
clear in all of those examined that the rupture of the integument
is as described above and in all it occurs at this same place.

DetacJimcnt of the Muscles. - -The detachment of the muscles
does not begin or progress far until after the formation of the

FIG. 10. C. feinornta. Section of the integument showing the relations of the
muscles to the cuticula and the method of the insertion of the muscles. When ecdysis
occurs the muscles break from the cuticula along the line AB.

exuvial suture and is then accomplished by a few violent con-
tractions aided no doubt by the corrosion of the tendinous por-
tion of the muscle exposed beyond the surface of the hypo-
dermis.

The normal insertion of a muscle in insects to the cuticula is
shown in Fig. 10 of C. fciuorata, where the modified tendinous
ends are seen to go to the very outer surface of the cuticula. This

OBSERVATIONS ON COLEOPTERA.

189

seems to be general for insects, almost identical conditions being
shown by Holmgren in Lepidoptera.

When the muscle ruptures the break always occurs along the
line AB (Fig. 10) level with the surface of the hypodermis, and
the long drawn out tendinous ends of each muscle are again in-

Fn;. II. C. feinorata. Section to show the relation of the muscle fibers immedi-
ately after separation from the old cuticula and after the new cuticula has been partly
developed.

corporated into the new cuticula as it is formed, cuticula being
developed between the tendons by specialized hypodermal cells
and the tendinous ends themselves becoming chitinous.

These muscle changes, as far as they are related to the hypo-
dermis and cuticula behave exactly in Coleoptera as Holmgren
has described for the Lepidoptera and the same behavior prob-
ably is true for insects in general.

Cnticula Structures Removed and Replaced at Ecdysis. — Con-
siderable diversity of opinion prevails as to what and how much
of the chitin is removed at ecdysis in insects. I find in L. dcciin-
liucata that ecdysis takes place in the following structures:

i. Body Wall. - - Cuticula entirely removed, including the chi-
tinous covering of all hairs, spines and scales, sensory pits and
sense organs, including chitinous lenses of both simple and com-
pound eyes.

I gO W. L. TOWER.

2. Alimentary Canal. — Entire cuticula lining of both fore and
hind gut, including all cuticular structures in the mouth and the
linings of the ducts of all the bucal glands.

3. Trachcal System. — All tracheal lining of the main tubes
and the smaller trachea, but not of the tracheoles. It is possible
that the lining of these latter is not chitinous, or is very soft and
does not harden and therefore does not need to be removed.

4. Internal Supporting Stntctnrts. - -Tentorium, apodemes and
the lining of all ducts and glands, including the reproductive
organs or such as exist.

Conclusion. — The observations and figures which I have given
show that ecdysis is a more important and far-reaching process
in the development of insects than might at first be supposed.
Inadequate as are the observations herein recorded, they show
changes comparable to other periodic phenomena in animals, in
the existence of the periods of preparation, active ecdysis, recon-
struction, and quiescence. We cannot of course conclude from
this that ecdysis is in any way related to other periodic phenomena
in animals, but rather that there exists in all a similarity, due in
all probability to the existence of deep-seated laws of growth
which control all of these phenomena.

All Arthropods and especially insects cannot grow or further
develop without the process of ecdysis, owing to the hard resistant
non-extensible chitinous outer portion of the integument. Among
insects ecdysis bears a direct relation as far as the specialization
of the process is concerned, to the degree of specialization of any
particular groups. Thus, in the lower forms like the Campodea,
according to Grassi, there is a single ecdysis, but this is fragmen-
tary. The integument is shed in small pieces, while in the Col-
embola the chitin is shed in bits throughout life. " A species of
Peripatus found in the rain forests of Vera Cruz, Mexico, when
about one fourth of an inch long did not shed its cuticula at all
during its growth. From these primitive conditions of ecdysis
found in the most generalized of tracheates and resembling pos-
sibly in this function their supposed annelid ancestors, there exist
all degrees of specialization up to the most highly developed con-
dition found in Coleoptera and Lepidoptera.

The specialization of ecdysis which goes hand in hand with

OBSERVATIONS ON COLEOPTERA. 19!

the specialization of the integument consists essentially, as I have
shown, in the development of a mechanical strengthening of the
body wall during ecdysis by the rearrangement of the hypodermal
cells in a very precise bridge-like structure ; the arrangement for
softening and dissolving a part of the cast chitinous covering, thus
decreasing resistance ; and finally the provision of a sufficient fluid
lubricator to enable the animal to slip out of its old covering with
the least danger of rupturing the body wall, or distortion of the
appendages. In these three provisions are features of great utility
which present many striking adaptations to the habits and habi-
tats of individual species.

That ecdysis is a crucial period in the life of insect larvae is the
generally accepted belief, but how crucial is quite an open ques-
tion. During several years in which I have reared larvae of
Leptinotarsa for the study of their evolution I have found that
ecdysis is one of the phases on insect breeding needing most care-
ful attention. This is especially true in experiments wherein there
are used greatly changed conditions of existence, and the failure
of many experiments can be attributed directly to improper care
and surroundings during ecdysis. Even under normal conditions
each recurring cycle acts to eliminate from one to eight per cent,
of the larvae and the average per cent, of individuals in my experi-
ments with L. decimlineata, which cover eleven years, that were
killed during ecdysis alone, is 13.69 per cent. Often the per-
centage runs far higher, frequently lower, but even 13 per cent,
of a population is a huge death rate to be directly due to one par-
ticular process. The mortality from ecdysis occurs largely as
the result of failure to rupture the integument, due primarily to
its not becoming softened owing to the failure of the cuticula
solvents to develop. The larvae dying from this cause in insect
breeding are often passed over as belonging to the death rate of
a certain stage, while in reality it is largely the elimination due
directly to ecdysis. Comparatively few larvae die during ecdysis,
but about from I to 1.5 per cent, become deformed and die in the
following stage. Apparently none are killed during the recon-
struction stage, excepting by their enemies. If we knew the
whole history of say L. decimlineata, I think we should be very
near to the truth if we hold that in any given population 15 per

IQ2 W. L. TOWER.

cent, are on the average eliminated by the process of ecdysis and
the accidents which may occur along with it.

Ecdysis is, in fact, extremely important in the physiological
activities of insects, each recurring period being accompanied by
great disturbances in all parts and functions of the body besides
those described in this paper. Muscles, digestive apparatus and
excretory system are deeply influenced by the process, and in
some forms at least are the seat of developmental dangers occur-
ring also in rhythmic cycles.

As far as my experience and observations go ecdysis is a pecu-
liar process which periodically reduces various organs and parts
of organs to a condition strongly resembling an undifferentiated
embryonic condition from which are again built up old or new
characters by exceedingly rapid reconstructive processes, each
passing through in its development, stages seen in the develop-
ment of the same or similar earlier larval characters. This
repetition of stages in the development of colors or other char-
acters is solely due to the recurrent rhythmic processes necessary
to further growth and development. Ecdysis is a process of in-
tense importance in the life of the individual and generation and
one upon which natural selection can work most effectively. This
preliminary paper suffices to call attention to the importance and
deep-seated nature of ecdysis in insect economy, the general fea-
tures of the internal changes accompanying this process, some
of its most striking anatomical changes and its direct selective
value in the evolution of insects.

HULL ZOOLOGICAL LABORATORY,
THE UNIVERSITY OF CHICAGO,
February 6, 1906.

/ W. X. 4 //>/-//, 1906. No.

BIOLOGICAL BULLETIN

TWO TRANSITIONAL STAGES IN THE DEVELOP-
MENT OF CYCLOPS SIGNATUS,
VAR. CORONATUS.

A PRELIMINARY NOTE.

ESTHER F. BYRNES.

While studying some of the inhabitants of a pool of spring
water at Cold Spring Harbor, Long Island, during the months
of July and August, I was perplexed by the constant occurrence
of two comparatively large Cyclops that I was unable to identify
as belonging to any known species.

The prevalent adult forms in the pool were C. sigiiatus var.
tcnnicornis (annulicornis} and var. coronatus and C. scrntlatus.
That the forms in question bore no relation to C. scrntlatus was
quite evident. In certain characteristics they agreed with C.
sigiiatus but for some time I hesitated to associate them with
C. signattis because of constant correlations in structure, and
because the larger of the two forms was occasionally found to be
sexually mature.

The smaller of the two cyclops, Plate VII., combined the follow-
ing characteristics : the antennas usually contained ten segments,
but occasionally nine segments were present ; the rami of the
swimming feet were two-jointed, with an armature that indicated
immaturity as follows :

First Foot. Second Foot.

Outer Ratnus.
3 outer spines,
I apical spine,
I apical seta,
3 inner setce,

Inner Ramus
I outer seta,
I apical spine,
I apical seta,
5 inner setie,

Outer Ramus.

3 outer spines,
I apical spine,
I apical seta,
4 inner seUe,

Inner Ramus.
I outer seta.
I apical spine.
I apical seta.
5 inner setae.

'93

194 ESTHER F. BYRNES.

Third Foot. Fourth Foot.

Outer Ramus.
3 outer spines,
I apical spine,
I apical seta,
4 inner setae,

Inner Ramus.
I outer seta,
I apical spine,
I apical seta,
4 inner setae,

Outer Ramus.

3 outer spines,
I apical spine
I apical seta,
4 inner setae,

Inner Ramus.

I outer seta.
I apical spine
I apical seta.
3 inner setae.

The rudimentary fifth foot was two-jointed, having an almost
square basal segment bearing an immature seta on its outer distal
angle, and a slightly elongated distal segment bearing three
hairs.

The larger cyclops, Plate VIII., combined the following charac-
teristics : the antennae contained eleven segments ; the rami of
the swimming feet were three-jointed with an adult armature, as
follows :

First Foot. Second Foot.

Outer Ramus. Inner Ramus. Outer Ramus. Inner Ramus.

3 outer spines, i outer seta, 3 outer spines, I outer seta.

I apical spine, I apical spine, I apical spine, I apical spine.

I apical seta, I apical seta, I apical seta, I apical seta.

3 inner setae, 3 inner setae, 4 inner setae, 3 inner setae.

Third Foot. Fourth Foot.

Outer Ramus Inner Ramus Outer Ramus. Inner Ramus.

3 outer spines, I outer seta, 2 outer spines, I outer seta.

I apical spine, I apical spine, I apical spine, I apical spine.

I apical seta, I apical seta, I apical seta, I apical seta?

4 inner setae, 3 inner set?e, 4 inner setae, 2 inner setae.

The rudimentary fifth foot was two-jointed with an almost
square basal segment, bearing an immature seta externally, and
a distal segment, longer than broad, armed with two immature
spines and a median seta.

The study of a large number of these forms showed that both
forms were always found together; that they always accompanied
C. signatus, var. coronatns, having seventeen antennal segments ;
that in the absence of strong color markings and hairs they gave
evidence of being young ; and that they were, as a rule, sexually
immature, although the larger cyclops occasionally contained
ova. I became convinced of their larval character, as well as of
their relationship to C. signatns, and determined to test the cor-
rectness of the conclusion by breeding experiments. Accord-
ingly, a few fertile females with appended embryos, were isolated
in a carefully prepared aquarium. At the end of two months
this aquarium was examined for the desired forms, when nu-

DEVELOPMENT OF CYCLOPS SIGNATUS CORONATUS. 195

merous individuals in both stages were found to be present.
Later, one cyclops with eleven antennal segments was found with
numerous large ova in the ovaries.

The rearing of the young from the adults proves conclusively
that the forms in question are transitional stages in the life his-
tory of a form with seventeen segments in the antennae. The
experiments show also that C. signatus may become sexually
mature in the larval state.

That this sexually mature young was a true larva, and not a
form in which the growth of only the antennae had been retarded,
is indicated by the relatively small size of the individual when
compared with the adults, by the incomplete number of segments
seen in the antennae, and by the transparency of the skin and the
general absence of hairs and spines.

In the formation of the antennal segments there is a constant
sequence. The nine-jointed antenna becomes transformed into the
ten-jointed antenna by the division of the third segment from the
base into two small ones, and the eleven-jointed condition arises
from the ten-jointed, by the formation of a short second segment,
the permanent second segment of the seventeen-jointed antenna.
Occasionally the antennae contain twelve segments, but inter-
mediate conditions between the eleven and the seventeen-jointed
stages are not frequent. In Signatus, prior to the completion of
segmentation in the antennae, two greatly elongated segments
(the seventh and eighth in the eleven -jointed stage) break up
simultaneously into the required number of small segments, four
and three respectively, while the third segment breaks into two,
thus providing seventeen segments, the number in the adult.

In a former paper, " Heterogeny and Variation in Some of
the. Copepoda of Long Island," I was of opinion that " some
facts point to the probability that the Cold Spring Harbor
forms . . . are morphologically undeveloped." I am now con-
vinced that those cyclops combining nine, ten and eleven joints
in the antennae, with a two -jointed fifth foot, having a distal seg-
ment armed with three undeveloped hairs, are transitional stages
in the life history of C. signatus, although the variety is less
easily determined.

The study of various larval stages of C. signatus shows the rudi-

196 ESTHER F. BYRNES.

mentary fifth foot to be the most characteristic organ by which
the species can be easily identified even in very early stages.
Whether the hairs on the fifth foot are setose or spiny depends
in large measure on the age of the individual, and is of little
importance.

There are two generally recognized varieties of the species C.
signatus, namely, var. coroiiatits (C. fuscus of Jurine) and var.
tenuicornis or annulicornis, called C. albidus by Professor Marsh.

Both of these forms occur side by side and are regarded by
some authors, Herrick and Brady, as transitional stages of one
and the same form, while other authors regard them as distinct
varieties.

I do not wish at this time to express an opinion on the rela-
tionship of C. coronatns and C. tenuicornis, but the correlations
indicated in this paper for C. signatus, var. coroualus, can also be
observed in the forms described as C. signatns, var. tcmiiconiis,
in which the hairs on the inner margins of the stylets are lacking,
as well as the serrations in the hyaline plate in the sixteenth and
seventeenth antennal segments ; and in which the second segment
of the antennule, as well as the basal segment of the rudimentary
fifth foot, are relatively long, while in coronatns they are conspic-
uously short.

The armature of the fourth pair of swimming feet, often shown
in keys to the Cyclops, is by itself of little value as a means of
identification, since several distinct species agree in the armature
of this appendage, although they have no agreement in any of the
other swimming feet, nor in the larger outlines of the body. It
is to be regretted that fuller details of the Cyclops are not given
as a means of securing a rapid acquaintance with common forms,
and with their later transitional stages, which occur in great
abundance side by side with the adults, and, like the adults, are
sometimes found with eggs.

COLD SPRING HARBOR,

LONG ISLAND, N. Y., July, 1905.

PLATE VII.

C. signatus, coronatus.

Leitz camera drawings. Reduced one half.

FIG. I. An outline showing the proportions of the thorax, abdomen, and antennae
when there are but ten segments in the antennae. 3X3-

The abdomen has but three undifferentiated segments, to the last of which are
attached the characteristic caudal stylets.

The antennae reach to the posterior edge of the second thoracic segment, and
contain ten segments each. Only the right antennule is represented.

FIG. 2. The external ramus of the fourth swimming feet. All the rami of the
swimming feet are two-jointed. The proximal seta of the inner margin is conspicu-
ously separated from the remaining setse, which become the inner armature of the
terminal segment after the formation of the third joint. 3X7-

FIG. 3. The rudimentary fifth foot is characteristic of the species C. signatus.
The hairs are neither setose nor spiny. 3X7-

The internal fringe of hairs on the caudal stylets, the short, broad, basal segment
of the fifth foot, and the short second joint of the antennule indicate that this is the
larval form of C. siynatits coronatus.

o

FIG. 4. An antenna with nine segments.

FIG. 5. An antenna showing the third segment dividing to form the tenth segment.

FIG. 6. An antenna with ten segments fully formed.

198

BIOLOGICAL BULLETIN, VOL. X

PUTE V

PLATE VIII.

C. signatus, corona/us.

Leitz camera drawings. Reduced one half.

FIG. i. An outline showing the proportions of the thorax, abdomen and antenna;
when the antenna contain eleven segments. 3X3-

The abdomen has acquired four segments in which there is still little differentia-
tion. The caudal stylets remain unchanged.

The antennae have acquired a new segment, the small second segment from the
base of the antenna, making eleven in all. Only the right antenna is represented.

FIG. 2. The external ramus of the fourth swimming feet. All the rami of the
swimming feet are three-jointed, and in their armature, have reached the condition
of the adult. 3 X 7-

FIG. 3. The rudimentary fifth foot remains practically unchanged. 3X7-

200

BIOLOGICAL BULLETIN VOL. X

PLATE VIII

THE MALE AND FEMALE EGGS OF PHYLLOX-
ERANS OF THE HICKORIES.

T. II. MORGAN.

The predetermination of sex in the egg has been demonstrated
only where small male and large female eggs occur, and it is a
very striking fact, so far overlooked I believe, that in these cases
the difference in the male and female eggs is connected with the
development of degenerate males. In Dinopliilns apatris the
male is smaller than the female, is degenerate, and the sexual
organs appear to be precociously developed. In Hydatina senta
the male is smaller than the female and degenerate. In the
Plivllo.vcra of the grape and of the hickories the male is small,
wingless, without digestive tract, and the sexual organs, as I shall
show for one species at least, develop very early, so that the
spermatozoa are fully formed before the male leaves the egg.

In Dionpliilus it is believed that both the male and the female
eggs are fertilized, but in Hydatina and in the phylloxerans the
male and the female eggs are not fertilized. Sex in these cases
is, therefore, determined independently of fertilization and pre-
exists in the egg. In cases of this sort it would obviously be of
great interest to discover what conditions determine that some

o

eggs become males and others females. Is the difference, for ex-
ample, connected with a visible difference of the nucleus, or of
the cytoplasm ? This question I believe I am able to answer, but
the deeper-lying problem as to the causes that lead to the differ-
ence observable in the egg I have not fathomed.

The material for study was collected from the galls of the
hickories in the spring and summer of last year, and included five
or six species, of which three only will be mentioned here. The
typical life history of these phylloxerans is the following : The
fertilized egg, attached to the bark, hatches in the spring pro-
ducing the stem-mother, who migrates to the young leaves and
attaches herself to one spot on the under surface of the leaf,
which becomes the center for the formation of a gall within which
she becomes enclosed. She lays a large number of eggs inside

201

2O2 T. H. MORGAN.

the gall that are all of one size, and give rise to the individuals
of the second generation, which, developing wings in most spe-
cies, ultimately leave the gall. These migrants contain two kinds
of eggs, larger eggs that give rise to females, and smaller eggs
that give rise to males. The eggs are deposited upon the bark

»'

FIG. I. Phvlloxcra globosum. Polar spindles of male and female eggs. The
first figure (to the left) is from a large female egg; the second, from a small male
egg ; the third and fourth figures are from eggs whose size was not determined.

of the hickory. From them the male and female individuals are
hatched. These soon pair and the female lays her single, large,
winter egg on the bark of the tree. This egg hatches in the
following spring, and produces the stem-mother.

A few species have a somewhat different cycle, and in one of
them the male and the female eggs are laid within the gall itself.
Owing to this condition a large number of the eggs in all stages
of development can readily be collected. I have been fortunate
enough to obtain a species of this sort, viz., Phylloxera sp. ? and
have obtained an abundance of developing male and female eggs.
The life history of a species of this sort according to Pergande is
as follows : The stem-mother lays eggs that give rise to a genera-
tion of wingless forms corresponding to the winged migrants of
other species. These wingless individuals contain large and small
eggs which they deposit within the gall. From these eggs the
minute males and females emerge. In a few cases a winged in-
dividual— a migrant --is found with the wingless individuals.
It seems probable that the wingless condition is the secondary
one, hence the occasional appearance of winged forms.

The Male and the Female Eggs of the Migrants. — The eggs
are mature in the migrant before it leaves the gall, and the
polar spindle is present. It is a difficult matter to find the
spindle, and the chance is also small of getting one cut parallel
to its equatorial plate. Nevertheless, I have found a fair num-

MALE AND FEMALE EGGS OF PHYLLOXERANS. 203

ber of such cases in which the number of the chromosomes could
be counted with perfect accuracy. In Phylloxera globosnm there
are only six chromosomes in the polar spindle (Fig. i), which
makes it a very favorable object for study. In all eggs in which
the chromosomes could be counted the same number was found,

FIG. 2. Phylloxera sp. ? Left hand figure is a spermatogonial equatorial plate.
Right hand figure is from a somatic cell of a male embryo.

and this is true both for male and for female eggs. In this re-
spect there seems to be no difference between the two kinds of

eggs-

In Phylloxera sp. ? the number of chromosomes in the male

and female eggs seems to be twelve (Fig. 2). It is more
difficult to count so many chromosomes with accuracy, at least
in the eggs that I have so far seen ; but as the clearest cases ob-
served showed twelve chromosomes, and as in other eggs eleven,
or twelve, or thirteen seemed to be present, and since the num-

••:

FIG. 3. Phylloxera sp. ? Spermatocyte divisions. Equatorial plates. Upper
row, first spermatocyte divisions ; lower row, second divisions.

ber of chromosomes in the spermatocytes is definitely six (Fig.
3), there can be little doubt that the number in the polar
spindle is twelve. In Phylloxera cary<z-globnli there are twenty-

204 T. H. MORGAN.

two chromosomes present of very different sizes, as seen in the
figures (Fig. 4).

In passing, it is worth while calling attention to the very dif-
ferent number of chromosomes found in these species of the same
genus ; species in fact, that are so similar that they can only be
distinguished with great difficulty. It does not seem probable

••

FIG. 4. Phvlloxcra caryac-^lobiili. Polar spindle-equatorial plate of egg of
migrants.

in the lig-ht of cases such as these that the absolute number of

o

the chromosomes can be a matter of any special significance. If
the chromosomes are all composed of the same identical sub-
stance it is difficult to account for their constancy in number and
sizes in each species. If the chromosomes are different in com-
position, as the conditions just mentioned would seem to indicate,
the differences can scarcely be associated with differences in the
structures of the body, since closely similar individuals are pro-
duced in one case with six chromosomes and in the other with
twenty-two chromosomes. This question is one of the most
puzzling problems in the whole range of cytology at the present
time, and it would be unwise to draw any conclusion from the
meagre facts known to us. I wish only to indicate as a possible
view that the chromosomes may be different chemically from each
other, and yet at the same time this difference may have no con-
nection with differences in the form of the body.

Further History of the Chromosomes. --Starting with the stem-
mother the history of the chromosomes through the life cycle is
as follows : The number of chromosomes present in the polar
spindle of the eggs laid by the stem-mother is the somatic or
whole number. The number in the somatic cells of the embryo
that develops from this egg is also the whole number. The
number in the polar spindle of the male and the female eggs is,

MALE AND FEMALE EGGS OF PHYLLOXERANS. 205

as has been stated above, the whole number ; and this is true for
the somatic cells of both male and female embryos. In the sper-
matogonial cells the whole number of chromosomes is present,
but in the two following spermatocyte divisions the half, or re-
duced number, occurs. I have seen the spindle in one winter
egg, which seems to have only the half-number of chromosomes.
It is evident, therefore, that the complete number of chromo-
somes is characteristic both for somatic and germinal cells
throughout the life cycle, except for the winter egg and the sper-
matocytes, where the reduced number occurs preparatory to fer-
tilization. After fertilization the number %of the chromosomes
would be the same as the whole number. Thus it is evident
that there is no reduction in the number of the chromosomes in
the parthenogenetic egg, but there is in the winter egg. The
former is not, the latter is, fertilized. The results for Phylloxera
are in these regards parallel to those that Stevens has recently
obtained in the aphids.

The Cytoplasm of the Male and Female Eggs. — Since no
discernible difference was detected between the chromosomes of
the small male and the larger female eggs, I examined with some
care the cytoplasm of these two kinds of eggs. In the male
eggs there is less yolk and the center of the eggs is occupied by
a clear mass of cytoplasm, while in the female egg there is more
yolk and no central cytoplasm. I have not observed any other
differences, and even those just noted can not be made out with
certainty for all eggs, but represent the extremes of the series.
Double and triple stains gave no differences. that I could detect.
In the ovaries the eggs showed no differences. In Dinophilus,
according to Richard Hertwig, the large female eggs are formed
by the union of several cells. It would be difficult to detect such
a difference if it existed in the case of Phylloxera, since the ovarian
eggs appear to be fused together at their inner ends, and each egg
as it leaves the ovary remains attached by a cord of protoplasm
to the fused center of the egg-mass.

There are striking differences in the development of the male
and the female embryos, and these I have followed in some detail,
but can not present the results at this time. One important dif-
ference must however be alluded to, namely, the precocious

2O6 T. H. MORGAN.

development of the relatively enormous reproductive organs of
the male. The testis is present as a large mass of cells at the
time when the blastoderm is first laid down, and the spermato-
gonial divisions occur at this time. The two spermatocyte divi-
sions occur when the embryonic plate is forming, and before the
fibers appear in the nervous system. The size of the mass of
spermatocyte cells is very large compared to the rest of the
embryo, and it fills the greater part of one half of the egg. The
spermatozoa are fully formed at the time when the embryo is
developed.

The precocious development of the large testicular mass in the
male egg suggests that a preexisting mass of cytoplasm from
which the testis develops may be present in the egg. The cen-
tral mass of cytoplasm might supply such material, but I have not
been able to make out any connection between the two. The
result is due rather to the more rapid development of certain em-
bryonic cells than to the presence of any large cytoplasmic mass
of preformed material. The question arises whether we are to
regard the large eggs that produce the females as doing so sim-
ply because they are large, or are they large because they are
already female eggs ? The latter alternative seems the more
probable. In the light of certain recent experimental results,
more especially those of Driesch and of Godlewski it seems
highly probable that the early development of the embryo is due
almost entirely to cytoplasmic influences. If this is true also in
the case of the eggs of Phylloxera, then I think we may safely
ascribe the difference in the size of the male and female eggs to
the difference in the kinds of cytoplasm that are present when the
egg is fully formed, so that the immediate determination of the
sex is a cytoplasmic phenomenon. Whether this cytoplasmic
difference can be traced to a preexisting cytoplasmic basis, or to
nuclear influence, or to the influence of external conditions is
quite unknown, but in the absence of any nuclear differences it
seem questionable whether we should assume that such exists.

THE ORGANIZATION AND EARLY DEVELOPMENT
OF THE EGG OF CLAVA LEPTOSTYLA AG.

CHAS. W. HARGITT.'

A recent paper by Harm ('02), on the development of Clava
sqnamata describes several features so unlike conditions which I
have found in the American species, that it has seemed worth
while to study anew, and in some detail, the entire life history,
including also some interesting peculiarities of the organization
of the egg not hitherto given any degree of attention among
ccelenterates.

The general facts concerning the morphology, habits, etc., of
Clava leptostyla Ag. were described long ago by Professor Agas-
siz ('62) with sufficient fulness to obviate the necessity of any
considerable details as to these points. Allman ('71) has likewise
given a somewhat similar account of C. sqnamata, though lacking
in embryological details.

The following account is based upon studies carried on at
Woods Holl during some two summers, chiefly upon living ma-
terial, though for the details of cytology specimens were killed
by several standard methods, among which were corrosive-acetic
acid, picro-acetic, Carnoy's solution, Petrunkevitch solution, and
corrosive formalin. Upon the whole, the first two of the reagents
named gave by far the better fixation. As will be noted in an-
other connection, for certain cytological results the picro-acetic
solution was exceptionally good.

The most common habitat of the species in this vicinity was
found to be the fronds of the common rock-weed, Fnciis nodosns.
Occasionally I have taken it from the piles of docks, or similar
situations. It occurs most frequently in shallower harbors, or
tide-pools, and is frequently found at low tide quite exposed, ex-
cept as the moist vegetation may afford a measure of protection,
lying thus for hours subject to the desiccating influences of sun
and air, and apparently without serious injury. Of course, this

1 Contributions from the Zoological Laboratory, Syracuse University.

207

2O8 CHAS. W. HARGITT.

vicissitude is common to many organisms of similar habitat, yet
few are so little protected by some skeletal feature as is this
wholly naked hydroid. The breeding season at Woods Roll is
apparently during June and early July, though colonies may be
found in seemingly thriving conditions both earlier and later.
Whether Clava like Pcnnana, Eudendrium and others, has its
phases of decline and recovery I have not been able to determine
from actual observation, though the fact that at recurring inter-
vals it appears in the same locations would seem to support this

view.

ORIGIN AND GROWTH OF THE EGGS.

Weismann ('83) was the first to give any direct attention to this
problem in Clava. According to him the ova arise in the ento-
derm in close relation to the supporting layer, though suggesting
that perhaps the most primitive stages may have an ectodermal
origin.

On this point Harm ('02) is quite specific in claiming their
origin in the ectoderm, and their later migration into the ento-
derm, where in the region of the gonophores they become defi-
nitely differentiated and complete their development. This author
even goes so far as to suggest that they may be distinguished in
the planula as primitive germ cells, --" urkeimzellen " (p. 47,
Fig. 48).

Of course, I am not prepared to discuss the matter so far as it
relates to C. sqnaiuata, not having seen this species, but so far as
it concerns C. leptostyla I have no hesitation in saying that eggs
probably never arise in the ectoderm, but always in the entoderm
of the peduncle of the gonophore, or in that of the polyp very
near the base of the gonophore. In thousands of sections studied,
both by myself and by several of my students, there has been no
exception to this statement of fact. Clava, like other hydroids,
has its breeding season, during which the germ-cells are ex-
tremely abundant, and at other times these cells are either entirely
absent or very scarce. Again they do not appear except in direct
relation to the forming gonophores or in that immediate region.
In fact in Clava leptostyla the morphology of the gonophores and
their development as dense, bud-like clusters from a single
peduncle to which they are attached by narrow pedicels, make it

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 209

almost certain that the vast majority of all egg-cells arise in the
gonophores themselves as, of course, is the case in almost all
free gonophores, or medusae, and as is invariably the case with
the spermatozoa.

Therefore, while it may not be impossible that " urkeimzellen >:
should perhaps exist in undifferentiated stages, still the proba-
bility is so extremely remote as to render doubtful to a degree
any but the most thoroughly substantiated claims.

It may be stated in passing, that the gonophores of C. lepto-
styla are extremely degenerate, hardly more, indeed, than sporo-
sacs, yet it is possible to distinguish rudiments of medusoid struc-
tures. They originate as buds of the hydroid, involving both
ectoderm and entoderm, and also supporting layer. Occasion-
ally this lamella seems to partially disintegrate at the terminal
portion during the outgrowth of the peduncle. The gonophores
remain entirely closed except at the time of rupture by the escap-
ing planula. From the primary peduncle secondary pedicels
arise, forming a racemose-like cluster, within each of which from
one to four eggs may develop, though the usual number is two
or three.

There seems to be considerable variation in the size of the eggs
in various specimens and in the eggs of various gonophores of
the same specimen. On this point it had occurred to me that
perhaps the number developing in a given gonophore might
naturally have some influence, but after comparing a considerable
number, whether growing singly or in clusters I have not been
able to convince myself that such is the case. I am rather in-
clined to believe that more depends upon the. start a given egg
may get in growth, and perhaps the state of nutrition in which
the given specimen may be at the time, than any other factors.

As a rule eggs which grow singly at the distal end of a gono-
phore are more nearly spherical, and in consequence I find the
cleavage of such eggs much more regular and symmetrical than
in cases where two or more are found in the same gonophore and
approach maturity at about the same time. This point may be
considered in more detail in a later connection under the subject
of cleavage.

Growth takes place, as in most hydroids, quite rapidly. With

2IO CHAS. W. HARGITT.

the origin of the gonophores the eggs appear in considerable
numbers in the entoderm of the peduncle, as previously stated,
and with the later budding of the gonophore bodies either migrate
into them, or as seems to be the case in many instances, originate
directly from the entoderm of the spadix or lateral walls of the
gonophore. At first they have the characteristic aspects of ordi-
nary ovocytes, namely, a very large germinal vesicle, with char-
acteristic chromatin network, a comparatively small proportion of
cytoplasm, which is more or less homogeneous in texture, and
staining quite uniformly with any of the ordinary plasma stains.
Evidences of growth are first indicated by the rapid increase in
the mass of the cytoplasm, while that of the nucleus for a time
remains apparently unchanged, though later also increasing in
mass likewise, though to a much less degree. Nutrition of the
ova is at first, indeed throughout so far as I can determine, by
direct absorption from the entoderm of the spadix, or to a less
degree also from the entoderm of the lateral walls of the gono-
phore. I find no evidence of the absorption of supernumerary
ovocytes involved in the matter of nutrition, and in this respect
C. leptostyla appears to differ somewhat from C. sqnainata. Ac-
cording to Harm there would seem to be involved both these
processes. He says that by the direct assimilation of yolk-like
granules from the entoderm cells of the hydranth, and by osmosis
from the walls of the gonophore the egg is nourished, and that
furthermore, the youngish egg-cells are also nourished by the ab-
sorption of ovocytes (op. cit., pp. II, 12).

The presence of the yolk-like particles to which he refers I
have also recognized in the entoderm of the hydranth body at this
period. They resemble in all essentials the pigmented yolk
granules later found in the fully grown egg, but I have found no
evidence that they are ever directly absorbed by the young egg.
On the other hand there is ample evidence to the effect that they
are gradually broken down and probably liquefied, in which con-
dition they may be easily transferred to the gonophores and ab-
sorbed by the young eggs. In an earlier paper ('04) I have
directed attention to similar phenomena in the growth of the
eggs of Pachycordyle, and it undoubtedly occurs in many
others.

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 211

HERMAPHRODITISM.

An interesting feature in the reproduction of this hydroid is
the fact that occasionally individuals, and perhaps colonies, are
found in which gonophores contain both eggs and spermatozoa.
Figs, i, 2 ; Fig. 5, PI. IX., show various phases of this somewhat
anomalous condition. As will be observed, the elements are in
various stages of development,
some of the eggs well along toward
full growth, and spermatozoa like-
wise well advanced. In most
cases the condition shown at Fig.
5, PI. IX., was the prevalent one,
namely, where one half of the
gonophore bore sperms and the
other half an egg. In a few cases,
however, a well developed egg
was found on each side nearly or
quite surrounded by sperm-
cells.

Of course, hermaphroditism in
itself is nothing strange among
animals, whether high or low.
Even among hydroids it is quite
familiar in the common Hydra,
though here it is not common to
find both organs in active function
at the same time on any given in-
dividual. I have also found a sim- FIG. I. Longitudinal section
ilar condition in Hydractinia, and throu§h youn§ g°i>°ph°re, showing

at b the development of spermary, at

Bunting ( 94), has likewise figured e and ^mixed gonads.
a single case though without giv-
ing any details concerning it. In the whole of his extended
researches on the " Origin of Sex Cells in Hydromedusse ''
Weismann makes no mention, so far as I have observed, con-
cerning hermaphroditism. It would seem somewhat remarkable
that he should not have observed some indications of such a con-
dition if it were at all common. Indeed, though having found
repeated cases of it in Clava, I am disposed to consider it as a
rather rare phenomenon in this group.

b

212

CHAS. W. HARG1TT.

One other feature in connection with the subject must be noted,
namely, that among the several cases, the ova were found in every
case in distinctly male gonophores. Among hundreds of female
gonophores examined there was not the slightest evidence of male
elements among them. Figs. I and 2 show sections through
two hermaphroditic gonophores. At b Fig. I is developing a

FIG. 2. Cross-section of hydranth showing development of gonophores. b, her-
maphrodite gonophore ; c, early stages of same ; d, spermary.

typical male gonad, while at c and d are shown what are evidently
destined to be mixed gonads. Similar conditions are also shown
in Fig. 2, which is a camera sketch of a section across the entire
hydranth and gonophores.

It has long been known that among actinians a form of herma-
phroditism, involving successive sexual rhythms, known as pro-
tandry or protogyny, occurs. Duerdon ('04), in his studies of
West Indian corals has found warrant for believing that protogyny
is the predominant condition, since " spermaries have never been
found alone, but always associated with large numbers of ova ;
on the other hand, many polyps have been found with ova alone,
often few in number, as if sexual maturity were but beginning."
At the same time he quotes from Mr. Stanley Gardiner, who
having studied a large number of developmental stages in Flabel-

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTVLA. 213

////// rnbnun, has been able to show that in this form spermaries
arise first on the mesenteries and that ova appear later, when the
production of sperm acini ceases. "The ova grow enormously
with the final result that the mass becomes entirely female, con-
sisting, usually of two or three large ova. flattened on their sides
against one another and occupying the whole area of the former
testes."

Hermaphroditism is also known among Scyphomedusae, Wright
having described in some detail the chief features in the case of
Chrysaora liyoscclla. He says, " Large individuals are herma-
phroditic, but smaller ones are found which are unisexual, the
male or female element being suppressed. Small Chrysaoras
(about four inches in diameter), have no ovarian bands in their
pouches, which only contain masses of the grape -like bodies
(testes), and tentacles before mentioned."

This would seem to imply that here again we have protandrous
hermaphroditism, the spermaries developing first, and later the
ovaries. Incidentally, it may be observed in passing, that this
author in his account of the development of the eggs of this
medusa was much impressed with the absence of any germinal
vesicle. "The ova of Clirysaora liyoscclla do not present, at any
stage, a trace of the germinal vesicle, — objects which are so
readily detected in the ova of other polypoid Zoophytes."

Haeckel ('79), who has also studied the development of the
medusa, was able to confirm Wright's account as to hermaphro-
ditism. " In Uebereinstimmung mit letzteren habe ich gefunden,
dass junge Chrysaoren rein mannlich vorkommen, solche mittle-
ren Alters meisters Hermaphroditen sind, und endlich ganz alte
Thiere meisters nur weiblich sind, oft noch mit Ueberresten
mannlicher Organe."

So far as I have been able to discern in the case of Clava, there
is no evidence to indicate the operation of either of these oscil-
lating phases of sexualism. I have specimens taken at all times
of the breeding season and have found no tendency toward the
one or the other. I am rather disposed to regard it as an
expression of a mutative impulse, in response to which in other
forms, such as those already cited, these interesting features
became established.

214 CHAS. W. HARGITT.

ORGANIZATION OF THE EGG.

It has long been known that simple though the egg may be
it must, nevertheless, be regarded as potentially highly complex.
Concerning the early views of His, Whitman, Flemming and
other earlier investigators, no attempt will be made here to give
special citations. The later experiments of Roux, on the devel-
opment of the frog's egg, supplemented by similar experiments
by Driesch '95, served to emphasize still further the general view
here stated.

The still later investigations of Wilson ('03, '04), and Conklin
('05', 'o52, 'O53), have given a new impulse to researches along
this line, and have clarified and measurably harmonized the con-
flicting views of earlier observers.

So far as I am aware, no one has shown any evidence of any-
thing of a similar character in the organization of the eggs of
codenterates. In commenting on the account given by the pres-
ent writer of the cleavage of the egg of Pcnnaria Conklin ('o53),
has suggested the probability that predetermining factors must
be present in some form. " Even in such eggs as that of Pen-
naria it is certain there must be determining factors somewhere,
if not in the cytoplasm then in the nucleus, which determines that
the egg shall develop into a pennaria rather than into some other
animal ; it is further evident that these determining factors must
be present in the cytoplasm at a relatively early stage, if not in
the very beginning of development " (p. 215).

Hence in my studies on hydroid eggs rather particular atten-
tion was given to this point with the hope that some evidence
might be found for or against the views in question. In my work
on Pcnnaria ('O43), while no direct evidence of cytoplasmic dif-
ferentiation was found in the living egg, there was found in some
cases after fixation by certain reagents, more particularly the
picro-sulfuric solutions, what seemed to be a stratified, or con-
centric arrangement of the cytoplasm. At first I was inclined to
regard this as probably significant of such locally differentiated
matter as might go to form the ectoderm. But since this condi-
tion was not found to be constant it was regarded as probably
an artifact, due to the action of the reagent. A recent examina-
tion of my earlier preparations has not suggested any change of

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 215

this opinion. Indeed, as pointed out before, the extremely er-
ratic and anomalous behavior of these eggs during cleavage would
seem to render extremely difficult, if not impossible, any prede-
termining factors in either cytoplasm or nucleus, whose influence
could be maintained during the varying and indeterminate process
of development. While we may readily admit that, as Conklin
has suggested, there must be factors present which determine
that the egg shall develop into a Pcnnaria, still this does not
compel the conclusion that therefore they must be definitely
localized. It is simply a matter of heredity ; and if it be true as
generally contended that this is a problem of chromosomes ; and
if, as I have shown in the case of both Pennaria and Eudendriuin,
the chromatin may be more or less dispersed throughout the
entire cytoplasm during maturation and early cleavage, then
definite localization in one or the other is not involved in Conklin's
sense of the term. However, it must be regarded as a question
of fact, and so far as evidence exists in the present case it would
seem to be opposed to the theory of localization.

During June and July of the past summer I carefully studied
the living eggs of Clava, and with this point still clearly before
me. As pointed out in an earlier section, the eggs originate in
the entoderm of the gonophore, and grow by direct nutrition
derived from the cells of that tissue. A study of the eggs in
various stages of growth revealed the appearance at a certain
stage of development of a delicate, bluish pigment, which grad-
ually accumulated in amount as the eggs approached maturity.
This was carefully observed in the living specimens and has since
been studied in sections after a variety of fixations. At first the
pigment makes its appearance in the immediate region of the
nucleus, about the time that body takes its place at the outer
periphery of the egg. This is shown in Fig. I, Plate I. From
the nuclear region the pigment extends as a crescentic disc out-
ward, forming later a peripheral zone which finally extends over
the entire egg, though this rarely occurs until cleavage has made
some progress.

An interesting fact observed in this connection was that the
amount, or at any rate the color-intensity, of the pigment differed
considerably in different specimens. This was particularly the

2l6 CHAS. W. HARGITT.

case in colonies which had been kept in the aquarium for a few
days. In these the pigmentation was appreciably less intense
than in specimens freshly collected. There were, however, not-
able exceptions in this respect among various specimens under
natural conditions. The same has been observed in Pennaria,
and is, indeed, a fact more or less well known in many animals.
Conklin ('05', p. 13) has cited similar cases among ascidians ob-
served by VanBeneden and Julin ; in some species two very dif-
ferently colored eggs being produced, one yellow, the other gray.
Both are said to develop normally and in the same manner, giv-
ing rise to larv?e whose entoderm cells are of the same respective
colors.

In the case of C/ava there is no such distinction as this, though
the presence of more or less pigment has apparently no effect
upon the normal development of the embryo. In another point
there is also some measure of similarity, namely, in that as
development proceeds, the pigment which was at first distinc-
tively peripheral in position, seems later to become transferred
to the entoderm of the larva. This, however, as will be seen
later, is due not to any shifting of this matter from one region to
another, but simply its resorption by the more rapidly developing
ectoderm, while as yet the entoderm is only partially differ-
entiated.

Whether this pigmentary zone formed a definite germinal area,
and its gradual development was an expression of the differen-
tiation of this area, or whether it might not be simply the results
of cytoplasmic metabolism associated with the formation of yolk
substance, or whether it might not, perhaps, be in some way asso-
ciated with the phenomena of maturation, seemed for a time some-
what uncertain. A study of the matter more critically in its
cytological aspects soon sufficed to discredit the last alternative,
namely, that it .was in any way associated with maturation. Fur-
thermore, the appearance of the pigment was too early in the
history of the egg to involve the operation of any maturation
phenomena.

Again, several series of facts conspired to discredit the prob-
ability of the first alternative, namely, that it was in any sense a
differentiation of germinal substance. Of these the following may

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 2 17

be particularly mentioned : (i) The entire absence of any correla-
tion between the pigmentary zone and the course of cleavage ;
(2) the continuous development of pigment, even after the com-
pletion of cleavage, and after the ectoderm has been clearly
established ; (3) the staining reactions of the granules after hard-
ening is comparable, point by point, with that of yolk granules ;
(4) finally, the granules are gradually reduced and resorbed with
the growth of the embryo. It is in this way that their disap-
pearance from the ectoderm above referred to is to be explained.
I think that we may therefore conclude that, at least so far as
the problem is concerned with Clnva, its solution is unquestion-
ably in the negative. If further warrant be needed for this view
I believe it will be found in the later history of these pigmentary
granules as given in the following account.

ORIGIN AND GROWTH OF PIGMENTARY GRANULES.

As already noted the origin of these pigmentary granules is in
the immediate region of the nucleus, and about the time this body
reaches the outer periphery of the egg. At first they are of ex-
tremely small size, about 0.5 micra in diameter, later growing and
reaching in some cases a diameter of 3 to 3.5 micra. For some
time they were entirely overlooked in prepared material, owing to
improper fixation. Only after fixation with picro-acetic (or to
less degree with Petrunkevitch) solutions and staining with iron-
haematoxylin on the slide were they adequately differentiated so
as to be readily studied. Wilson gives a similar account of this
technic in reference to the eggs of Dentalinin (cf. Jr. Exp. Zoo/.,
Vol. I., p. 9, Figs. 10—13, explanations).

As the granules continue to grow larger there may be distin-
guished within their substance what appear to be vacuoles,
usually a single one within each granule, occupying an eccentric
position. In some respects they exhibit nucleolar-like features,
especially in their staining reactions and in their vacuolation.
Montgomery ('98) has made similar observations on nemerteans
and believes they indicate some sort of genetic relationships be-
tween nucleoli, yolk balls and granules. My own observations
have not seemed to confirm this last point, though in the case of
Hydra there are not lacking evidences which I think would give

2l8 CHAS. W. HARGITT.

strong confirmation of Montgomery's views. This point may
have further consideration in connection with observations upon
the history of the nucleolus.

Concerning the real nature of these granules there arises the
query, are they katabolic products, associated with some vital
wastes incident to the cytoplasmic activities of growth, or are
they not rather anabolic in character, highly nutritive proteid
bodies, analogous to yolk matter and of similar import ? The
latter is by far the more probable view, though there are points
of difference as compared with the usual formation of such nutri-
tive matters. For example, in most hydroid eggs which have
come under my observation the development of yolk granules
has no appreciable relations to nuclear influence, and seems to be
for the most part developed and deposited chiefly at the vegetal
pole of the egg ; while in the present case, as has been shown,
they seem to arise and develop chiefly in the nuclear area and
only at a late period are found at the vegetal pole. However, I
am inclined to believe that this is not a serious difficulty. It will
be observed that the vegetal pole lies in immediate contact with
the spadix of the gonophore, and that the reception of nutritive
matter by the growing egg is from this source. Of course, this
nutritive matter is in the form of liquid, and so long as the egg
is continuously receiving it in this way there is no occasion for
further metabolism into the more solid reserve of yolk. The
remoter animal pole where such reserve would first be needed
begins the anabolic process first, and with the gradual suppression
of the nutritive activities of the spadix the process of proteid anab-
olism would extend into that area.

According to Harm ( op. cit., p. 19), the development of this
yolk matter in C. squaniata presents some rather sharp contrasts
as compared with the above. For example, instead of develop-
ing more or less gradually, and spreading from the nuclear region
over the surface of the egg, he finds it arising somewhat suddenly
and equally throughout the entire egg. " Bald nachdem die
Eizellen Glockenkern erreicht haben, beginnt in ihnen die Dotter-
bildung, die an alien Stellen zu gleicher Zeit einsetz." He also
finds that before the formation of yolk granules the eggs are red-
dish in color. " Wahrend die lebende Eizelle vor der Dotterbil-

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 2 19

dung rothlich erscheint, zeigt nacht derselben eine blaugraue
Fahrbung." This point is confirmed in the case of C. leptostyla,
in which essentially the same process takes place.

Perhaps a few words may be added as to the significance of the
development of pigment in connection with these yolk granules.
I have in recent papers ('O41, 'O42), submitted certain views as to
this subject and it may suffice in general to refer to those discus-
sions. In the former it was said that pigments in organisms
might appear under three aspects : (i) Those directly serviceable
to the organism, as in chlorophyll, haemoglobin, etc. ; (2) as
waste products, which embrace probably the more numerous of
organic pigments, such as guanin, melanin, etc. ; (3) as reserve
products, of which the lipochromes are typical. In all probability
the various pigmentary matters found in eggs belong to the third
of these classes. And here undoubtedly should be classed the
pigmentary granules of Clava, and other similar pigments of
hydroid ova. In the second paper attention was directed to a
special case, that of Pacliycordyle, already referred to, in which
one may trace the various stages in the growth of the egg and
the formation of the pigmentary bodies. Here as in Clava there
can hardly be reasonable doubt that the process is a gradual and
progressive anabolism, so far as the granules themselves are con-
cerned, but it must still be a somewhat open question as to the
exact relation of pigmentation thereto. May it not be probable
that here, as in many of the more active phases of metabolism in
which pigments are more or less evident expressions of excretory,
or waste products, the pigment itself, though associated with
anabolic activities, is an expression of the correlative process of
catabolism ? In other words, that even in those constructive
processes involved in the storage of reserve matters, whether as
proteids, fats, or whatever they be, there is involved the insep-
arable process of energy bestowed, and that as one of the signs
of such energy its imprint is left in these pigmentary elements ?
Such I am inclined to believe is what actually happens. And as
the nature of these processes differ more or less in various
organisms so the pigmentary signs of waste will likewise differ.
Hence the purple pigment of the eggs of Clava, the pinkish of
Pennaria, the reddish of Eudendrium, etc.

22O CHAS. VV. HARG1TT.

MATURATION.

Concerning this phase of development there is comparatively
little to be said in the case of Clava. I have studied as critically
as the nature of the egg would allow the behavior involved in
maturation both in living and preserved material, and in a very
large number of preparations, but with almost wholly negative
results. That is, I have found the phenomena to be so obscured
by the opacity of the cytoplasm, or by the pigment matter in the
yolk, or as seems to me a still further probability, namely, the
extremely fugitive character of the phenomena, as to render them
indistinguishable. I have had occasion to emphasize this matter
in several earlier papers dealing with the subject. The observa-
tions made upon the eggs of Pennaria have been duplicated,
almost point by point in the present case. Of course, in the egg
of Clava there is the added di c'ulty that all the phenomena
occur within the closed gonophore. In sections, however, this
fact ought to offer no serious obstructions to their detection, yet
the results, as in the former, are quite as uncertain and in most
cases absolutely lacking.

In the case of CLwa squainata, Harm (op. cit., p. 23) describes
the phenomena in some detail, and gives almost diagrammatic
drawings of the several stages. However, as will be further
shown in connection with the cleavage phases, there are so many
points of difference between these species that the ova may per-
haps belong to very different classes so far as their character and
texture are concerned. From the fact that in several particulars
I have been able to confirm the observations of Harm it does not
seem probable that they are so greatly different as might be im-
plied.

I shall briefly describe the principal features which it has been
possible to certainly determine, leaving others open to further in-
quiry or study.

As pointed out in an earlier section, about the time the ova
approach full size the nuclei are to be found close to the outer
periphery as will be seen in several of the accompanying photo-
graphs. It will also be observed that the eggs occupy closely
the entire space of the gonophore, and that, therefore, the nuclei
in coming in contact with the gonophore wall become more or

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTVLA. 221

less flattened. It is at this stage that the first indications of
maturation becomes apparent, namely, the shrinking and gradual
disappearance of the nuclear membrane. This is particularly
well shown in Fig. 2, PI. IX. About the same time, or in some
cases slightly before, there is also an evident dissolution of the
chromatin network, unaccompanied by any indications of chro-
mosomes. In this respect these eggs closely resemble those of
Endcndriuin and Pcnnaria. The most painstaking attempts to
differentiate these bodies by staining operations have as fre-
quently failed. It seems to me we are forced to the conclusion
that the appearance of dissolution is indeed a fact, and that at
this stage there is a general dissipation of chromatin, and per-
haps other nuclear matter, into the cytoplasm.

Behavior of the Nucleolus. — Usually at about this stage marked
changes take place in the nucleolus. In several instances it has

o i

been found to migrate bodily from the germinal vesicle into the
cytoplasm where it is gradually dissipated and probably assimi-
lated. At this point I find my observations closely in accord
with those of Harm (op. cit., p. 24). " Wahrend also hier der
Nucleolus in toto vor der Polkorperchenbildung aus dem Keim-
blaschen heraustritt, urn vom Eidotter aufgenommen und resor-
birt ZLI werden, verbleibt er in anderen Fallen in demselben und
zerfallt dort in mehrere Kugelchen."

Just prior to this migration of the nucleolus into the cyto-
plasm it was found to show varying degrees of vacuolation, in ad-
vanced stages of which it was often seen to partially shrink and
collapse, as if there had been a loss of nucleolar substance.
Something of this may be observed in several of the photographs
already referred to.

Conditions very similar to these I have elsewhere described in
connection with other hydroid eggs (op. cit., O44, p. 562). Mont-
gomery has cited many similar features which had come under
his own observations, as well as observations of a like character
made by many others. Those who are particularly concerned
will find his discussion exceedingly interesting and suggestive, as
well as including a valuable summary of evidence bearing on this
problem. I may say, however, in passing that I have failed to
find any indications in the present case of the metamorphosis of

222 CHAS. W. HARGITT.

the vacuolated portions of the nucleolus into yolk granules, such
as Montgomery has described. On the other hand the whole of
the nucleolar substance seems to be directly dissipated through-
out the cytoplasm and indistinguishably assimilated by it.

Polar Bodies. --With the phenomena already described there
have been found in a few cases what seemed to be polar bodies.
But it was impossible to distinguish the presence of any mitotic
mechanism. In every case the nuclear matter was devoid of any
trace of chromatin, or, as just suggested, mitotic figures. In a
few cases the formation of these bodies was observed in living
eggs, and these showed essentially the same features. The nu-
clear matter being in close contact with the gonophore wall,
small rounded portions seemed separated from the larger part,
and could be distinguished close under these retaining mem-
branes for only a short time, when they gradually disappeared,
apparently resorbed, as I have suggested in former studies upon
Etidendriiiin and Pennaria. In this respect also my observations
correspond with those of Harm, though I have in no case been
able to confirm his account of the phenomena of mitosis. " Die
Resorption der Polkorperchen durch die eizelle sehr schnell
vor sich gehen, da ich dieselben zur Zeit, wenn der weibliche
Pronucleus besteht, nicht mehr habe nachweisen konnen."

The results in the present case, as in those of Eudendriinn and
Pennaria, already referred to, as well as in certain others, serve
to suggest the query whether, indeed, there may not be great
variation as to phenomena of maturation, even perhaps to the
extent of the suppression of the more conspicuous aspects asso-
ciated with it in higher forms. It may be regarded as the wild-
est biological heresy to even remotely suggest that in some eggs
these physical phenomena might be entirely absent, but such an
impression has grown upon the present writer for some time and
increases with each further case, such as that under considera-
tion. If we may have normal nuclear division without mitosis
in the early embryonic history, and this I believe to be fairly be-
yond doubt ; and if we may have differentiation without cleavage,
now likewise admittedly true ; and if, furthermore, we may have
prior to either of these phenomena in development the dissipa-
tion of both chromatin and other nuclear matter throughout the

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 223

cytoplasm, another fact of which I can no longer doubt, then it
does not seem a far call to the assumption that the reduction
phenomena of maturation may well be accomplished without any
of the complex and spectacular processes of mitosis. Why should
the egg go through with the physical process of extruding the
polar globules if they are to be immediately resorbed by the
cytoplasm ?

I do not overlook the circumstance that the opacity, and in
the present instance pigmentation, of the egg may establish an
obstacle so formidable as to render accurate observation ex-
tremely difficult, as I have previously admitted, still the fact re-
mains that carefully preserved and stained sections of hundreds
of eggs have failed to afford convincing evidence of the presence
of these bodies in distinguishable form, except in doubtful cases
already referred to. It would seem somewhat remarkable that
there should be all the differential results of staining upon other
cell features while it should be uniformly lacking in these, usu-
ally among the most readily demonstrated.

Yet another feature bearing on this point will be found in the
details of cleavage, during which there is every appearance of
spontaneous nuclear reorganization throughout large portions of
the egg at almost the same time, the last detail of which may be
followed. The facts will be described in a later section, though
it seems well to call attention to their significance in relation to
the matter under consideration.

CLEAVAGE.

In a general way cleavage in these eggs corresponds with that
of other hydroids having similar gonophores, such as Tubularia,
Hydractinia, etc. The earlier work of Ciamician ('79), and
Brauer ('91), on Tubularia mesembryantkemum, and the later
work of Allen ('oo), on T. crocea, afford good examples of the
type of cleavage here referred to. In a recent paper ('O44), the
present writer has briefly reviewed the results of Ciamician and
Brauer, and it is unnecessary to discuss here in any detail these
features. Suffice it to say that Brauer found what he regarded
as two rather distinct types of cleavage. The first more or less
regular and equal ; the second irregular and indefinite, involving

224

CHAS. W. HARGITT.

for a time an internal nuclear proliferation, followed later by the
spontaneous division of the cytoplasm into a corresponding num-
ber of blastomeres. While my own observations (op, cif.} did
not fully confirm those of Brauer they abundantly proved the
general proposition that cleavage is not uniform, either in mode or
progress. The work of Allen (op. cit.} showed even more con-
clusively the variation of this feature in T. crocea.

The same thing is true in the eggs of C. leptostyla. While
there is here much more uniformity than in either of the former
cases the range of variation is still very considerable, as a glance
at the various figures will abundantly show.

FIGS. 3, 4, 5. Camera sketches of phases of cleavage, from life.

The account of Harm (op. cit., p. 28) concerning cleavage in
C. squainata is in very marked contrast in its general aspects.
Both in his descriptions and figures there is a most striking regu-
larity and definiteness in the cleavage. He very briefly refers to
a condition which he considers exceptional and abnormal, but
which unless I greatly mistake must.be much more common than
he is disposed to think. It is quite similar to a condition referred
to by Brauer, and regarded by him as likewise exceptional, but
which I have showed to be more or less common.

Again, Harm describes in some detail what he regards as a
more or less definite rotation of the blastomeres during earlier

o

cleavage, in some measure comparable with phenomena familiar
in the eggs of annelids and molluscs. When attention is directed
to the fact that in these hydroids the eggs are usually quite
closely confined within the closed walls of the gonophore, and
that, whether there be two or even more in a single gonad, they
are always more or less flattened against each other or against

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 225

the spadix, and that the entire course of development takes place
within these walls it is difficult to see how any considerable move-
ment of the blastomeres upon each other can be possible. Or,
furthermore, how it is possible to have anything like the regula-
tion type of cleavage characteristic of the forms just referred to,
unless, perchance, the gonophores of C. squainata differ very
greatly from those of C. Icptostyla, which hardly seems probable.
At any rate, I have not found it possible to trace any close cor-
respondence to such features in our species, as will be seen from
the following account.

As stated in an earlier connection, the gonophores of C. Icp-
tostyla may contain only a single egg or as many as four, though
the usual number is two or three. In many cases where but two

FIG. 6. Camera sketch of gonophore FIG. 7. Camera sketches of two

containing two eggs at slightly different eggs, e and/", as seen from side,
stages.

are found they will be upon opposite sides of the spadix, but fre-
quently also supported side by side at the terminal portion. In
either case the eggs are flattened very much on one or more
surfaces, as shown in several of the figures (Fig. 7, c and /). In
some cases they may even come to have a biscuit shape, or may
be crescentic disks, as also shown in Figs. 3-5. In many cases
as the eggs approach full growth there is a tendency to become
more or less spherical, especially when occupying singly the ter-
minal portion of the gonophore. Now, I have found that these
various conditions have a more or less marked effect upon the
mode of cleavage. Where the egg is spherical, for instance,
cleavage is usually more or less symmetrical, as shown in Fig. 8.
On the other hand, where the conditions maintain a continued

226

CHAS. W. HARGITT.

pressure the cleavage is very irregular and unequal, as shown
in Figs. 6 and 7, drawn from life with the aid of the camera. In
Figs. 10 to 1 8 are shown conditions found in sections of an egg,

FIG. 8. Sketch of section of egg, showing nearly regular cleavage. In two cells
are resting nuclei, and in others various phases of nuclear reorganization.

which are almost exactly equivalent to similar sections obtained
in the study of T. mesembryanthemum ('(H4).

I think it will be evident, from even a cursory study of the
several figures and photographs, that the cleavage in this egg is,
like that of Pennaria, more or less erratic and indeterminate, and
conforms to none of the regular types.

FIG. 9. Sketch of two cells of egg section, showing in one a dumb-bell shaped
nucleus in amitotic division.

Another feature of cleavage remains to be considered, namely,
one which involves chiefly, perhaps in some cases wholly, the

DEVELOPMENT OF THE EGG OF CLAVA LEFTOSTYLA.

227

nuclei. In Eudendrium and in the species of Tubularia referred
to above, it has been shown that in a considerable number of
eggs there was formed by nuclear proliferation an evident syncy-
tium, and that from this there was later a differentiation of the
embryonic tissues without the process of ordinary cleavage. I
have found something of the same kind in the case of Clara.
Occasionally an egg was found among serial sections which
showed no evidence of cytoplasmic cleavage, but where internal
nuclear proliferation was clearly evident, and the specimens were

15

FIGS. 10-18. Outline sketches of sections of single egg, showing the various dis-
tribution of nuclei, shape of egg, etc.

numerous enough to enable one to definitely determine the phases
of tissue differentiation and the formation of the embryo. In these
cases were found the same evidences of the origin and organiza-
tion of nuclei dc novo which were found in Eudendrium and Pcn-
naria. It was possible to trace almost every phase of this nu-
clear organization, from the appearance of the smallest particles of
chromatin and their segregation into larger masses to the fully
formed resting nucleus with its typical elements in normal rela-
tions and proportions.

- linitosis and Nuclear Organization. — During the early cleav-
age, even up to the sixteen-cell stage, no evidence of mitosis has

228 CHAS. W. HARGITT.

been found. I have already directed attention to its entire ab-
sence during maturation also. Now it is interesting to find
abundant mitoses during the later cleavage, and during the growth
of the embryo. Is it not somewhat remarkable that among the
eggs of a given gonophore cluster in various stages of cleavage
and later embryonic development there should be found in certain
cases abundant mitoses while in others their entire absence ? This
cannot be attributed, under such circumstances, to differences of
fixation or other details of technique, for this has been identical
throughout. To what, then, may it be attributed, or how ex-
plained ? As already suggested in connection with the subject of
maturation, it seems to me we are forced to assume the operation
of some extremely obscure bio-chemical changes which neutralize
alike both acid and basic stains, or that certain phases of the mi-
totic mechanism may be disguised or actually lacking.

In the light of cumulative evidences along this line I think one
may safely assume that the second of these alternatives is the
more probable, and the facts here submitted strongly support this
conclusion.

During the earlier aspects of development the nuclear phenom-
ena associated with these eggs are extremely indefinite and obscure.
During maturation there seems to be an actual fragmentation and
dissipation of the nuclear matter throughout the cytoplasm.
The first signs of its reorganization appear in the segregation of
chromatin-like masses during the early phases of cleavage, about
which the cytoplasm becomes organized into cell-like masses, as
suggested in an earlier connection. At this time these chromatin
masses are extremely irregular in form and size, stain very
densely, and appear as indefinite, flocculent patches occupying
the center of cytoplasmic aggregates, or cells. They usually
assume an elongate, or often dumb-bell shaped form and seem to
divide amitotically, followed by the division of the cell. In Fig.
8 is shown a camera drawing of a section of an egg, in which
these features are evident. These are very common features
during this period of cleavage, and must, it seems to me, be re-
garded as more or less typical. In two of the cells of Fig. 8
will be observed typical resting nuclei. Whether this is prelimi-
nary to the beginning of the later aspects of mitosis I am unable

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 229

to say, though it seems altogether probable. At any rate, the
facts seem clearly to justify the general conclusion that for a time
in the early history of the development of this egg nuclear ac-
tivity differs very greatly from the ordinary forms of mitosis, and
appears to involve direct, or amitotic division. During later
cleavage abundant mitosis clearly indicates the prevalence of this
form of cell division, if, indeed, it may not wholly supersede the
other, though its appearance during regenerative activities shows
the possibility of its recurrence under various conditions.

I have called attention to similar nuclear phenomena in the de-
velopment of Endcndriiun and Pcnnaria, and in the earlier paper
('O42, p. 267), cited observations of similar sort from several
sources. More recently still other cases have come to light, and
it seems altogether probable that as facts multiply and attention
is focused upon the phenomena cytologists will be forced to take
cognizance of this form of cytogeny and give to it something
more than a merely incidental place in cellular activities, and
assign to it something more than senescent significance.

Among the more recent data bearing upon this point may be
cited, first, the observations of Osborn ('04), in connection with
the development of Fasciolaria ; and second, similar observations
by Glaser ('05), on the same organism, which go to substanti-
ally confirm the facts noted by Osborn, though with somewhat
different phases of interpretation. Still a third series of facts are
brought to light by Child ('04) in a paper on " Amitosis in
Moniezia," in which he clearly shows the prevalence of this form
of cell division in the growth of the reproductive organs and the
development of the sexual cells, and expresses the belief that
" future investigation will probably show that amitosis is at least
as important in the life of the cell as mitosis."

As I have elsewhere pointed out, it is well known that cell
division in Protozoa exhibits very different cytological features
than do cleavage cells in early ontogeny.

In many, mitosis seems to be entirely lacking, while in most
its features are difficult to correlate with the more typical features
in metazoa. Just why we should insist upon finding among a
class like ccelenterates all the details of cytogenic mechanics
more or less familiar in vertebrates or other higher groups of

230 CHAS. W. HARGITT.

metazoa does not appear to be quite obvious. Moreover, why
mitosis should have come to be regarded as absolutely cardinal
in biologic faith is likewise uncertain. At any rate, the repeated
revisions of creed as to centrosome, chromosomes, prelocaliza-
tion, etc., should suggest a spirit of tolerance toward facts, what-
ever their significance.

SYRACUSE UNIVERSITY,

THE ZOOLOGICAL LABORATORY,
March I, 1906.

LITERATURE CITED.
Agassiz, L.,

'62 Contr. Nat. Hist. United States, Boston. Vol. IV.
Allen, C. M.

'oo Development of Tubularia crocea, Biol. Bull., Vol. I.
Allman, J. G.

'71 A Monograph of the Gymnoblastic Hydroids, London.
Bunting, M.

'94 Origin of Sex-cells in Hydractinia, Jour. Morph., Vol. IX.
Child, C. M.

'04 Amitosis in Moniezia, Anatom. Anz., Bd. XXV.
Conklin, E. G.

'05' Organization and Cell-Lineage of the Ascidian Egg, Journ. Acad. Nat. Sci.,
Philadelphia. Vol. XIII.

'os2 Organ-Forming Substances in Eggs of Ascidians, Biol. Bull., Vol. VIII.

'OSM Mosaic Development in Ascidian Egg, Jour. Exp. Zool., Vol. II.
Driesch, H.

'95 Von der Entwickelung einzellner Ascidienblastomeren, Archiv. fur Ent-

wicklm., Bd. I.
Duerdon, J. IT.

'04 The Coral Siderastrea Radians and its Post- Larval Development, Carnegie

Inst. Washington, 1904.
Glaser, 0. G.

'05 UberdenKannibalismusbei Fasciolaria, etc., Zeits. f. wiss. Zool., Bd. LXXX.
HaeckM, E.

'79 Das System der Medusen, Jena, 1878.
Hargitt, C. W.

'04' Some Unsolved Problems of Organic Adaptation, Science, Vol. XIX.

JO42 The Early Development of Eudendrium, Zool. Jahrb., Bd. XX.

*043 The Early Development of Pennaria, Arch. f. Entwick., Bd. XVIII.

'04* Some Hydromedusae from the Bay of Naples, Mitt. Zool. Sta. Neapel,

Bd. XVI.
Harm, Carl.

'02. Die Entwicklungsgeschichte von Clava squamata, Inaugural Dissertation,
Leipzig. Also Zeits. f. wiss. Zool., Bd. LXXIII.

DEVELOPMENT OF THE EGG OF CLAVA LEPTOSTYLA. 231

Montgomery, T. H.

'98 Cytological Studies, Jour. Morph. Vol. XV., p. 421.
Osborn, H. L.

'04 Amitosis in the Embryo of Fasciolaria, Am. Nat., Vol. XXXIII.
Weismann, A.

'83 Entstehung d. Sexualzellen bei d. Hydromedusen, Jena.
Wilson, E. B.

'03 Experiments on Cleavage Localization in the Nemertine Egg, Arch. Ent-
wicklm. Bd. XVI.

'04 Experimental Studies on Germinal Localization, Jour. Exper. Zool. Vol. I.
Wright, T. S.

'61 Hermaphrodite Reproduction in Chrysaora hyoscella, Ann. Mag. Nat. Hist.
III. ser. Vol. VII., p. 357.

.

'

.

232 CHAS. W. HARGITT.

EXPLANATION OF PLATE IX.

FIG. I. Photograph of section of egg under one-twelfth oil immersion, showing
the crescent of pigment granules extending peripherally from the nuclear region.
This area is very imperfectly shown as compared with the actual condition as seen
under the microscope.

FIG. 2. Photograph of section of an egg about the period of beginning maturation.
The flattened nucleus at the outer margin is evident, as is also similar change in the
nucleolus. There may also be observed the general dissolution of the nucleus. Mag-
nification as in Fig. I.

FIG. 3. Photograph of section through two eggs in a single gonophore, showing
the flattening of the eggs along the line of contact. The cleavage masses may be
observed as about the same stages in each egg. While this is often the case, there
are exceptions, as shown in some of the text figures.

FIG. 4. Section of egg photographed under one-twelfth oil immersion, showing
the more or less syncytial character of the egg at this stage of development.

FIG. 5- Section through a male gonophore, showing on one side the egg, and on
the other the mass of spermatozoa, with the body of the spadix occupying the median
region of the section.

FIG. 6. Photograph of portion of an embryo about ready to be liberated, showing
the entoderm and ectoderm well differentiated, and with cell-like masses of pig-
mented yolk spheres in the enteron. It may also be observed that the ectoderm is
practically free of pigment matter, as pointed out in the text.

I am under obligation to my colleague, Dr. Rogers, for the photo-micrographs
illustrating the above features. All were made under the one-twelfth oil immersion
lens, with arc light illumination.

BIOLOGICAL BULLETIN, VOL. X

PLATE IX

REGARDING THE RATE OF GROWTH OF THE
AMERICAN LOBSTER.1

PHILIP B. HADLEY.

\_From the Biological Laboratory of Brown University and the Experiment Station of
the Rhode Island Commission of Inland fisheries. ~\

At the present time when artificial propagation is bidding fail-
to at least partially check the ever-increasing depletion of many
forms of marine animals whose economic value has long sustained
a many-sided fishing industry, any facts which may bear directly
or indirectly upon the life, habits or development of such forms
might seem to be of value. This is especially true of the Ameri-
ican lobster (Homams americanus), a knowledge of whose de-
velopment must influence not only methods of artificial propaga-
tion, which is in these days becoming more common, but also
state legislation in determining the size and season at which the
taking of lobsters shall be allowable.

THE FREQUENCY OF MOLTING AND THE PERCENTAGE OF

INCREASE.

The r.ate of growth of a lobster depends primarily upon two
factors, the frequency of the molting periods and the amount of
increase in length at each molt. To date, the most complete and
only satisfactory account of the development of Homarus is pre-
sented by Herrick 2 who made many observations at Woods Hole
on all of the earlier and many of the later stages. Herrick found
that young lobsters (stages 2 to 10) in confinement gain from 1 1
to 15.84 per cent, at each molt, the average in 66 individuals
being 13.5 per cent. He assumes that 15.3 per cent, is the
average rate of increase both for the young lobsters which grow
up in natural environments, and for adults under normal con-

1 This paper is presented with the purpose of giving in brief the main facts of a
more detailed report on this subject to the Rhode Island Fish Commission. The full
account will appear in The Thirty-Sixth Annual Report of the Rhode Island Com-
mission of Inland Fisheries, 1906. Reprinted as Special Paper No. 23, iqo6.

2 HERRICK, U. S. F. C. Bulletin, vol. 15, 1895.

233

234

PHILIP B. HADLEY.

ditions. The difference in the above percentages he attributes to
the unfavorable conditions of aquarium life. Taking this fact as
a basis, and assuming the average length of the first stage lobster
to be 7.84 mm., Herrick then constructs the following scheme to
show the probable relation between the stage and the size of
lobsters from the time of hatching through the thirtieth molt :

Stage.

Length.

Stage.

Length.

Stage.

Length.

I

7.84 11

32.55

21

135.17

2

9.04 12

37-54

22

155.86

3

10.42 13

43-28

23

179.70

4

12.02 14

49.90

24

2O7.2O

5

13.86 I5

57-53

25

238.90!

6

15.98 16

66.34

26

275-4S2

7

18.42 17

76.49

27

3I7.59

8

21.24 18

88.19

28

366.16

9

24.49

19

101.68

29

422.21

10

28.23

20

117.24

30

486.813

'9.5 inches.

2 inches.

3 19. 1 inches.

Regarding the probable frequency of molts, Herrick assumes
that a lobster molts fourteen to seventeen times during its first
year of life, and that in this time it attains a length of two to three
inches. From this and other detailed considerations, Herrick
finally concludes that a lobster ten and one-half inches long is
between four and one-half and five years old, the higher degree
of probability being in favor of the lower estimate.

The observations made by the writer and others at the experi-
ment station of the Rhode Island Fish Commission at Wickford,
R. I., though differing to some extent from the results obtained
by Herrick at Woods Hole, may serve to throw further light on
the rate of growth of lobsters in their natural environment, and
give some hint as to the conditions which modify it.

The record of the rate of growth of the early stages (one to
ten) include observations upon several hundred young lobsters
whose definite stage and approximate age was, for the most part,
known. Individual records were started immediately after the
molt from the third to the fourth stage and were carried on as
long as either the weather conditions or the term of life of the
young lobsters permitted. In most cases the young individuals
were confined in separate compartment cars which furnished a

KATE OF GROWTH OF THE AMERICAN LOBSTER.

235

very natural environment, and rendered it easy to make observa-
tions at any time. The facts concerning the development of the
early stages may be more tangible when presented in the fol-
lowing table :

WICKFORD LOBSTERS.

Stage.

Length.

Stage-period.

Per cent. Increase.

I

8.2 mm.

2 days

2

9-6

4 days

17.0

3

II. 4

5 days

ICJ.2

4

13-5

1 2 ' days

18.6

5

16.0

1 1 ' days

15-0

6

iS.S

12.5 days

19.9

7

22 5

14 days

21. 0

8

26.5

15.5 days

17.0

9

32.0

2 1 days

21. 0

10

37-9

25 days

. 17.0

Total average 18.3

WOODS HOLE LOBSTERS.

Stage.

Length.

Stage-period.

Per cent. Increase.

I

7.84 mm.

I- 5 days

2

9.2O

2- 5 days

17-3

3

II. 10

2- 8 days

20.6

4

12. 60

10-19 days

13-3

5

14 20

11-18 days

12.7

6

16. 10 14 days

13-3

7

1 8. 60

15.5

8

21.03

12.5

9

24.50

16.5

10

28.03

14.0

Total average 15.3-

1 The fifth stage-period is often shorter than the fourth because the water at Wick-
ford is usually the warmest during the fifth stage-period.

2 This percentage is greater than in the case of another group of 66 lobsters in
which Herrick obtained a result of 13.67 per cent.

The foregoing demonstrates the fact that the average stage-
period of the Wickford lobsters is less than the stage-period of
the Woods Hole lobsters. Further that, while the percentage of
increase at each molt for the Woods Hole lobsters (kept in
aquaria) was only 13.5 or 15.3, the amount of increase for the
Wickford lobsters was over 18 per cent, in the individuals recorded
above, while in the case of a group of individuals which had been
especially selected (/. e., the smaller and weaker specimens were

136

PHILIP B. HADLEV.

thrown out) the average amount of increase for the first ten stages
was 20.9 per cent. In tin's last instance the lobsters in the re-
spective stages were consequently much larger than those in the
group of Wickford lobsters tabulated above. Their average
measurements, however, are presented in the following table :

Stage.

4

5

6

7

8

9

10

Length (mm. ).

14.4

17.0

20.5

24.6

3i -3

37-0

4S-o

Stg.-per. (days).

II.7

II. 2

12.2

13-5

»5-'

21. 0

25-1

Per cent, increase.

18.0

20. 6

20. o 27.2

18-5

21.6

These and other observations would demonstrate that there are
great variations in the rate of development of lobsters, not only
in different localities, and under different conditions of environ-
ment, but also in the same locality and under identical conditions.
Furthermore, that there is a tendency manifested in those indi-
viduals which are slightly above the normal in size and strength,
to increase the advantage which they have already gained. This
advantage, lodged in the fourth stage lobster, may be no more
than a millimeter, but this slight gain compounded through nu-
merous successive stages gives, even the tenth and eleventh stage,
a decided lead which is never again lost and which may be ob-
served in the last mentioned group.

Continued observations upon the later stages (from the tenth
on) prove that not twelve to seventeen stages, as calculated by
Herrick, but an average of twelve stages are passed during the
first year of the lobster's existence. We may trace the future
development of the young lobster through the later successive
stages as follows : September finds the average individual, hatched
the previous June, in the ninth stage and with an average length
of 32 mm. He passes into the tenth stage in the latter part of
September, with a corresponding length of 37.9 mm. In the
latter part of October or the first of November he enters the
eleventh stage with an increase to 45 mm. Through the months
of November, December, January, February and March he lies
dormant, passing into the twelfth stage some time in April or the
first part of May. Thus it appears that a lobster one year old is
in the twelfth stage and has an average length of 53 mm. There
are always exceptions to this rule, - - instances where an individual

RATE OF GROWTH OF THE AMERICAN LOBSTER. 237

may occasionally pass into the twelfth stage before the winter
months. Such specimens sometimes manifest an increase of 28
per cent, in passing a single stage. These lobsters are, however,
usually among those which were hatched early in the season, and
are not very common.

From observations upon the yearling lobsters it becomes ap-
parent that the young creature molts on the average of four times
during its second year of life. The thirteenth stage is entered
some time in July or August, with a corresponding length of
62 mm. In the latter part of August he molts for the thirteenth
time and now covers 73 mm. The entrance to the fifteenth
stage occurs in October of the second year. No further change
takes place until the following April ; that is to say that the aver-
age lobster passes its second winter in the fifteenth stage, length
86 mm. (33/3 inches). By the middle of June we find the young
lobster, now approximately two years old, in the sixteenth stage,
and with a length of 102 mm. (4^ inches).

Observations on the molting periods of lobsters over two years
old make it apparent that the entrance to the seventeenth stage
takes place some time in the late summer of the third year. The
lobster generally molts again before the winter months of the
same year into the eighteenth stage with a length of 141 mm.
(55/6 inches). No further change is experienced until the follow-
ing April.

After the seventeenth or eighteenth stage the percentage of
increase at each successive molt undergoes a gradual diminution
as the molting periods become less frequent. The amount of in-
crease for lobsters about 6 inches in length appears to be in the
neighborhood of I 5 per cent. Thus continuing, we find that the
young lobster passes its third winter in the eighteenth stage,
molts again in the spring (usually in April) and by June, when
approximately three years old, has a length of 162 mm. (6^
inches).

In lobsters of 7 inches and over we find a still smaller per-
centage of increase at each molt ; 1 1 or 1 2 per cent, represents
with a fair degree of accuracy the average percentage of increase
in length for lobsters between 7 and 10 inches.

Further observation reveals the fact that lobsters over 6 inches

238 PHILIP B. HADLEY.

in length do not molt oftener than twice in a year ; once in the
spring or early summer and once in the autumn. Thus the aver-
age lobster enters the twentieth stage some time in the autumn of
his fourth year and at this molt increases from 162 mm. to 181
mm. (7% inches). In the late spring or early summer of the
following year the lobster, now approximately four years old,
enters the twenty-first stage with a corresponding length of 200
mm. (8 inches).

If the case is not one of a young female bearing external eggs
(very rare in lobsters of this length), we may expect another molt
the following autumn and consequently find the lobster in the
twenty-second stage now with a length of 222 mm. (8^j inches).
In all probability the molting periods of the male and female re-
main the same until past the nine-inch length. Therefore, the
entrance to the twenty-third stage probably takes place just be-
fore, or at any rate soon after, the lobster becomes five years old.
The corresponding length is 247 mm. (9^5 inches).

By the time this length is reached many of the female lobsters
are sexually mature and are bearing external eggs. Owing to
this fact, from this time on the rate of growth of the females must
be much diminished. This is due no doubt to the checking of
the growing process, a phenomenon which very naturally pre-
cedes the spawning period ; also to the length of time (ten to
eleven months) the eggs are carried. The male lobsters, on the
other hand, maintain their former rate of development so that by
the twenty-fourth stage the average male lobster has a length of
275 mm. (11 inches) and cannot be much less than six years
old. In the case of the females, however, which have borne eggs
since the nine-inch stage, the eleven-inch limit cannot be attained
in a shorter period than eight years.

This discrepancy in the rate of growth of the male and female
lobsters from this time on, is undoubtedly the explanation of the
fact that, in nearly all individuals in which the sex has been ob-
served, the " giant ' lobsters have been of the male variety.
There are few data on the rate of growth of large lobsters but it
is probable that after the ten-inch size has been attained, the lob-
ster does not molt oftener than once in a year ; and after the fif-
teen-inch stage not oftener than once in two years. Regarding

RATE OF GROWTH OF THE AMERICAN LOBSTER.

239

Stage No.

Sex.

Age.

Length, mm.

I

Male and

8.2

2

Female.'

3 days.'

9.6

3

• "

7 days.

II. 4

4

1 1

12 days.

13-5

5

«

24 days.

16.0

6

«

36 days.

18.8

7

1 1

7 weeks.

22.5

8

0

9 weeks.

26.5

9

t <

3 months.

32.0

10

1 t

5 months.

37-9

ii

t«

9 months.

45-o

12

< t

I year.

53-o

13

> i

i yr. i mo.

62.0

H

1 1

I yr. 3 mo.

73-o

IS

«

i yr. 6 mo.

86.0

16

n

2 yrs.

IO2.O

17

1 1

2 yrs. 3 mo.

121. 0

18

it

2 yrs. 6 mo.

I4I.O

19

a

3 yrs.

162.0

20

K

3 yrs. 6 mo.

iSo.O

21

( £

4 yrs.

2OO.O

22

Male.

4 yrs. 6 mo.

222.0

C (

Female.

4 yrs. 6 mo.

1 1

23

Male.

5 yrs.

247.0

. "

Female.

6 yrs. 5 mo.2

t i

24

Male.

6 yrs.

275.0'

(C

Female.

8 yrs. 4 mo.

t c

25

Male.

7 >'rs.

3OO.O

t C

Female.

10 yrs. 4 mo.

it

26

Male.

8 yrs.

327.0

11

Female.

12 yrs. 4 mo.

i I

27

Male.

9 yrs.

356.0

(1

Female.

14 yrs. 4 mo.

( i

28

Male.

10 yrs.

380.0

1 1

Female.

1 6 yrs. 4 mo.

< c

29

Male.

12 yrs.

4O6.O

t i

Female.

iS yrs. 4 mo.

it

30

Male.

14 yrs.

431-0

ii

Female.

20 yrs. 4 mo.

4 t

31

Male.

17 yrs.

457-0

32

i t

2O yrs.

480.0

33

t *

23 yrs.

505-0

34

1 1

26 yrs.

525-0

35

* i

29 yrs.

546.0

36

1 1

33 ys.

568.0*

1 Age at entrance to the stage.

2 Assumes that the lobster spawns for the first time in the summer of its sixth year,
and that the eggs hatch the following summer.

3 1 1 inches. 4 22% inches.

the growth of " giant " lobsters, it appears reasonable to believe
that the molting process does not occur oftener than once in
three years ; and this is a small estimate. The amount of in-
crease in these specimens at a single molt cannot be over four or

240 PHILIP B. HADLEY.

five per cent, and is often inappreciable. The shells of these
huge lobsters present every appearance of great age and give
testimony to a life of inactivity. Usjng as a basis the observa-
tions which led to the foregoing conclusions, the writer has com-
piled a table showing the estimated rate of development of lob-
sters from the time of hatching to the attainment of the greatest
known size. While the data on the first twenty stages have
their ground in actual observation, the records of the later stages
have been deduced from less positive evidence, and are, to a great
extent, speculative. The great variation in the size of lobsters,
even of the same age and stage, render it well-nigh impossible
to tell off-hand the age of any adult lobster. On the other
hand, the size of large numbers of individual lobsters of a certain
age must remain not far from a general average, on a basis of
which, the approximate age of large numbers of individuals can
be determined with a fair degree of certainty. It is this average,
together with the correlated age, that the writer has attempted to
formulate in the preceding table :

INFLUENCES ON THE RATE OF GROWTH.

Among the influences which modify the rate of growth of
young lobsters under natural or artificial conditions, are to be
mentioned especially the following : temperature, food supply,
light, parasites, injuries and individual physiological peculiarities.
It is probable that the water temperature and physiological con-
dition are the most influential for young lobsters in the ocean.
The others enter largely into consideration in the problems of
artificial propagation. The frequency of the molting periods
and, secondarily, the amount of increase in length at each molt,
is directly dependent upon and determined by the prevailing tem-
perature of the water ; a difference of twelve degrees may cause
the period of growth to the fourth stage to be over twice as long
as normally. Thus we find great variations in the rates of growth
of lobsters at different points on the Atlantic coast. For this
reason it is very probable that the lobsters in the warm waters of
Narrangansett Bay may attain marketable size (lOj/z inches in
Massachusetts) much sooner than do the Massachusetts or Maine
lobsters.

RATE OF GROWTH OF THE AMERICAN LOBSTER. 24!

It is apparent through other observations that the effects of
strong lights and, as shown by Emmel,1 the mutilation of ap-
pendages, exert an influence detrimental to the development of
the lobster in the early stages. Excessive sunlight in cases
where the lobsters were exposed superficially in the water, ap-
pears to cause not only a marked increase in the duration of the
stage-periods, but also a diminution in the percentage of increase
in length at molts ; and furthermore, a generally less healthy
condition in the lobsters themselves. This may be brought
about either directly, by inhibiting the body processes and gen-
eral metabolism, or indirectly, by favoring the excessive growth
of diatoms, algae and protozoa which, under certain conditions,
may accumulate on the body and appendages, to such an extent
as to prevent nearly all activity. It is also observable in this
connection, that food supply may play an important role in de-
termining the size of the young lobster.

1 V. E. EMMEL; "The Regeneration of Lost Parts in the Lobster," The Thirty -
Fifth Annual Report of the Rhode Island Commission of Inland Fisheries,

NOTE ON THE INFLUENCE OF TEMPERATURE

UPON THE RATE OF THE HEART-BEAT IN

A CRUSTACEAN (CERIODAPHNIA).1

T. BRAILSFORD ROBERTSON.

1. Arrhenius has shown2 that the velocity of a chemical
reaction increases very much more rapidly with increasing tem-
perature than any known physical phenomenon. The velocity
of a chemical reaction increases about 10 per cent, per degree
centigrade rise in temperature while molecular velocity, the elec-
tric conductivity of a wire, the elasticity of a solid, the viscosity
of a fluid, surface-tension, etc., are much less affected by the same
rise of temperature. The viscosity of a fluid, for example, only
diminishes about 2 per cent, per degree centigrade rise in tem-
perature. The viscosity of a gas, according to Maxwell's equa-
tion for air3 should increase about 6.7 per cent, per degree centi-
grade rise in temperature at 15° C.

2. The fact of the extreme variability of chemical reactions with
temperature has been applied to ascertain whether given biological
phenomena involve chemical reactions or not. Arrhenius 4 and
van t'Hoff5 have shown that the quotient

velocity of reaction at Tn+lo
velocity of reaction at Tn

is equal to about 2, Tn being the absolute temperature. That is
to say, the velocity of a chemical reaction increases about 100
per cent, per ten degrees. Hertwig 6 has found that the rate of
development of frog's eggs increases very rapidly with rise of
temperature. At Dr. Loeb's suggestion C. D. Snyder" investi-

1 From the Rudolph Spreckels Physiological Laboratory of the University of
California.

2 Zeitsch rift f. Physik-Chemie, 1899, Bd. 4, s. 226.

3 "Theory of Heat," 1872, p. 279.
« Loc. cit.

5 " Etudes de dynamique chimique," p. 112, etc.

6 " Die Zelle und die Gewebe," Bd. II., S. 119.

7 University of California Publications, Physiology, Vol. 2, p. 125.

242

RATE OF THE HEART-BEAT IN A CRUSTACEAN. 243

gated the influence of temperature upon the rate of the heart-
beat and found that the rate of the beat in the heart of the ter-
rapin (Clemmys inannoratii) is almost exactly doubled by ten
degrees rise in temperature between the temperatures 10° C. and
32.5° C. — at lower temperatures the rate is more than doubled
by a rise of 10° while at higher temperatures the rate is some-
what less than doubled by the same rise in temperature. Loeb 8
has found that the velocity of artificial maturation in the eggs of
Lottia is more than doubled by raising the temperature from 8°
C. to 1 8° C. Other investigations on the influence of tempera-
ture upon biological phenomena are being carried out in this
laboratory. These observations do not prove that the above-
mentioned biological phenomena are entirely chemical in char-
acter, but they afford indication that chemical reactions are
involved although not to the exclusion of possible concurrent
physical changes.

3. It appeared to me of interest, in connection with experiments
on the influence of electrolytes on the rate of the heart-beat, of
which an account will appear at an early date, to ascertain the
influence of temperature upon the rate of the heart-beat in the
transparent fresh-water crustacean Ceriodaphnia (?). The organ-
ism, after washing in tap-water, was laid in a drop of tap-water in
the depression on the glass top of an Englemann gas-chamber,
the temperature being regulated by running warm or cold water
through the chamber. A thermometer was fitted into the chamber
so that the bulb lay directly under the depression in which the
organism was placed. A few minutes was always allowed to
elapse before the rate of the beat was registered in order to allow
the organism to attain the same temperature as the bulb of the
thermometer. The beats at room-temperature and at higher tem-
peratures are so rapid that they cannot be counted but have to
be recorded by tapping a key which completes a circuit includ--
ing a signal-magnet, which thus registers a mark upon a revolv-
ing drum for every beat, the time being taken with a stop-watch.

4. The following are the experimental results.

8 University of California Publications, Physiology, Vol. 3, p. I.

244

T. BRAILSFORD ROBERTSON.

Temprrature.

24.75°
20.75°

15-75°

Temperature.
15°

EXPERIMENT i.

EXPERIMENT 2.

Peats per Second.

6-53

4-43
2.47

Beats pet 'Second

2-53

27°

Temperature.
15°

17°
19°

21°
23°

29°

Rate at 25'
Rate at 15'

EXPERIMENT 3.

Rate at 27'

= 2.

4-39
6. 1 1

Beats per Second
2.63

2-95

3-67
4.87
5.05
6.59

6. 10
7-375
Rate at 29°

Temperature.
13°

15°

17°
19°

21°
23°

25°
27°
29°

Rate at 23°
Rate at 13°

Rate at 27°
Rate at 17°

Temperature
11°

13°

I9°
21°

25°
29°

31

Rate at 17° " °7 ' Rate~~aT7cT
EXPERIMENT 4.

= 2.OI

= 2.25 ;

= 1. 80;

Rate at

Beats per Second.
2.22
2.88
3-46

2-95

3.91

5

5-7i

6.23

615

Rate
Rate

Rate

EXPERIMENT 5.

Beats per Second.
2-31

at 29°

- = -> 08

f^. — "• — *\j{j .

°

4.12

4-79
6.15
6.22
6.17
7.01

f

at 19

Remarks

Irregular pauses in
diastole accompanied
by very slight quick
beats.

Still irregular.
Regular.

Gills stopped

RATE OF THE HEART-BEAT IN A CRUSTACEAN. 245

At 33° the whole animal went into convulsive tremors, and at
35° the heart stopped permanently.

Rate at 2 1 ° Rate at 2 5 °

Rate~atTi i ° = 5 ' Rate at 15° =

Rate at 29° Rate at 31°

Rate at 19° ? ' Rate at 2!°

EXPERIMENT 6.

Temperature. Beats per Second. Remarks.

11° 2. 1 1 Beat irregular

13° 2.53

15° 2.53 " regular

19° 2.67 " "

21° 4.05

23° 4.80 "

25° 5.56

29° 6.50 " "

29° 5-33 After a fall and subsequent

rise in temperature.

Rate at 21° Rate at 23° Rate at 25°

Rate at 1 1 ° = l'g2 ' Rate at 13° = I-9° ' Rate at 15"° =

Rate at 29° Rate at 29°

— 0- = 2.3o; =r- - (second observation) = 2. oo.

Rate at 19° Rate at 19°

EXPERIMENT 7.

Temperature. Beats per Second.
11° 1.64

15° 2.38

17° 2.63

19° 2.98

*2I° 4.53

21° 4-55

21° 4.48

25° 5-33

27° 5-4

29° 5-15

Average rate at 2 1 ° Rate at 2 5 °

"RateiTiT5' Rate at 15°

Rate at 27° Rate at 29°

RatTatTp" 2'°5' Rate at 19°

The beat was regular throughout the observations in this ex-
periment. The observations were made in the following order,

246 T. BRAILSFORU ROBERTSON.

21°, 21°, 29°, 27°, 25°, *2i°, 19°, 17°, 15°, 1 1°-- hence the
rate of beat at 21° was counted three times, twice at the begin-
ning of the experiment and once after taking the animal through
a series of temperature-changes none of which were extreme —
it will be observed that the three observations at 21° are in very

good agreement.

EXPERIMENT 8.

Temperature. Beats per Second.

11° 1.82

15° 2-95

17° 3-78

19° 4-09

*2i° 4-375

21° 4-151 .

> Average = 4.40

21° 4-65 '

25° 4.67

27° 6.67

Average rate at 21° Rate at 25°

Rate at 11° '4I ' Rate at i 5°

Rate at 27°

Rate at 1 7

= 1.76.

The beat was regular throughout the experiment. The organ-
ism was taken through the following temperatures in the follow-
ing order: 21°, 21°, 29° (beats not counted), 27°, 25°, *2i°,
19°, 17°, 15°, 11° —hence the rate of beat at 21° was counted
three times, twice at the beginning of the experiment and once
after taking the animal through a cycle of temperature-changes
none of which were extreme — the three determinations of the
rate at 21° are in good agreement with each other and with
those in experiment 7.

EXPERIMENT 9.

Temperature. Beats per Second.

19° 3-2?

21° 5-12

29° 6.25

The organism was taken through the following temperatures
in the following order, 31° (not counted), 29°, 21° and 19°.

Rate at 29°
Rate at 19° =

RATE OF THE HEART-BEAT IN A CRUSTACEAN. 247

EXPERIMENT 10.

Temperature. Beats per Second.
29° 6.15

21° 3.06

21° 3.23

19° 2.86

Rate at 29°

Rate~at 7g° =

The beat was regular throughout, but the temperature had
been raised above 29° before the observations at 21° were taken.

EXPERIMENT n.

Temperature. Beats per Second.

17° 2-38

21° 4.04

27° 5

Rate at 27°
Rate at 17° =

The observations were taken in the following order, 21°, 17°,

27°.

EXPERIMENT 12.

Temperature. Beats per Second.

13° 2.28

21° 4-57

23° 5

Rate at 23°
Rate at 13° =

Rate at T

5. The average value of the coefficient -^- ".t10 for all the

Rate at fn

observations made was 2.03. The results are therefore such as
to confirm Snyder's observations, referred to above, in so far as
they apply in general to rhythmically contracting tissues and to
lead us to conclude that a chemical reaction is involved --as I
have assumed in previous papers.1

6. The rate of the beat in different individuals at 21° is
remarkably constant, provided they have been treated in the
same manner and have not been subjected to any extreme tem-
peratures— this is particularly well shown in experiments 7 and

1 Transactions of the Royal Society of South Australia, Vol. XXIX., p. 47;
Plilger'1 s Arch. f. d. ges. Physiologie, Bd. IIO, s. 610.

248 T. BRAILSFORD ROBERTSON.

8. I have repeatedly confirmed the fact that the rate of the beat
at 21° is about 4.5, with a possible variation of about 5 per cent.
The organisms were all taken from the same vessel, in which
they were kept ; but they differed widely in size and develop-
ment, some containing eggs and others containing well-developed
embryos. Ceriodaphnia therefore affords homogeneous material
for experiments on the heart-beat.

TWO DISTOMES.

L. B. SEELY,

•
HAVER KORD COLLEGE.

i. PNEUMONGECES COMPLEXUS, N. S., a NEW FROG DISTOME.

This description is based on the examination of several pre-
served specimens which were received from North Carolina
marked " From the mouth of Rana pipicns."

F. coinplexus is a rather elongated worm, between 5 mm. and
8 mm. in length, and in the widest portion, about 1.7 mm. wide.
The thickness is about .71 mm., a cross-section being elliptical
in outline. The widest region is just forward of the middle, and
from here it tapers very slightly toward the hinder end, which is
rounded and blunt. Toward the forward end it tapers more
rapidly, sometimes giving the worm almost the appearance of
having a neck. The oral sucker is subterminal and .4 mm. in
diameter. The acetabulum is 2.4 mm. from the anterior end of
the animal and very slightly smaller than the oral sucker, being
.38 mm. in diameter. Both suckers are sessile. All the speci-
mens examined were smooth, /. e., without spines. This may
have been due, however, to maceration. Of the five species of
this genus described by Stafford (1902), all are covered with
spines except one, P. longiplcxns, which is smooth. The deter-
mination of this point with respect to P. complc.vns will depend
finally upon the examination of a living or freshly-killed speci-
men. In the specimens examined, however, there were patches
where the cuticle seemed to be intact. It was very thick and
there was no trace of spines. Thus the evidence seems almost
conclusive that the worm is a smooth one.

The mouth is easily discernible in the center of the oral sucker,
and leads to the muscular, bulb-like pharynx. The oesophagus
is about .5 mm. in length. It then branches, forming the two
intestinal caeca, which extend to within .4 mm. of the posterior
end of the body. The caeca are rounded tubes about . 1 5 mm.
in diameter, and are without branches.

249

250 L. B. SEELY.

The excretory vesicle, which is tubular and in cross-section
about the size of one of the intestinal caeca, is extremely long.
The median portion extends from the excretory pore, which is
terminally located at the posterior extremity, to a position just
posterior to the acetabulum. Here it branches, forming the two
crura, which extend outward and forward to about the plane of
the posterior end of the cesophagus. The excretory vesicle is
extremely thin-walled, and almost impossible to be seen except
in sections. It is dorsal in position throughout its entire extent.

The testes are two large, very prominent organs just posterior
to the middle of the animal. They are irregularly ovate in out-
line and are somewhat lobed. They do not both lie in the same
plane, the left one usually being posterior to, but overlapping,
the right. The posterior one is usually somewhat longer than
the anterior, being 1.4 mm. long, while the anterior one is I.I
mm. long. Each is about .92 mm. wide and about .78 mm.
thick. The vas deferens from the left testis runs forward to the
left of the right testis, lying between it and ovary. Just anterior
to the shell gland it is joined by its fellow from the right testis,
and from here a slightly convoluted tube extends to the genital
pore. There is no cirrus or cirrus sac. The genital pore is a
minute opening on the ventral side, in the same transverse plane
as the pharynx and to the right of the median plane.

The ovary is a somewhat elongated organ considerably smaller
than either testis. It is usually on the left side and slightly pos-
terior to the acetabulum. Occasionally it is on the right side,
and when it is so placed the testis on that side is posterior to the
other. It is about .7 mm. in length, .32 mm. in width, and .32
mm. thick (dorso-ventrally). The oviduct runs from the ovary
into the shell-gland where there is an ootype. The shell-gland
lies in the middle of the body just posterior to the acetabulum.

The receptaculum seminis is a sac about one-half as large as
the ovary. It lies just forward of and lateral to the right testis.
Its duct runs forward and to the left, meeting the oviduct at its
entrance into shell gland. There is no Laurer's canal.

The yolk glands are numerous, being placed in clusters along
both sides of the body from the position of the pharynx to the
posterior extremity of the hinder testis. There are five or six

TWO DISTOMES.

251

distinct groups or clusters on each side, each group being com-
posed of 6 to 20 glandular bodies. They are ventrally placed
and follow the general course of the intestine.

The tube from the forward clusters on each side meets that
from the rear clusters just inside the intestinal caecum and in a
transverse plane just posterior to the ootype. The transverse
tubes meet near the median plane, well toward the dorsal body-
wall, and from here a short tube leads to the yolk reservoir.

I

Tes
Ut(d).

FIG. I. PntitntotKvces complexus, ventral aspect. O. s., oral sucker ; A., acetabu-
lum ; Ph., phar)nx ; Int. , intestinal crecum; JLxc. v., excretory vesicle ; Lt. , uterus ;
7>.f. , testes ; K. s. , receptaculum seminis ; Gen. p., genital pore ; Ov., ovary; Sh.g.,
shell gland ; /*. def., vas deferens ; Y. g., yolk gland.

FIG. 2. Diagram of female genital organs in P. complexus. Ov., ovary; Sh. g.,
shell-gland ; Ut., uterus ; R. s., receptaculum seminis ; T. y. d., tranverse yolk duct.

FIG. 3. Cross-section of Renifer e/ongafus in region of testes. Ut. (</), uterus,
descending; Ut. (a), uterus, ascending; Tcs., testis ; Int., intestinal caecum;
Exc. i'., excretory vesicle.

From the yolk reservoir a tube extends diagonally forward and
towards the ventral body wall joining the oviduct after its entrance
into the shell gland. The yolk reservoir is a pear-shaped sac
about .12 mm. in length.

252 L. B. SEELY.

The uterus is a continuation of the oviduct, running to the
posterior end of the anima-1, then forward to the genital pore. It
winds about, filling the spaces between the other organs, and in
the case of adult worms often nearly covers them. It is circular
in cross-section, and about .15 mm. in diameter. In the adult
animal this tube is filled with eggs which as they ripen take on
a dark brown color, and give the worm a peculiar mottled ap-
pearance. There are no lateral longitudinal folds as in the case
of most of the other species of this genus.

P. coinplexus is closely related to P. siini/iplcxus, described by
Stafford (1902) under the name of H&matolcechus similiplexus.
The principal differences between the two species are that in P.
coinplexus the testes are very much larger than in P. similiplexus ;
there are no lateral longitudinal folds of the uterus ; and the eggs
are considerably smaller, being .029 mm. by .014 mm. as com-
pared with .039 mm. by .019 mm. in P. similiplexus. In outer
characters, --size, shape, suckers, etc., — there is a close simi-
larity between the two species. P. similiplexus is covered with
spines, while P. coinplexus is probably smooth. In this respect
P. coinplexus resembles P. longiplcxus. In respect to the folds
of the uterus it most closely resembles P. medioplexus. The
ascending and descending coils seem however to have less regu-
larity that in that species.

The other species of this genus are universally reported as
being found in the lungs of frogs, while, as has already been
noted, these specimens were labelled as having been found in the
mouth of the leopard frog. Upon inquiry I found that these
distomes are rarely or never found in the mouth of a living or
freshly-killed animal. If, however, the frogs are killed by
chloroform, and especially if they are left for some time in the
killing jar, the worms, still alive, may frequently be found in the
mouths of the dead animals. They may be thrown from the
lungs by the struggles of the dying animal, or they may find
their way to the mouth in their efforts to escape from a dead
host. At any rate the above facts explain their being found in
the mouth of the frog, and leave every probability that P. com-
plexus is a lung parasite.

TWO DISTOMES. 253

II. RENIFER ELONGATUS PRATT.

This worm was first described by Pratt (1903) from a single
specimen, mature but not yet fully grown. The object here is
simply to supplement that description, by a comparison with a full-
grown specimen. The worm from which the following measure-
ments were taken was selected as the largest of about 40 or 50
specimens at hand. This worm is 5.5 mm. in length, and 1.7
mm. wide at the widest part, while that described by Pratt was
3 mm. in length, and .68 mm. in width. The thickness of the two
worms is .93 mm. and .4 mm., respectively.

The oral sucker is round and about .35 mm. in diameter;
the acetabulum is decidedly elliptical in outline, extending .82
mm. transversely and .6 mm. longitudinally. The corresponding
measurements given by Pratt are the following : oral sucker ;
length, .32 mm., width, .33 mm. ; acetabulum ; length, .4 mm.,
width, .36 mm. The relative positions of the suckers on the ani-
mal are of course the same. The relative sizes of the various parts
of the digestive tracts of the two worms correspond to relative
sizes of the worms themselves.

The most marked difference between the two worms are the
greatly increased size of the ascending limb of the uterus, and
the consequent displacement of the other organs in the larger
worm. In the region just back of the testes, the uterus occupies
nearly two thirds of the space within the body walls, and the
excretory vesicle is pressed flat against the dorsal wall, extending
from one lateral margin to the other. In the region of the testes
there is but little diminution in the size of the uterus, and the
testes are pressed outward toward the lateral body walls, they
themselves being apparently somewhat flattened (see Fig. 3).
Forward of the acetabulum, the uterus rapidly decreases in size
as it approaches the genital pore.

The eggs were approximately the same size as those in the
smaller worm.- The measurements given by Pratt are .035 mm.
by .02 mm. The measurements in the larger worm were .034
mm. by .021 mm.

254 L- B- SEELY.

BIBLIOGRAPHY.
Looss, A.

'99 Weitere Beitrage zur Kenntniss der Trematodenfauna Aegyptens. zugleich

Versuch einer natiirlichen Gliederung des Genus Dislomum Retzius. Zool.

Jahrb. Abt. Syst., Bd. 12, pp. 521-784.
'02 Uber neue und bekannte Trematoden aus Seeschildkroten, nebst Erorterungen

zur Systematik und Nomenclatur. Zool. Jahrb. Abt. Syst., Bd. 16, pp.

411-894.
Pratt, H. S.

'03 Mark Anniversary Volume, 1903, pp. 28-38.
Stafford, J.

'02 On American Representatives of Distomum Variegatum. Zool. Jahrb. Abt.

Syst., Bd. 1 6, pp. 895-912.
'05 Trematodes from Canadian Vertebrates. Zool. Anz., Bd. 28, pp. 681-694.

Vol. X. May, 1906. No. 6.

BIOLOGICAL BULLETIN

THE YPSILOID APPARATUS OF URODELES.

INEZ L. WHIFFLE.

INTRODUCTION.

Articulated to the anterior margin of the pubis of many sala-
manders, there is a cartilage which, from its peculiar Y shape, is
known as the Cartilage ypsttoides. This cartilage has in general
been assumed, by those who have described and figured it, to be
identical with the anterior paired or unpaired cartilaginous proc-
ess, the Processns epipubicus of the ProtcidcB and Derotremata.
Thus Wiedersheim in his Grundriss (1893, p. 165), says, after
speaking of the paired nature of the anlage of the pelvic girdle,
" Dies hervorzuheben ist namentlich auch wichtig im Hinblick
auf die Morphologische Bedeutung jenes Abschnittes, den ich als
Processns cpipnbicns bezeichnet habe. Die urspriinglich paarige
Natur desselben tritt bei dem Becken von Proteus und Am-
pJiimna zeitlebens deutlich hervor.

"Auch bei den Derotremen und Salamandrinen findet sich am
vorderen Beckenrand in der Medianlinie ein Knorpelfortsatz, der
als Epipnbis zu betrachten ist, und in manchen Fallen lasst sich
dessen directer Zusammenhang mit der eigentlichen Beckenplatte
noch deutlich nachweisen. Insofern aber liegt bereits ein Fall
von sogenannter abgekurzter Entwicklung vor, als der Processes
epipubicus hier nicht mehr paarig, sondern als ein unpaarer Aus-
wuchs sich an seinem Vorderende erst secundar gabelig theilt."

Gegenbaur, also, in his " Vergleichende Anatomic " (1898, Vol.
I., p. 55°). after describing -the pelvic girdle of Nccturns with its
long anterier process says that " Derselbe Teil bei Salamandrinen
als medianer terminal gegabelter Fortsatz erscheint, das soga-
nannte Epipnbis. Aus der Vergleichung dieser beiden Zustande
geht hervor, dass das Epipnbis bereits in der Platte des Pubis

256 INEZ L. WHIPPLE.

besteht und nicht als besonderer Fortsatz auftritt. Seine Ent-
stehung geht sonach aus einer bilateralen Reduction eines Teiles
der ventral Beckenplatte hervor."

Hoffmann (Bronn's " Their-reich," Bd. 6, II., Amphibien,
1873—8, p. 77), says in his description of the pelvic girdle of
Uroddes, " Der vordere Rand des Schamsitzbeins verlangert sich
nach vorn in cine mediane Spitze. Bei Proteus und Meno-
brancJius ist diese Spitze mit der ventralen Platte continuirlich ver-
bunden, vvahrend sie bei den anderen Urodelen durch Syndesmose
damit innig zusammenhangt. Dieser ventrale Fortsatz, welcher
bei Proteus, Menobranchus, und Amphiuma nur sehr kurz ist,
verlangert sich bei den anderen Urodelen in die ventrale Muskel-
masse und spaltet sich vorn gabelformig in zwei divergirende
Schenkel {Cartilago ypsiloides]. Bei Cryptobranchns japonicus*
wo dieser ventrale Fortsatz der knorpeligen Epipliyse der Scham-
beinplatte aufsitzt, ist desser rechter Schenkel in drei Sipfen \_sic\
gespalten, vvelche jedoch am linken Schenkel fehlen."

In all of the above quoted authorities it is implied or assumed
that the epipubic cartilage (Cartilago ypsiloides) occurs universally
in Urodclcs. Wiedersheim ('75), however, in his work on
Salamandrina perspicillata and Geotriton fnscus (p. 142), under
the heading Cartilago ypsiloides expresses himself as greatly sur-
prised at finding no trace of the cartilage in Gcotriton, adding,
" Wo also die Erklarung zu suchen ist, ist mir dunkel gebleiben,
doch ware vielleicht von der Untersuchung der Larven, welche
mir im Augenblick nicht bei der Hand waren, noch etwas zu
ervvarten " ; and recently Miss Emerson in her work on the anat-
omy of TypJilouwlge rathbnni (1905) mentions the absence of the
Cartilago ypsiloides in that species.

The following tabulated results of my own dissection of the
pelvic region of salamanders * shows, moreover, that the Car-
tilago ypsiloides is far from being universal in occurrence :

Cartilago ypsiloides, present. Cartilago ypsiloides ^ absent.

Triton alpestris, Desmognathus fusca,

Triton helveticus, Batracoseps attenuatus,

Dieniyctylns viridescens, Spelerpes mber,

1 The term salamander is used here as synomymous with the suborder Salamandrida,
comprising all of the Urodeles except the Derotremata and Perennibranchiata.

THE YPSILOID APPARATUS OF URODELES. 257

Cartilago ypsiloideS) present. Cartilago ypsiloides, absent.

Amblystoma opacj/m, Spclerpcs bi tin ea tits,

Amblystoma punctatnm, Spelerpes guttolineatiis,

Amblystoma jeffersoniannm, Spelerpes porpliyriticus,
Amblystoma talpoidcum, Spelerpes longecandns,

Salamandra macnlosa, Manc2t/us quadridigitatus,

Salamandrina perspicillata. Pletlwdon cinereus,

Plethodon erytlironotus,
Antodax lugubris.

In those forms in which the cartilage is present it exhibits the
typical Y-shaped character, and is movably articulated to the
anterior edge of the pubis. In the species given in the right-
hand column there is no trace of the cartilage whatsoever.
These results show that the presence of a Cartilage ypsiloidcs is
closely correlated with the presence of lungs, although there
seem to be two exceptions, viz., Amblystoma opaciim and Sala-
mandrina perspicillata. The first of these, however, rests upon
an error, since the Amblystoma opaciim, like the others of its
genus, possesses large and well developed lungs, although
through the authority of Lonnberg ('96) this species has for sev-
eral years been placed among lungless forms. It is impossible
to say through what appearances Lonnburg was deceived, but
the presence of well developed lungs has been repeatedly demon-
strated by me through dissection and physiological experiment.
The most plausible hypothesis is that he was mistaken in his
species, an error extremely likely to occur in the study of this
group.

Marked differences in the activities of lungless and lunged sal-
amanders when in the water, a subject which will receive full
discussion later, further corroborate the view that the function of
the ypsiloid cartilage is correlated with that of the lungs. In fact,
it was this difference in activity which first called my attention to
the difference in anatomy and led me to make an extensive study
of the ventral pelvic region of Urodeles.

258 INEZ L. WHIFFLE.

PART I. ANATOMY OF THE YPSILOID REGION IN URODELES.
A. Description of Adult Forms.

i . Diemyctylus viridescens. — In Dicmyctyhis, as in all Urodelcs,
the muscular abdominal walls are differentiated into the four typi-
cal layers, viz., Mnsculus obliqnus extermts, M. obliqims interims, M.
rcctus abdominis and M. transversalis. The external oblique is
strongly developed, showing the typical outer and inner laminae.
The outer has lost its metamerism except in its origin ; its fibers,
which form a strong, continuous, muscular sheet, are mainly in-
serted into a wide, ventral aponeurosis, and the posterior bundles
into ilium and pubis. The rectus abdominis consists of a narrow
band of muscle, lying on either side of and contiguous to the linea
alba and covered ventrally by the aponeurosis of the obliquus
externus. It is very primitive in character, consisting of a series
of myotomes, separated by well formed myocommata. The
obliquus internus is almost vestigial and in some individuals I was
unable to find it. Its very thin layers of fibers arise mainly by
digitations from myocommata beneath the deeper layer of the ex-
ternal oblique, although the more posterior bundles have their
origin in the anterior edge of ilium and pubis. The fibers of this
muscle extend very obliquely, anteriorly and ventrally and are
inserted into myocommata. At the edge of the rectus abdomi-
nis muscle the fibers of the internal oblique seem in some cases
to become continuous with those of the rectus, in other cases to
pass dorsal to the rectus, which, however, they only slighly
overlap.

The transversalis, even more than the obliquus externus, con-
sists of a continuous muscular sheet, its metamerism being evident
only in its origin. It is inserted ventrally into an aponeurosis
which is somewhat narrower than that of the obliquus externus
and which lies on the inner (7. <\, dorsal) surface of the rectus ab-
dominis. This aponeurosis narrows abruptly in the two meta-
meres anterior to the pubis, and ends at the pubic symphysis in a
point.

In the two myotomes immediately anterior to the pelvic girdle
there is considerable muscular differentiation, which may be most
easily shown by a series of dissections. Fig. I shows a ventral

THE YPS1LOID APPARATUS OF URODELES.

259

view of the posterior abdominal ]region, with the skin and exter-
nal oblique muscle removed from the left side. In the mid
line, extending anteriorly from the pubic symphysis through
the width of a single myotome
is the stem portion of the Carti-
lago ypsiloidcs (r), the two lat-
eral arms of which are hidden by
the rectus abdominis. The first
myotome of the rectus abdomi-
nis, counting anteriorly from the
pelvic girdle, is differentiated
superficially into two portions,
the more Ventral and lateral of
which (ra} extends outward with
its fibers converging to their in-
sertion into a process at the

OUter edge of the pubis, which I FIG. I. Ventral view of Diemyc-

Shall call the lateral process of h'llts virutescens, posterior abdominal
,, , • TU- 4.- • 11 region, V T.. The external oblique

the pubis. This portion is well

muscle has been removed from the left

differentiated into a distinct mus-
cle, and in its insertion is closely
associated with the posterior por-
tion of the external oblique. The
second portion (rb} is partially
covered by the first, and its fibers
slant medially to their insertion
into the ventral surface of the

stem of the ypsiloid cartilage and the anterior margin of the
pubis. When the more ventral of these two muscles is removed,
it is found that the medial portion of the deeper muscle takes its
origin from the lateral arm of the ypsiloid cartilage, which ex-
tends obliquely nearly across the second myotome. In Fig. 2
both of these muscles have been removed, as well as that portion
of the second myotome which lies over but unattached to the
Cartilago ypsiloides. On the right side in the figure may be
seen two muscles connected with the cartilage. One of these,
which may be designated the M. ypsiloidcns posterior (yp\ is a
strongly developed muscle which arises from the anterior margin

side. Abbreviations : ex, M. obliquus
externus ; _/£, M. pubo-ischio-femoralis
externus ; yf, M. pubo-ischio-femoralis
internus ; in, M. obliquus interims;
/, lateral process of pubis ; /, pubis ;
ra, and rb, differentiations of the first
myotome of the rectus abdominis ; y,
ypsiloid cartilage ; ya, M. ypsiloideus
anterior.

260

INEZ L. WHIFFLE.

of the pubis, its thickest portion having its origin in the lateral
process. The fibers of this muscle diverge slightly, and are in-
serted into the edge of the lateral arm, and the inner (/. c\, dor-
sal) surface of the stem of the Cartilago ypsiloides. Lying, as it
does, dorsal to the main mass of the rectus abdominis, this mus-
cle is almost continuous at its lateral boundary with the obliquus

internus. It is the muscle which
has been often called the pyramid-
alis. Thus Hoffmann described it
as a part of the general urodele
musculature and homologized it
with the pyramidalis of human
anatomy.

The other ypsiloid muscle, M.
ypsiloidciis anterior {yd}, arises from
the medial portion of the second
myocomma, and from the linea
alba of the first metamere, and its
fibers extend obliquely outward
and posteriorly to their insertion
into the anterior edge of the lateral
arm of the cartilage.

The transversalis muscle shows
a wide range of individual varia-
tion in its relation to the ypsi-
loid cartilage. In some of the
specimens dissected none of its fibers were inserted into the car-
tilage. The posterior narrowing of the aponeurosis of the trans-
versalis in this region, however, causes the edges of the aponeu-
rosis to lie closely parallel with the outer edges of the ypsiloid
cartilage, and in the majority of the specimens dissected some of
the fibers of this muscle had become inserted into the cartilage,
the ends of the lateral arms being the region where this insertion
most frequently occurs. The extent of the insertion of this
ypsiloid portion of the transversalis varies, however, from one
which involves the outer portion of the arm only, to one which
extends along nearly the entire length of the cartilage, both arm
and stem (Fig. 3, tb}. Almost invariably a few fibers of the

FIG. 2. Ventral view of Diemyc-
tylus vtridesce>is,-posle.r\or abdominal
region, X 3- All of the abdominal
muscles have been removed except
the tranversalis muscles and the yp-
siloid muscles of the right side.
Abbreviations not previously ex-
plained : /, M. transversalis abdom-
inis ; _)'/, M. ypsiloideus posterior.

THE YPSILOID APPARATUS OF URODELES.

26l

muscle immediately anterior to the cartilage extend farther
medially than those of the rest of the muscle and may even be
inserted into the second myocomma in association with those of
the M. ypsiloideus anterior (Fig. 3, to).

It should perhaps be emphasized that the entire musculature
of the ypsiloid cartilage, while in all cases strongly developed,
shows a considerable amount of individual
variation, as if the apparatus were one
of relatively recent origin and still in the
experimental stage of physiological adap-
tation. The most definite of the muscles
involved is the M, ypsiloideus posterior.

In addition to the above described mus-
culature of the Cartilage ypsiloides, the
posterior portion of the stem of the car-
tilage forms the origin of the anterior
portion of the M. pubo-ischio-femoralis in- Fic" 3 • Dorsal ™w ( i. ,.

from within body cavity

tennis. This muscle takes its origin mainly showingthe muscles attached
from the mid ventral line of the pubo- to the ypsiloid cartilage of

ischium, and that portion which arises Diemyctyiusviridescens,K$*

On the right side the ante-
from the ypsiloid cartilage extends be- . f ., A, ,

J* nor part of the M. pubo-

tween the posterior ypsiloid muscle and ischio-femoralis intemus has
the transversalis (Figs, i, 2, and 3,/). been cut awa7- Abbrevia-

T* -, it.- -r /t < • o/ tions not previously ex-

2. Inton helveticus. 1 . albcstns, sa/a-

plamed: /<?, the portion ot

inandra jnacnlosa, Ainblystoma cpacinn, A. the transversalis inserted

pnnctatllUl. Although representing dif- into the second myocomma;

ferent subfamilies, these forms so closely ^ tbf portio1: of lhe trans;

versahs which is inserted

resemble Diemyctylm in the anatomy of into the ypsiioid cartilage,
the ypsiloid region that they may be

grouped together in this comparison. The Tritons most closely re-
semble Diemyctylus, the correspondence part for part being almost
exact. Salamandra and the Amblystomas have a relatively less;
strongly developed external oblique, while the rectus abdominis is.
broader and more powerful and the internal oblique more strongly
developed. The differentiations in connection with the ypsiloid
cartilage (Fig. 4), are, however, practically the same as in
Diemyctylus. The M. ypsiloideus anterior presents less deviation
from the longitudinal course of the rectus abdominis, almost all

262

INEZ L. WH1PPLF.

of its fibers proceeding from the myocomma instead of the linea
alba. The transversalis shows about the same range of variation
that was described in Dieinyctylus ; the edges of its aponeurosis
coincide more exactly with those of the cartilage and in one very
muscular specimen of Amblystoma pitnctatuui, this muscle was

strongly inserted into the cartilage
along the whole length of its arm.
In the same individual the carti-
lage was very large, as its arm
crossed the second myotome, and
the lateral portion of the M. ypsi-
loidcns anterior rose from the third
myocomma thus giving this muscle
an origin from two myocommata.

In all of these genera as in Dic-
invctylns the M. pubo-ischio-femoralis
interims takes its origin partly from
the posterior end of the ypsiloid
cartilage.

3. Salamaiidrina pcrspicillata. —
This species, which, so far as is known, is the only lungless mem-
ber of the subfamily Mecodonta, is sufficiently different in its ab-
dominal musculature from the lunged forms already described, to
require a separate discussion. The outer, deeper portion of the
rectus abdominis (the rectus profundis of Maurer's nomenclature),
is highly specialized into a retractor of the tongue (M. pubo-hyoi-
deus}. It lies in a sheath, in which it moves freely, since although it
possesses myocommata corresponding to those of the remaining
portion of the rectus abdominis, these are wholly disconnected from
those of the main mass of somatic muscles and thus do not corre-
spond with the latter during all phases of muscular contraction.
Aside from this highly specialized region, the abdominal muscles
of Salamandrina are poorly developed. The external oblique
muscle seems reduced almost to a mere sheet of connective
tissue, and is evidently functional more as a support for the
abdominal wall than for any muscular activity. The internal
oblique I have been unable to demonstrate. The rectus abdomi-
anis is thin and its two halves are separated by a very wide line

I'"IG. 4. Ventral view showing
the ypsiloid apparatus of Anihly-
stoma punctatnin, X 2- Dissection
and abbreviations as in Fig. 2.

THE VPSILOID APPARATUS OF URODELES.

263

alba. It does not show in the myotome anterior to the pubis
the superficial differentiation noted in the forms already described.
Neither is any of this superficial portion attached to the ypsiloid
cartilage.

The musculature of the Cartilage ypsiloides is otherwise quite
similar to that of Dicniyctylns. The anterior ypsiloid muscle, is,
owing to the great width of the linea alba, inserted into the distal
portion only of the arms. The relation of the M. pubo-ischio-
fcinoralis interims to the ypsiloid
cartilage is the same as that already
described in the preceding cases.

4. Desmognathus brimleyonnii,
D. fuse a, Spclcrpcs ruder, S. bilinc-
atits, Plethodon erythronotus, P.
cincrcns. — In none of these forms
is there a Cartilago ypsiloides. The
musculature of the posterior ab-
dominal region gives, however,
decided evidence of the former ex-
istence of an ypsiloid apparatus.

The abdominal wall shows a
more primitive musculature than
Diemyctyhts, the metamerism being
much more strongly marked. The
rectus abdominis is broad and powerful. In the myotome
(Fig. 5, r) immediately anterior to the pubis there is the same
superficial differentiation that Dieinyctylns and other lunged forms
exhibit. There is one difference, however, which is due to the
fact that the anterior border of the pelvic girdle is deeply incurved
on either side of the ventral midline so that the border presents
a conspicuous median and two lateral processes (Fig. 6, / and ;;/).
As a result of this the medial portion of the rectus abdominis
extends much farther posteriorly than the lateral portion, since
it is inserted into the ventral surface of the girdle somewhat
posterior to this incurved edge.

In all of the lungless forms dissected the deeper portion of the
first myotome of the rectus abdominis shows a well defined
though not at all strongly developed portion (Fig. 6, ,)'/<') which

FIG. 5. Ventral view of the pos-
terior abdominal region of Desmog-
•n ai hits brimliyoriiiii, X 2- The M.
obliquus externus has been removed
from both sides and the M. pubo-
ischio-femoralis externus from the
right side.

264

INEZ L. WHIPPLE.

FIG. 6. Ventral view of the poste-
rior abdominal region of Desitiog-
nathus briinleyonim, X 2- All of the
abdominal muscles have been removed
except the transversalis muscles and
the vestigial ypsiloid muscles of the
left side. Abbreviations not previ-
ously explained : r>it median process
of pubis ; ya,., vestige of anterior yp
siloid muscle ; yp, , vestige of posterior
ypsiloid muscle.

arises from the lateral process of the pubis. From this origin
the fibers diverge to their insertion mainly into the linea alba,
although in most cases a few are inserted into the first myocomma.
When we consider the Cartilago ypsiloidcs of the lunged forms,
with its stem in the linea alba and its arms diverging at the first
myocomma, it seems very probable that this muscle in the lung-
less forms is a vestigial M. 'ypsiloidens posterior which persists

even after the disappearance of
the cartilage to which it was orig-
inally attached. In some speci-
mens a few fibers of this muscle
were found to extend to the sec-
ond myocomma and in one speci-
men of Dcsmognatlius brimleyorum
a separate little muscle (Fig. 6,
jw) consisting of a few fibers only
was found in the region of the
second myotome arising in the
linea alba and converging ob-
liquely outward and posteriorly.
These variable evidences of differ-
entiation from the second myo-
tome suggest, of course, the
probable vestiges of the M. ypsi-
loidens (Ulterior.

Still further evidence of the for-
mer existence of an ypsiloid car-
tilage in lungless forms is fur-
nished by the fact that in its absence the anterior portion of the
M. pubo-ischio-femoralis interims originates from the linea alba in
the exact region corresponding to the origin from the posterior
part of the stem of the ypsiloid cartilage in lunged forms.

5. Cryptobranchus allegJieniensis. — In Cryptobranclms the ypsi-
loid cartilage is very well developed and its articulation with the
pubis displays marked mobility. It differs somewhat in form
from that of the salamanders, in that its lateral arms are very
much longer and rapidly broaden toward the outer ends so that
they are spatulate in shape (Fig. 7). The cartilage begins to

THE YPSILOID APPARATUS OF URODELES.

265

fork slightly anterior to the first myocomma and the lateral arms
extend obliquely across the third myotome.

The musculature of the cartilage is interesting. Superficially,
the first myotome of the rectus abdominis is imperfectly differen-
tiated into two portions, and the medial fibers not only of the

FIG. 7. Dissection of Cryptubranckus allegkeniensis showing the ypsiloid carti-
lage and the muscles associated with it, X 2^. The external oblique and rectus
abdominis muscles have been removed from the left side. Abbreviations as in
previous figures.

first but of the second myotome are inserted into the stem of
the cartilage, the more anterior fibers of the second monotome,
however, being inserted into the ventral surface of the cartilage
at the very base of the arms.

The deeper layers show a strongly developed M. ypsi-
loideus posterior, differing from that of the forms already de-
scribed in that its insertion into the Cartilage ypsiloides extends
only about two thirds the length of the arm, the thin,
expanded, outer third of the arm being free from all muscu-
lar attachment. The most medial fibers of the third myotome

266

INEZ L. WHIFFLE.

extend as a small bundle from the third myocomma to an inser-
tion into the middle of the ventral surface of the lateral arm of
the cartilage. This differentiation from the third myotome is the
only representative of a M. ypsiloides anterior, and except for its
insertion it is not at all distinct from the adjoining fibers of the •
myotome.

None of the fibers of the transversalis are inserted into the
ypsiloid cartilage. The edges of its aponeurosis in this region,
however, are parallel with the cartilage and separated from it by

a space about equal to the
width of the cartilage itself.
Moreover, the stem and proxi-
mal third of the arms of the
cartilage are firmly bound to
the aponeurosis ; the portion
of the cartilage involved in
this attachment is much
thicker than the free, distal,
expanded region of the carti-
lage which lies in the third
myotome in a sort of sheath
between the aponeurosis of
the transversalis and the deep-
er layers of the rectus abdo-
minis and has neither muscu-
lar nor aponeurotic attach-
ments.

The origin of the anterior portion of the M. pubo-iscliio-fenw-
ralis interims does not involve the ypsiloid cartilage, although it
extends to the extreme anterior margin of the pubo-ischium.

It should be mentioned in connection with this description of
Cryptobranchus allegheniensis that Hyrtl ('65) has described the
ypsiloid cartilage of Cryptobranchus japonicns as lacking bilateral
symmetry in that the right lateral arm is subdivided into three
branches while the left is simple. This description was, I suppose,
based upon a single individual and the condition was quite prob-
ably an abnormal one.

FIG. 8. Dissection of Nectitrus macu-
lalus showing a ventral view of the pelvic
region, with the M. pitbo-ischio-femoralis
extern us of the left side removed, X *•
Abbreviations as in previous figures.

[From an unpublished drawing by H.
H. Wilder.]

THE YPSII.OID APPARATUS OF UROUELES. 267

6. Nccturns niaculatus. — In Ncclitrits there is no vestige what-
ever of a Cartilago ypsiloidcs. The pubo-ischium is produced
anteriorly into a long median point (the epipubic process of
previous writers), the sides of which slant gradually posterio-
laterally to the outer angle of the pelvis where a lateral process
is slightly developed (Fig. 8). The first myotome of the rectus
abdominis muscle is correspondingly narrow at the ventral mid-
line, and much wider toward its lateral boundary. Its fibers are
inserted along the entire anterior margin of the pubo-ischium
and show a differentiation into a medial (rH) and a lateral portion
(ra), the latter inserted into the lateral portion of the pubis. In
some specimens this lateral portion is more distinct than in others,
and occasionally assumes the character of a semi-independent
muscle, as in Diemyctylus. There is no indication whatever of a
muscle dorsal to the main mass of the rectus abdominis and in-
serted into the linea alba, to suggest a vestigial ypsiloid muscu-
lature. The M. pubo-ischio-femoralis intcrmts moreover, has its
origin wholly from the pubo-ischium instead of arising in part
from the linea alba as it does in the case of the lungless sala-
manders. There is therefore absolutely no indication of the
previous existence of a Cartilago ypsiloidcs in this species.

7. Amphiuma means. — Ampliinma means shows a similar failure
of all trace of an ypsiloid apparatus. The pubo-ischium lacks a
mid-ventral symphysis in this form and there is only a slight an-
terior prolongation of the cartilage on either side of the mid-ventral
line where the two halves of the girdle are in contact. The
muscular differentiation in this region is, however, practically
similar to that already described for Nccturns.

8. Siren laccrtina. - • This species not only shows no trace of
a pelvic girdle and appendages, but, as might be expected, there
is also no muscular differentiation to indicate the former presence
of an ypsiloid apparatus.

B. Development of the Yfisi/oiil Apparatus.

My material for the study of the development of the ypsiloid
apparatus was somewhat limited. It consisted of (i) larvae of
Amblystoma opacum — various stages from 37 to 50 mm. in
length ; (2) a series of horizontal sections of the larvse of Triton

268 INEZ L. WHIPPLE.

alpestris, length from tip of snout to cloaca, 13 mm.; (3) small
specimens of Diemyc tyliis viridcsccns in the terrestrial stage, 32—
68 mm. in length ; (4) larvae of Spclerpcs rnber, and Spelerpes
bilineatns ; (5) larvae of Desmognatlms fnsca from 17.5—25 mm.
in length.

The methods employed in the study of these larval forms were
(i) dissection of the larger ones, prolonged staining in methylene
blue being used in some cases to bring out the cartilage ; (2)
staining in borax carmine and clearing in toto the ventral wall of
the posterior part of the body cavity including the pubo-ischium
and proximal portion of the femur ; (3) horizontal serial sections
of the ventral body wall ; and (4) transverse serial sections of the
posterior part of the body. In the case of each of the younger
stages all three of the latter methods were used.

In 37 and 42 mm. long Amblystoma opacum larvae there is no
trace of the ypsiloid cartilage. The two halves of the pubo-
ischium are quite separate. The muscular abdominal walls show
the two primitive laminae (obliquus externus profundus and ob-
liquus interims) with the rectus abdominis as a ventral continua-
tion of both. In the larger specimens the obliquus externus
superficialis and the transversalis appear as secondary develop-
ments in the form of very thin laminae. There is a noticeable
difference in size of fiber between the medial well differentiated
portion of the rectus abdominis and the more lateral region
which grades imperceptibly into the deep external oblique on the
outside and the internal oblique within. The latter region of the
rectus abdominis (/. e., the rectus abdominis profundus), like the
two primitive laminae with which it is continuous, consists of
large fibers two or three times the diameter of those constituting
the medial portion of the rectus. The latter are thus easily
recognized both in sections and in the in toto preparations.

All of these early stages show that in the somite immediately
anterior to the pubo-ischium, the inner portion of the rectus
abdominis is differentiated into a muscle which deviates sharply
from the general longitudinal course of the rectus abdominis.
Its fibers, which are small like the rest of the medial portion of
the rectus abdominis, arise in the lateral processes and along the
anterior edge of the pubis and extend obliquely medially to be

THE YPSILOID APPARATUS OF URODELES. 269

inserted partly into the linea alba of the first somite. From the
point where the first myocomma joins the linea alba, the insertion
of the muscle follows an outwardly curving line which ends
about half way across the second myotome, and thus with the
insertion of the corresponding muscle of the opposite side maps
out the exact location of the future Cartilago ypsiloides. This
muscle is evidently the M. ypsiloideus posterior and is at this stage
the only definite indication of an ypsiloid apparatus.

In Ainblystouia opacmn larva; of 50 mm. I find that the
ypsiloid cartilage has appeared. It possesses practically the
adult form and relationship, but is very thin, especially toward
the ends of the lateral arms. The stem of the cartilage is quite
separate from the pubo-ischium which at this stage still consists
of two wholly separate lateral halves. I was unable to obtain
larvse of Aniblystoma opacitm between 42 and 50 mm. in length
and can therefore make no statement concerning the earliest
appearance of the ypsiloid cartilage. Its entire absence in the
42 mm. stage, however, considered in connection with its com-
plete formation in the 50 mm. stage in which the two halves of
the girdle are still separate, points conclusively to the origin of
the ypsiloid cartilage independently of the pelvic girdle as a
chondrification of the linea alba of the first somite and of the
deeper portion of the first myocomma. The possibility of such
chondrification of regions of muscular attachments upon which a
special strain is brought is well established, and the lack of
correspondence in this case between the transverse direction of
the myocomma and the curved form of the arms of the cartilage
may be looked upon as expressing a resultant of forces, since the
arms of the cartilage tend somewhat to follow the direction of
the edge of the aponeurosis of the transversalis, the posterior
portion of which, we have already seen, is usually eventually
inserted into it.

The horizontal series of sections of the Triton alpestris larva
show a well developed ypsiloid cartilage. This is wholly sepa-
rate from the pubo-ischium though, as in Amblystoma, articu-
lated with it. The two halves of the pubo-ischium are at this
stage still quite separate.

All of the specimens of terrestrial Dicmyctylns which I dis-

270

INEZ L. WHIFFLE.

_1

sected show the ypsiloid cartilage well developed and with prac-
tically the same muscular attachments as in the adult.

From my necessarily limited study of larval forms, it appears,
therefore, that (i) ypsiloid cartilage is of later origin than the
differentiation of the muscles associated with it, and (2) that it
arises as an unpaired structure in association with these muscles
and at a time when the girdle itself still exhibits its paired nature.
// cannot, therefore, be interpreted as having arisen from the epi-
pubic process, but must be regarded as an independent cliondrifica-
tion in association with differentiations of the innermost portion of

tJic rectus abdominis muscle in
the two somites immediately an-
terior to the pelvic girdle.

Larvae of both Spelerpes
and Desmognathus show, as do
those of lunged forms, an
early differentiation of the M.
ypsiloidcus posterior. As in
the adult lungless forms the
fibers of this muscle insert, in
the absence of the ypsiloid car-
tilage, into the linea alba of
the first somite.

I had thought it possible that these larval stages might even
show a vestige of the ypsiloid cartilage itself, and there is, in
fact, some indication that such may be the case though I have as
yet been unable to obtain the stage necessary to absolutely prove
it. The union of the two halves of the pubis begins at the ex-
treme anterior end, thus forming the future median anterior proc-
ess of the pubis. This union (Fig. 9) appears to occur, not as
a direct fusion of the two halves by the process of chondrifica-
tion of the connective tissue between them, but rather by a fusion
of each half with a median unpaired anlage, which lies in the
linea alba anterior to the girdle. Thus in transverse section this
median portion shows no trace whatever of a paired nature, a
fact which is especially significant when the condition is compared
with that of Necturus larvae in which the median anterior process of
the pubis (Proccssus cpipubicus} shows a paired nature even to its

FIG. 9. Ventral view of the pubo-
ischium of Desmognathus fusca larva (21
mm.) obtained by maceration, < 25-
Drawn with camera. Abbreviations : /,
femur ; /, lateral process ; m, median proc-
ess, at this stage the only median part.

THE YPSILOID APPARATUS OF URO DELES. 271

extreme end. However, I have not succeeded in ^obtaining
either a Spclcrpcs or a Desmognathus of a stage just previous to
the formation of this connection between the two halves of the
pubis, and I am not sure, therefore, that the median unpaired
portion ever exists as a separate cartilage arising like the ypsiloid
cartilage anterior to the pelvic girdle and independent of it. If
it has such a separate origin it is undoubtedly a vestige of the
stem of the ypsiloid cartilage.

C. Tlic Hoinology of the Ypsiloid Apparatus.

There are two diametrically opposed views as to the homology
of the ypsiloid cartilage. One of these is that indicated by the
quotations given earlier in this paper, viz., that the Cartilago
ypsiloidcs is the homologue of the median anterior process of the
pubis (Processus epipnbicns] such as is found either single or
paired in certain of the Perennibranches and Derotremes. More-
over, this homology is extended to include the similarly situated
process in various Selachians, Ganoids and Dipnoans. This
opinion as to the homology of the ypsiloid cartilage has been
held very strongly by C. K. Hoffmann ('73-' 78) and R. Wieders-
heim ('92) and corroborated by certain observations of H. Riese
('91). Aside from general similarity of location of the ypsiloid
cartilage and the Processes epipnbicns the homology is apparently
based upon a continuity of the cartilage tissue of the pubis with
that of the stem of the ypsiloid cartilage. This condition is, as
has been shown, not the usual one in the adult, although Riese
('91) found it to exist in Tylototriton verntcosits, and Wieders-
heim has noted in the case of old individuals of other species a
condition which he designates as a secondary fusion. Wieders-
heim says, however, regarding the adult condition, that " Man
bei histologischer Untersuchung in Allgemeinen viel haufiger
auf verbindende Knorpelbriicken zwischen der Hauptmasse des
Beckens und dem Epipubis stosst, als man dies nach der ein-
fachen Preparation mit Messe und Pincette ervvarten sollte."

In Triton alpestris larvae Wiedersheim found, as I have done,
that the ypsiloid cartilage arises by an independent anlage, but in
larvae of Triton Jielveticus and in a 26 mm. Axolotl he found
the cartilage element continuous, although he adds : " Diese

272 INEZ L. WHIFFLE.

Verbindungszone bestand im vorliegenden Fall nur ventral warts
und werde welter dorsalwarts d. h. gegen das Cavum pelvis zu
durch zellreiches Bindgewebe ersetzt." He finds, moreover, that
" Um diese Zeit stellt das Epipubis eine auf dem Vorderand der
Beckensymphyse aufsitzende spitzhockerige, durchaus unpaare
Vorwolbung dar, welche nur langsam zapfenartig nach vorn
auswachst, und sich erst verhaltnismassig spat in die schon er-
wahnten zwei Aste gabelt." The movable articulation of the
Cartilago ypsiloides is then, according to Wiedersheim's interpre-
tation, a secondary condition.

That there is, on the other hand, an apparent inconsistency
between this idea of the homology of the Cartilago ypsiloides
and the well-established fact of the paired nature of the anlage of
the pelvic girdle, Wiedersheim at least tacitly admits when he
says : " Die Verwischung des urspriinglichen Verhaltens pragt
sich namlich bei Salamandrinen in dreifacher Weise aus, erstens
darin, dass hiervon einer paarigen Anlage des Epipubis ontogen-
etisch nichts mehr nachweisbar ist, zweitens, dass zwischen diesem
und dem iibrigen Becken haufig eine Kontinuitatstrennung be-
steht, und drittens endlich, dass das kop warts schauende Ende
des Epipubis eine secundare Formanderung, eine Gabelung, er-
fahren hat."

A further, and, in my opinion, insurmountable objection to
this homology lies in Wiedersheim's own statement that the
ypsiloid cartilage is of later origin ontogenetically than the girdle
and makes its appearance as late as at the time when the mid-
ventral symphysis of the halves of the pubo-ischium is taking
place.

The second view as to the homology of the Cartilago ypsiloides
is that held by Bunge ('80) and Baur ('91) that the structure is
developed whojly independently of the pelvic girdle to which it
becomes secondarily articulated. With regard to this homology
Bunge says :

" Es (the epipubis) ist eben eine Gebilde sekundarer Art, das
ausschliesslich den Amphibien zukommt, wie ja Ahnliches auch
bei anderen Wirbelthieren beobachtet werden kann, z. B. das
Hypoischium der Saurier . . . Der Ansicht Wiedersheim's dass
der Epipubis als ein erst sekundar von der knorpeligen Pars

THE YPSILOID APPARATUS OF URODELES. 273

pulnca, resp. deren Verlangerung zur Symphysenbildung abge-
gliedertes Gebilde sein kann, da dasselbe sich als einheitlicher
Knorpel vor dem proximalen Ende der Symphyse anlegt, gleich-
falls nicht zugestimmt werden."

Baur, basing his reason upon the fact of the completely paired
origin of the true epigastroid (/. e., epipubic) process, as shown
by Ncctnnts, for example, draws the following very definite con-
clusion with regard to this unpaired, more anterior structure, the
Cartilago ypsiloidcs :

" I believe the ypsiloid cartilages are of secondary origin, de-
veloping independently from the gastroid (i. c., pubic) cartilage.
The long epigastroid of the Chclyiidce is homologue to the
short epigastroid in Testudinidce ; homologue to the anterior
portion of the gastroid cartilage in Ncctrtnts ; homologue to that
portion of the gastroid in salamanders and Dactylctra to which
the ypsiloid cartilages are connected. I consider these cartilages
as a later acquisition and they may develop in any group, Batra-
chia, Pterosaur ia, Monotreniata, Marsiipnlia" So far, even, as
similarity in location between the Cartilago ypsiloides and the
epipubic process of the lower Urodeles is concerned, I have been
unable to find any ground for the homology. In none of the
Urodeles in which the Cartilago ypsiloidcs is lacking have I found
the epipubic process crossing even a single myotome of the
rectus abdominis. In Necturus, the form in which the epipubic
process is most conspicuously developed, there is merely a cor-
responding narrowing of the posterior myotomes of the rectus
abdominis, particularly the first one (Fig. 8).

Whether the ypsiloid cartilage is ever in any case continuous
with the pubis or not, it is very evidently a separate structure,
an independent chondrification of the linea alba of much later
origin than the pelvic girdle to which it sooner or later becomes
articulated. Thus considered, the ypsiloid cartilage presents no
obstacle to the idea of the paired nature of the anlage of the
pelvic girdle. Its existence is moreover explained quite in ac-
cordance with the principle which accounts for the origin of
similar structures (e. g., the sternebrae of Nccturus] in those con-
nective tissue regions where especially strong origin or insertion
of the muscle fibers is necessary. I have not had the opportunity

274 INEZ L- WHIPPLE.

to study either the anatomy or the habits of Dactyletra in which
an apparently similar cartilage to the Cartilago ypsiloides is formed,
and can therefore express no opinion as to this homology.

That the Cartilago ypsiloides of salamanders is homologous
with the marsupial bones of Monotremata and Marsupulia is a
view which has been considered so completely established that
Duges ('55) named this cartilage in salamanders the "marsupial
cartilage." Huxley also accepted this idea of its homology and
it is one of the principal points made by Wiedersheim ('92) in his
Phylogenie der Beutelknochen, in which, of course, this interpre-
tation is quite consistent with his idea that both the ypsiloid
cartilage and the marsupial bones are differentiations of the epi-
pubic process. Moreover, the acceptance of this homology is
indicated in the various names which have been given to the M.
ypsiloidcus posterior, such as pyramidalis (Hoffmann). Leaving out
of account the question as to whether the ypsiloid cartilage and
marsupial bones are of similar origin so far as the pelvic girdle
is concerned, the supposed homology between the two is dis-
proved by their relations to the rectus abdominis muscle. The
ypsiloid cartilage lies dorsal to the main mass of this muscle ;
the marsupial bones are, of course, ventral to it. A comparison
of musculature, therefore, shows the lack of homology of the
ypsiloid apparatus with the marsupial. The musculature of the
ypsiloid cartilage is derived from the deeper layers of the rectus
abdominis and from the transversalis ; from the very poistion of
the marsupial bones, on the other hand, it is evident that the
musculature of this apparatus is derived from the superficial ab-
dominal muscles.

The pyramidalis, which has been homologized with the M.
ypsiloideus posterior is, for example, the most superficial portion
of the rectus abdominis. Further, it is on the wrong side of the
marsupial bone to make the homology a consistent one through-
out, since, if the marsupial bones correspond to the lateral arms
of the Cartilago ypsiloides, a muscle to be the homologue of the
M. ypsiloideus posterior must extend from the outer edge of the
marsupial bone to the pelvis, not as does the pyramidalis, from its
medial side to the linea alba or sternum. The pyramidalis and
the posterior ypsiloid muscle are then homologous only in the

THE YPSILO1D APPARATUS OF URODELES. 275

very general sense that both are differentiations from the rectus
abdominis ; they are differentiated from different layers and in
connection with structures which are not themselves homologous.
They are, in other words, independent differentiations occurring
in widely separated forms .and in response to absolutely different
physiological needs.

PART II. THE FUNCTION OF THE YPSILOID APPARATUS.
A. Respiratory Habits of Lunged Salamanders.

Beyond the statement of the very evident fact that the
ypsiloid cartilage furnishes the attachment for certain of the
abdominal muscles (Wiedersheim, '75), I have been unable to
find, in the literature upon the subject, any explanation of its
function. There is, however, as has already been said, so appar-
ent a correlation in the Salamandrida between the presence of
the apparatus and that of the lungs, that the explanation of its
function will involve, first of all, a discussion of the respiratory
habits of lunged salamanders.

The more obvious respiratory movements of lunged salaman-
ders when breathing air have been very clearly described by
Bruner ('96). In brief, two forms of aerial respiration occur, one
merely a bucco-pharyngeal, the other a pulmonary respiration.
Both of these may be readily observed in the case of any lunged
salamander. The first takes place almost constantly and with
great rapidity. It begins with an enlargement of the bucco-
pharyngeal cavity by lowering the hyobranchial apparatus ; this
results in air being drawn in through the nares. Following this
inhalation is an exhalation in which the floor of the mouth rises
again. These movements follow each other so quickly that the
visible effect is a rapid fluctuation of the throat. The mouth
remains tightly closed during the entire process, and the respira-
tory currents make use of the nasal passages alone.

At frequent, though irregular intervals, during bucco-pharyn-
geal respiration, acts of pulmonary respiration occur. These are
easily distinguished externally from the bucco-pharyngeal form
by the fact that the depression of the floor of the mouth is a
prolonged and exaggerated one, during the latter part of which
a contraction of the M. constrictor naris occurs. According to

276 INEZ L. WHIPPLE.

Bruner, the effect of this contraction is to completely close the
external nans. My own observations of Diemyctylus and the
Ainblystomas made by the aid of a lens do not, however, corrob-
orate this statement, since I have frequently seen the external
nares fail to close completely during pulmonary respiration,
although there is always an almost complete closure.

As a result of the prolonged depression of the floor of the
mouth air is first drawn in through the open nares, as in bucco-
pharyngeal respiration. This part of the process is known as
aspiration. During the latter part of the act of depression, how-
ever, when the external nares are closed, air is drawn from the
lungs into the mouth through the opened glottis and the air in
the mouth thus becomes a mixture of pure and impure air. This
part of the process is termed expiration. When the floor of the
mouth rises again some of this mixed air is forced into the lungs,
the external nares being still closed. This constitutes the proc-
ess of inspiration. Finally the external nares are opened again,
and the fluctuating movements of bucco-pharyngeal respiration
are resumed.

In addition to these two methods of aerial respiration the
lunged salamanders which have come under my observation pos-
sess, when in the water, an aquatic bucco-pharyngeal respiration.

The Gages ('86^, '91) have reported such an aquatic respira-
tion for Diemyctylus as well as for some of the lower Urodcles.
Their statements are, however, indefinite as to the exact method
by which the water is alternately taken into and expelled from
the mouth. O. P. Hay ('89) seems to have made more exact
observations upon Amblystomas, of which he says that " streams
of water are drawn in through the nostrils and this water is then
expelled at intervals by the mouth." This is precisely the
method of bucco-pharyngeal respiration of water which I have
many times verified with a lens by the aid of solid particles (car-
mine or sediment) suspended in the water in which specimens of
Dicmyctylus and Amblystoma were submerged. The muscular
act seems to be exactly the same as in the bucco-pharyngeal
aerial respiration, but owing to the heavier fluid the act is a much
slower one, though varying in depth and rapidity with the activity
of the specimen, as the accompanying tabulation of observations

THE VI'SILOID APPARATUS OF URODELES.

277

shows. Moreover, the expulsion of water takes place through
the slightly opened mouth as well as through the nares. Ambly-
stoina opacnin, which is said to be the most terrestrial of all the
Amblystomas, showed the least readiness to adopt this aquatic
mode of respiration, the pharyngeal movements being very feeble
as if they occurred in response to an almost forgotten instinct.
They are probably not of sufficient respiratory value to support
life, since these specimens die in a short time if compelled to re-
main in the water.

RECORD OF OBSERVATIONS OF AQUATIC BUCCO-PHARYNGEAL
RESPIRATION IN DIEMYCTYLUS VIRIDESCENS.

(In each experiment a different individual was used.)

a

Hi

*** is

<°

o

K G

| Q.< .

^ s

0. 2 3

• JJ

.2 3 fcs

Vl_ C
O -"

Nature of Respiratory

Activity of Specimen Dur-

i 4,.1-G

S

«S

cS

Acts.

ing Experiment.

£ J? «5

£ u

J "

.2 B .
2'Ei!

0

>K ft

M -*

C ^

3 CX3

15 <

J

Cfl

Q

No. I. 21.4

23 19

10

Shallow.

Slightly active.

No. 2. 15.14

20 II 7

Deep.

Inactive.

No. 3. 10.04

13 9

5

[Not recorded.]

Slightly active.

No. 4.

s

18

3

10

Detailed Record for Each Minute of Experiment Aro. 4.

4

Very shallow.

Inactive.

9

One very deep. Walking slowly.

3

Very deep. Inactive.

I^

Shallow. Active.

4

Deep.

Inactive.

9

Shallow.

Inactive, after a period

of activity.

18

Shallow.

Active.

6

One very deep, end- Inactive.

ing in a prolonged

gape.

3

Very deep.

Inactive.

12

Shallow.

Slightly active, follow-

ing a period of great

activity.

RECORD OF OBSERVATIONS OF AERIAL BUCCO-PHARYNGEAL AND
PULMONARY RESPIRATION IN DIEMYCTYLUS VIRIDESCENS.

No. of
Experi-
ment.

Average No.
of Bucco-
phar. Resp.
per Min.

* Largest
No. for One
Minute.

Smallest
No. for One
Minute.

Average No.
of Pulmo-
nary Resp.
per Min.

Largest
No. for
One Min.

Smallest
No. for One
Minute.

Duration of
Observa-
tions in
Minutes

No. i.

No 2.

I87.7
125-75

213

146

H5
104

I.I

4
it record e

0

d.]

10

278 INEZ L. WHIPPLE.

The method of change of respiratory habit necessitated by the
transition from one median to the other is interesting. When a
Dicmyctyhts which has been breathing air is submerged in water,
bucco-pharyngeal respiration of water begins almost at once.
Amblystomas, being less thoroughly aquatic, postpone this change
of habit for a longer or shorter time. The nares in this case are
at first tightly closed and if the animal is kept submerged for
only a few minutes it may not establish aquatic respiration at all.
When a specimen which has fully established the aquatic habit
of respiration is taken from the water there is evinced more or
less mechanical difficulty in reestablishing the aerial habit. This
difficulty arises from the fact that the nasal passages are filled
with water which must be removed before rapid, unimpeded res-
piration of air can occur. The efforts to do this involve forced
and greatly exaggerated depressions of the floor of the mouth, a
device which may prove efficacious in two ways, first, by draw-
ing the water from the nasal passages into the mouth, and sec-
ond, by drawing from the lungs a supply of air which can be
used to force the water out of the nasal passages through the ex-
ternal nares. The transition from aquatic to aerial respiration
may thus involve much effort and a considerable loss of time.
I have observed specimens of Diemyctylus to consume ten min-
utes or more before perfectly normal aerial respiration was estab-
lished. Amblystomas make the transition more quickly.

In connection with this mechanical difficulty of rapid transi-
tion from aquatic to aerial respiration, some lunged salamanders,
notably Diemyctylus, have acquired for use when in the water a
modification of the ordinary method of pulmonary respiration.
Frequently air must be taken into the lungs during the brief
period when by a rapid swimming to the surface, sufficient mo-
mentum has been acquired to force the head for an instant out
of the water. It is evident that under these conditions the nos-
trils are utterly useless as air passages, as they are filled with
\vater. Moreover, even if they were empty of water, or could be
emptied in so brief a time, the ordinary method of drawing in air
through such narrow passages is far too slow to be made use of
here. The method employed is, therefore, a quick, gulping
motion by means of which the water in the mouth is replaced by

THE VPSILOID APPARATUS OF URODELES. 2/p

air. This is immediately followed, as the head again returns
into the water, by a forcible swallowing motion as a result of
which the air is forced from the mouth partly into the lungs and
partly out through the nostrils. Of those spe'cies the habits of
which I have studied, Dicuiyctylus viridescens accomplishes with
the greatest ease the act of taking air into the lungs in this way,
an observation quite in harmony with the fact that Dicmyctylus
has the reputation of being the most thoroughly aquatic of our
American salamanders.

The Gages ('86tf, r) have shown that, in general, an animal
having a mixed aquatic and aerial respiration depends mainly
upon the latter for its supply of oxygen. I am not convinced,
however, that this is of necessity true in the case of lunged sala-
manders living under aquatic conditions. Diemyctylus and the
Amblystomas, it is true, not only swim frequently to the surface
and take in air, but, if it is possible, will partly crawl out of the
water and for a shorter or longer time each day will breathe air
normally. However, to test the absolute physiological necessity
for aerial respiration, I experimented as follows : Several speci-
mens of Diemyctylus were enclosed in small wire cages which
were immersed to a depth of about 15 inches in a small tank of
running water. To prevent the collection of bubbles of air upon
the inside of the wire, the cages were frequently shaken to remove
the bubbles while they were still too small to be used in breath-
ing and thus vitiate the experiment. To ensure this frequent
agitation during the night when personal attention to the matter
was inconvenient, I used the simple device of placing a large
and lively specimen of Necturus in the tank with the cages. The
Nectiinis, being nocturnal in its activities, accomplished quite as
efficiently the duty of keeping the cages free from air as was
done during the day time by my own exertions. The precaution
was taken, moreover, to make it impossible for any activity of
the Necturus to lift the cages out of the water.

For periods varying from seven to ten days specimens of Die-
myctylus were thus kept completely submerged and they remained
in an active condition and apparently suffered no inconvenience as
a result of the experiment. The capillaries of the skin, however,
as observed by means of a lens, were much more distended with

28O INEZ L. WHIFFLE.

blood than those of a specimen which had meanwhile lived a free
aquatic life with access to the air. Apparently the skin, which,
being supplied with capillaries may be looked upon as an acces-
sory respiratory apparatus, had proved itself, in the emergency,
equal to the extra demand made upon it.

An interesting effect of the disuse of lungs in this experiment
showed itself in the great difficulty with which specimens, thus
confined to the water, reestablished the habit of filling the lungs
with air when they were released from their imprisonment. Or-
dinarily, when a Diemyctylus swims to the surface and takes in
air by the gulping process already described, there is an abun-
dant visible proof of the fact that air has entered the lungs in the
increase in girth of the body and especially in the immediate in-
crease in buoyancy to such an extent that the specimen which
before had been able to sustain itself in the water only by active
swimming, suddenly becomes lighter than water and passively
floats. When, however, specimens which had for several days
been prevented from using their lungs were once more set free in
the water and swam to the surface, although great gulps of air
were taken, there was not the usual subsequent increase in buoy-
ancy and the air escaped immediately in large quantities from
both nostrils as the head sank again below the surface of the
water. The effort to fill the lungs was repeated many times in-
terspersed with intervals of rest lasting 15 or 20 minutes, so that
several hours elapsed before any effect seemed to be produced
upon the disused lungs. One could from the fruitless efforts of
the animals imagine the lungs in a collapsed condition, the inner
surfaces in contact with each other, and therefore resisting the
entrance of air ; and such, indeed, was found to be the case in
other specimens which had been similarly confined under water
and then killed without having had access to the air.

B. The Hydrostatic Habits of Lunged Salamanders.
Having discussed the various methods of respiration of lunged
salamanders we are now prepared to describe those particular
habits which involve, as will be shown, the use of the ypsiloid
apparatus. Since my more extended observations have been
made upon Diemyctylus viridesccns, this species is the one which

THE YPSILOID APPARATUS OF URODELES. 28l

will be referred to almost exclusively in the following discussion.

As has been said, the adult Dicinyclylus is the most aquatic of
our American salamanders. To anyone who has observed, even
casually, the activities of this little animal in the water, its abso-
lute ease under aquatic conditions must have been evident. If
the specimens are in a deep aquarium this physical ease is very
readily observed. Occasionally they may be seen at the bottom
where they walk about or take rapid little swims to higher levels
from which, as soon as the swimming motions cease, they pas-
sively sink to the bottom again. More often they may be found
at the very surface of the water where they float with great ease
or rest upon the aquatic plants, sometimes supporting themselves
upon these by the fore limbs and lifting the entire head above
the water. Frequently, moreover, they will be seen suspended
in the water at a greater or less depth, where they have the power
to swim lazily to and fro with hardly a perceptible muscular
action, to paddle about, using all four feet as propellers, or to
dart swiftly through the water by means of a rapid lashing with
body and tail, the legs meanwhile being closely pressed to the
sides of the body. For many minutes they will sometimes re-
main absolutely motionless in the water, the body kept in place
by the mere contact of a foot, or even of a single toe, with some
plant or other stationary object.

These facts indicate that while the specific gravity of Dicniyc-
tylus is never far from one, it varies slightly, as shown by the
passive sinking, suspension or floating of the body at different
times. What are, then, the mechanical means by which these
changes in buoyancy are accomplished ?

If our observations begin with a Diemyctylits at the bottom of
the aquarium, with a specific gravity greater than one, it will be
found that sooner or later the animal will swim rapidly to the sur-
face, and, by the modified process of pulmonary respiration already
described, will take into the lung sufficient air to cause the body
to float (Fig. 10, b\ the minute portion of the back which appears
above the surface bearing witness to the fact that the specific
gravity has now become slightly less than one. The animal
may remain in this condition for a longer or shorter time. If he
swims to a lower level, the moment his motions cease his body

282 INEZ L. WH1PPLE.

rises slowly to the surface again. Usually, however, the increase
of buoyancy is soon followed (within a minute or two) by the
emission from the lungs through the mouth of one tiny bubble of
air after another, seldom more than two or three in all, until his
buoyancy is so perfectly adjusted that his specific gravity is ex-
actly one. In this condition he can go about at ease, or remain
motionless at any depth, and it is apparently only when he de-
sires to sink to the bottom and remain there with some stability
that by the emission of still more air the specific gravity is made
sufficiently great to serve the purpose. There always occurs,
however, a gradual loss of buoyancy even when there is no
further emission of air, a loss which I have never observed to be
made good until the animal again swims to the surface and takes

in more air.1

Diemyctylus shows a still further delicacy of adjustment to its
aquatic environment, since, under any condition, whether floating,
suspended in the water or resting on the bottom, there is the
power to change, without the slightest swimming motion, the
direction of the long axis of the body. This adjustment may be
best observed when the animal is suspended motionless in the
water ; since then all other factors which produce change of po-
sition are eliminated. The usual position of Diemyctylus when
thus suspended is one in which the anterior end of the body
slants slightly downward (Fig. 10, b\ From this position the
whole body, without the slightest bending, may swing through a
vertical angle of perhaps 30° until the head is directed upward
instead of downward. This change of direction is accompanied
by a striking change in 'the shape of the animal. When the
poise is such that the head slants downward, there is a pro-
nounced bulging of the lateral and ventral walls of the posterior
third of the body cavity, particularly noticeable in the angle be-
tween the ilium and the vertebral column, as though some
mechanism within were exerting an outward pressure. As the

1 This phenomenon of loss of buoyancy without the emission of air is worthy of
careful investigation. It is probably not attributable to a mere compression of air (as
in the case of the air-bladder of the fish), since there is no subsequent increase of
buoyancy without taking in more air. It must be due to an actual loss of gas from
the lungs, probably owing to the excess of the volume of oxygen used over that of
CO2 and other gases given back to the lungs.

THE YPSILOID APPARATUS OF URODELES.

283

body swings upward, however, there occurs a marked constric-
tion of the posterior part of the abdominal cavity, often so pro-
nounced that the ventral wall (that is, the ypsiloid region) is
drawn sharply upward (xFig. 10, a). In this condition the ventral
contour of the body exhibits an angle between this posterior and
the more anterior region. These changes of shape may best be

FIG. 10. Diemyctyhts ririJesctns ; (a) showing the body directed upward in
swimming as a result of the compression of the posterior portion of the body cavity
through the action of the ypsiloid apparatus; (^ showing a characteristic floating
position with the posterior portion of the body cavity expanded and the anterior end
of the body depressed.

seen in a specimen which has not been fed for several days, since
they are partially masked by the presence of masses of food in
the digestive tract.

They occur, moreover, not only during this inactive change in
the direction of the long axis of the body, but changes of direc-

284 INEZ L. WHIPPLE.

tion during active swimming involve a constant exhibition of cor-
responding changes of contour ; the constriction of the posterior
abdominal region occurs when the swimming motion is upward
(Fig. 10, a], the prominent bulging when the motion is down-
ward. Further, a sudden change from a downward to an upward
direction is preceded by an exaggerated constriction of the pos-
terior part of the cavity, an act which conspicuously involves the
sudden vigorous inpulling of the ypsiloid region. Evidently there
is in operation some mechanism for controlling tlie direction of tJie
body, ivhet/icr at rest or in motion, through the control of the rela-
tive buoyancy of anterior and posterior ends.

C. The Hydrostatic Mechanism of Lunged Salamanders.

Turning now to the anatomy of the ypsiloid region, the
explanation of this hydrostatic mechanism becomes very simple.
The contraction of the M. ypsiloideus posterior exerts a strong
pull upon the whole ypsiloid cartilage. The origin and insertion
of this muscle are, however, so nearly in the same plane as the
fulcrum (the articulation of the ypsiloid cartilage with the pubis)
that it seems at first a question as to whether the cartilage would
be bent upward or downward (/. e., dorsally or ventrally) by the
contraction of this muscle alone. It must be remembered, how-
ever, that the muscle is inserted into the upper (dorsal) surface
of the stem of the cartilage and also that the more strongly
developed portion of the muscle has its origin in the lateral por-
tion of the pubis, a region which owing to the convexity of the
body is slightly higher (more dorsal) than the insertion of the
muscle and the articulation of the cartilage. Moreover, outside
of the whole apparatus there are muscular walls (external oblique
and rectus abdominis) which would resist any tendency to bend
the cartilage downward, and with origin and insertion on so nearly
the same plane as the fulcrum it requires only a slight resistance
of this sort to turn the scale. Other muscles, moreover, are
attached to the ypsiloid cartilage and cooperate with the ypsi-
loideus posterior to determine the direction of motion. The con-
traction of the ypsiloid portion of the transversalis exerts a decided
upward (dorsalward) pull upon the ypsiloid cartilage while the
anterior ypsiloid muscle, pulling upon the arms of the cartilage

THE YPSILOID APPARATUS OF URODELES. 285

from the linea alba, at least lends a certain steadiness to the
apparatus while at the same time it cooperates with the ypsi-
loideus posterior and the ypsiloid portion of the transversalis to
pull the arms strongly inward.

Corroboration of the above explanation is furnished by those
occasional specimens which happen to have been preserved with
the ypsiloid muscles contracted. These cases show that the
effect of the concerted contraction of the muscles associated with
the ypsiloid cartilage is not only to bend the stem of the cartilage
upward (dorsally) at its articulation with the pelvic girdle but to
curve the flexible arms upward and inward (medially). Evidently
the result of the contraction of the three pairs of muscles con-
nected with the ypsiloid cartilage is a decided constriction of the
posterior region of the abdomen and a consequent compression
of the organs contained within^ it.

To understand in what way this action of the ypsiloid appa-
ratus controls the relative buoyancy of the anterior and posterior
regions of the body, the shape and position of the lungs must
be considered. The lungs of Dicinyctylns are exceedingly simple
structures, mere sacs with no trace of the usual amphibian con-
dition in which the cavity is subdivided by partial partitions. It
seems impossible, in fact, that such very simple structures with
so small a supply of blood can justify their existence merely as
respiratory organs. In shape, also, the lungs of Dicinyctylns are
peculiar. Narrow anteriorly, they widen gradually and round off
quite abruptly at the posterior end. The whole form is most
adequately described, perhaps, as club-shaped. The statement
often given as to the size of the lungs (viz., one third to one half
of the length of the body cavity) I find quite incorrect when the
observations are made upon freshly killed specimens. If the
lungs of such a specimen be inflated through the glottis not even
sufficiently to float the body in water (and therefore not unduly),
subsequent dissection shows that the lungs extend the entire
length of the body cavity so that their rounded, bulging, free
ends lie on either side in the angle between the ilium and the
vertebral column.

It is thus easy to see the cause of the bulging of the lateral
and ventral walls of the posterior part of the body cavity.

286 INEZ L. VVH1PPLE.

Moreover, it is also evident that with the lungs inflated and with
no muscular constriction of this posterior region this portion of
the body will possess greater relative buoyancy than the anterior
portion, or in other words the long axis of the body will assume
its ordinary position with the anterior end slanting downward.
As soon, however, as the ypsiloid apparatus is brought into
action, the resulting pressure upon the posterior abdominal
organs becomes exerted upon the clavate ends of the lungs thus
forcing the air in them forward. The effect is to immediately
increase the buoyancy of the anterior region of the body and
diminish that of the posterior region.

On the other hand, when the muscles relax, the pressure of
the air in the lungs, as well as the elasticity of all the parts con-
cerned, causes the return of the air to the posterior region again
and the bulging of the body wall in this region occurs as before.
It seems probable that the superficial portion of the rectus
abdominis which is attached to the ventral side of the ypsiloid
cartilage (its fibers extending from the lateral arms to the stem
and to the anterior margin of the pubis), may assist in straighten-
ing the curved ypsiloid cartilage, since when the cartilage is in
the bent condition these fibers lie upon its convex (ventral) side.

Thus the ypsiloid cartilage and the muscles connected with it
constitute, together with the lungs, the mechanism by means of
which the relative buoyancy of anterior and posterior ends of the
body may be controlled. One needs only to witness the constant
use of this hydrostatic apparatus by Dicinyctylits to understand
how completely the absolute ease of the animal under aquatic
conditions is due to its power to control the direction of its body
by means of the rapid adjustment of the relative buoyancy of
anterior and posterior ends. The ypsiloid apparatus is thus of
vital importance in the free-swimming aquatic life of a species
which, like Diemyctylus, depends for its food supply upon its ease
of movement in water at any depth.

Opportunity has not been afforded me to study extensively the
aquatic activities of lunged forms other than Diemyctylus. Seve-
ral specimens of Ainblystoma punctatnui and Amblystoma opacum l
have, however, been observed with regard to this point. Both

1 See p. 257 for statement with regard to the lungs of Amblystoma ofacimi.

THE YPSILOID APPARATUS OF URODELES. 287

of these species, although capable of much less perfect adjust-
ment to aquatic life, resemble Diemyctylus in the fact that the
lungs are used as hydrostatic organs. Almost the first act of an
Amblystoma when it is placed in deep water is to swim to the top
and take in the air sufficient to float the body. I have not ob-
served an Amblystoma opacum to become sufficiently at home in
the water to do more than to remain floating at the surface.
Amblystoma punctatnm will, however, after a little while, appear
quite at ease, crawling about the bottom, floating at the surface,
or swimming around with much freedom.

Although the greater thickness and breadth of the rectus
abdominis of Amblystomas prevent the visible exhibition of the
action of the ypsiloid region during aquatic life, changes of shape
of the posterior lateral walls of the body are often observable.
Upon one occasion a specimen of A. pnnctatnin was observed
floating in a horizontal position at the surface of the water. Sud-
denly there was a violent contraction of the posterior abdominal
walls particularly noticeable in the lateral region, and immediately
the position of the body became so nearly vertical that the head
was sufficiently protruded from the water to make aerial respira-
tion through the nostrils possible. This observation not only
proved that the ypsiloid apparatus is functional in the control of
the hydrostatics of Amblystoma punctatum, but it suggests the
application of its action as a means for bringing the floating body
into such a position that the respiration of air may occur. In the
case of imperfectly aquatic forms this use of the mechanism might
at times be extremely important. For example, Amblystoma
opacum will frequently, if compelled to remain in the water, take
this same almost vertical position at the surface with the nostrils
out of the water and is thus able to breathe air.

That the lungs of Amblystomas are of greater importance as
respiratory organs than are those of Diemyctylus is evidenced by
the fact that they are more complicated in structure and there-
fore present a much larger respiratory surface. They are, how-
ever, like the lungs of Diemyctylus, of sufficient length when
moderately inflated to extend the entire length of the body cavity
and would therefore lend themselves readily to the hydrostatic
function in connection with the ypsiloid apparatus.

288 INEZ L. WHIPPLE.

With regard to the relative importance of the lungs of sala-
manders as respiratory and as hydrostatic organs, it is a signifi-
cant fact that in no case have I found that a Diemyctylus or an
Amblystomct which was out of the water and using its lungs
normally in air-breathing, had sufficient air in the lungs to float
the body when it was dropped into water. Almost the first act
under these circumstances is to swim to the top and take in a
quantity of air sufficient to float the body. This indicates plainly
the secondary adaptation of the lungs as organs of buoyancy and
it is easy to see how in the case of a species like Diemyctylus
which has become thoroughly aquatic, the hydrostatic function
might become of so much greater importance than the respira-
tory as to account for the apparent degeneration of the lungs as
respiratory organs which is indicated by their simplicity of struc-
ture. Moreover, it is easy to understand how a mechanism such
as the ypsiloid apparatus for controlling relative buoyancy of the
anterior and posterior ends of the body, while useful to any lunged
form for the longer or shorter periods during which it normally
stays in the water, would become especially perfected in its action
in the case of a thoroughly aquatic species.

D. Negative Evidence Furnished by Litngless Salamanders.

In corroboration of the above conclusions as to the function of
the ypsiloid apparatus of lunged salamanders, we have the
negative evidence furnished by the habits of lungless forms in
which, with the single exception of Salamandrina perspicillata,
the Cartilago ypsiloides is apparently lacking.

These forms have, of course, no hydrostatic poxvers. They
are thus, unlike the lunged salamanders, incapable of a comfort-
able, free-swimming existence at any depth, but owing to lack
•of hydrostatic organs they must remain for the larger part of the
time at the bottom. As Camerano ('94, '96) has pointed out,
although certain lungless species may be more or less aquatic,
their activities, even when in the water, are terrestrial. Various
species of Spelerpes, PletJiedon and DcsinognatJins, for example,
will at first, when placed in an aquarium, swim to the surface,
then around and around the edge of the aquarium, as if seeking
a means of escape, but the instant that active swimming ceases,

THE YPSILOID APPARATUS OF URODELES. 289

the body sinks clumsily and heavily to the bottom where they
remain until disturbed, or until another effort is made to escape.

Consistently with the lack of hydrostatic apparatus, lungless
forms show on the whole, little power to adapt themselves to
aquatic life. Most of them are terrestrial in habit, some of them
as, for example, PlctJwdon cinerens and P. glutinosus, being found
far from any water supply, while the arboreal Autodax furnishes
an extreme illustration of total abandonment of aquatic life.
Those species, which, like Desmognathus, live along the banks
of small streams, apparently never seek deep water, nor do they
remain long submerged in shallow water, but often are found
lying with the body in the water and the head (or at least the
nostrils) out.1

Lungless forms, moreover, exhibit less adaptation to aquatic
life in their respiratory powers, since unlike the lunged forms
there is practically no aquatic bucco-pharyngeal respiration.
When the animal is submerged, the nostrils, which have been
widely open during aerial bucco-pharyngeal respiration, close at
once and, so far as I have been able to carry my observations,
the nares remain closed as long as the animal is in the water.
In a few cases I have observed occasional feeble movements of
the floor of the mouth, which were undoubtedly attempts at
bucco-pharyngeal respiration, but even then the external nares
were closed and the water was both drawn in and expelled
through the slightly opened mouth.

Spelerpes ruber proved to be the most aquatic of all the
lungless forms with which I experimented. One specimen lived
for weeks at the bottom of the aquarium and was never observed
to attempt to come to the surface except when disturbed. On
the other hand, specimens of Desmognathus fnsca invariably
escape from the water when not caged, while Plethodon glutinosus,
Spelerpes guttoliiieatus and Spelerpes bilineatus make frantic attempts
to do so, but since they do not possess the power to crawl up
the surface of the dry glass as Desmognathus does, their efforts
are unsuccessful. This aversion to aquatic life, is, however,
apparently not due to an actual physiological need, for speci-
mens of DesinognatJius fusca, Plethodon glutinosns and Spelerpes

1 See my article now in press on the "Naso-labial Groove of Lungless Sala-
manders."

290 INEZ L. WHIFFLE.

guttolineatus suffer no physiological inconvenience when com-
pelled to remain under water as in the experiment described
above in which they were confined in wire cages immersed in
running water for a week or more without access to air. They
invariably, however, as soon as released, swam to the surface of
the water and tried to escape, thus showing a strong instinct to
seek terrestrial conditions even though their physiological needs
were satisfied. It cannot, therefore, be argued that the aversion
to aquatic life is due to lack of lungs and correlated ypsiloid
apparatus but rather that the long continued terrestrial habit has
resulted in the loss of these structures. An aquatic lungless
form like Spclerpes ruber must then be regarded as having sec-
ondarily reacquired its aquatic habits.

Salamandrina, a lungless form which possesses an ypsiloid
apparatus, is an interesting exception but by no means an em-
barrassing one, since it belongs to a wholly different group of
salamanders and thus represents a case of analogical resemblance.
It might be expected a priori to show less divergence in structure
from the lunged salamanders than do the members of the families
Plethodontid(Z and Desmognathidce, since its departure from the
habits of the rest of its own family, the Pleurodelidce, is presum-
ably comparatively recent. Thus we find that it still possesses
arytenoid cartilages and rudiments of lungs. Similarly the ypsi-
loid apparatus persists, though Wiedersheim ('75) called attention
to the fact that the cartilage is less strongly developed than in
the case of Triton in which it usually undergoes more or less
calcification. There are additional evidences in the condition of
the muscles of the region (already described) that slight degen-
eration of the apparatus has taken place.

Of course a secondary adaptation to some other function might
tend to preserve the apparatus, but as I have not yet had the op-
portunity to observe the living Salamandrina I can make no
statement as to the probabilities of such secondary adaptation.

E. The Hydrostatic Functions of Perennibranches
and Derotremes.

The following table expresses briefly the conditions of lower
Urodeles with reference to the possession of lungs and ypsiloid

THE VPSILOID APPARATUS OF URODELES.

291

apparatus, together with a general description of body form
and habits : l

Name.

Lungs.

Ypsiloid Appa-atus.

Form of Body.

Habits.

Net turns.

Present.

Wholly lacking.

Short and broad
with compressed tail.

Lives largely on
bottom. Floats
only under abnor-
mal conditions,
i. e., when water
becomes foul.

Proteus.

Present.

Cartilage lacking;
facts concerning
musculature un-
known.

Slender with com-
pressed tail.

Facts unknown.

Siren.

Present.

Wholly lacking.

Slender and eel-
like.

Burrows and
swims.

Typhlo-
molge.'2

Lacking.

Cartilage lacking ;
facts concerning
musculature un-
known.

Slender.

Crawls about at
bottom.

Axulotl.

Present.

Present.

Short and broad ;
tail compressed.

Swims.

Crypto-
branchus.

Present.

Present.

Short and broad ;
tail compressed.

Swims; is said
to come frequently
to the surface for
air.

Amp hi-
ii in a.

Present.

Wholly lacking.

Slender and eel-
like.

Burrows, swims
with great ease
and comes fre-
quently to the sur-
face for air.

A comparison of the facts given in this tabulated form shows
that the following classifications may be made :

1. Forms with lungs and ypsiloid apparatus — Axolotl, Cryp-
tobranchus.

2. Forms with lungs but without an ypsiloid apparatus -
Siren, Nee turns, Proteus, Amphiuma.

3. Forms with neither lungs nor ypsiloid apparatus- - Typh-
lomrtge. Of the first group it may be said that since Axolotl is
a larval form of a salamander which has lungs and an ypsiloid
apparatus, its condition is exactly what one would expect to find

1 The statement of facts relating to thi habits of these forms is the result of my
own observation only in the case of Necturus and Amphiuma. In other cases the in-
formation has been derived from various scientific works.

2 Miss Emerson (1905) has shown valid reasons for regarding Typhlomolge the
permanent larval form of one of the Plethodontidtz.

INEZ L. WHIFFLE.

from the facts already given as to the development of the ypsi-
loid apparatus in other Amblystoma larvae. Moreover, the lungs
are doubtless functional as hydrostatic organs and the ypsiloid
apparatus probably serves its usual purpose in controlling this
function.

I have not had the opportunity to observe the habits of Cryp-
tobranchns. From the descriptions which have been given of its
habits, however, one can readily believe that its lungs and ypsi-
loid apparatus are important, functionally, as hydrostatic organs.
The large size of the ypsiloid cartilage and the well developed
state of its muscles is, in itself, an indication of its functional
value. Moreover, the body is relatively short and the three pos-
terior somites of the trunk, that is, the region .which would be
constricted by the action of the ypsiloid apparatus, form a suffi-
ciently large proportion of the entire length of the body to
render such constriction effective.

With regard to the second group two general types of body
form may be observed - - the short, stout, heavy body of Nec-
tnrus, and the eel-like form such as Amphiuma and Siren. Cam-
erano ('96), has expressed his belief that in all these forms the
lungs have an important hydrostatic function. My own obser-
vations of the living animals have been confined to Necturus and
Amphiuma. The former I have never observed to float except
upon one or two occasions when the water has become very
foul. Under ordinary conditions the Nectnnis in captivity stays
at the bottom of the aquarium, often hiding in crevices between
rocks. Occasionally, especially if much disturbed, it will swirn
to the surface and take in air through the mouth by a gulping
motion. This is usually followed by an immediate escape of air
through mouth and gill-slits as the animal sinks slowly to the
bottom. Undoubtedly the natural habitat of Nectnnis is at the
bottom ; it has, therefore, no use for an apparatus controlling the
hydrostatic function of the lungs. Moreover the total lack of
all traces of an ypsiloid apparatus indicates at once that the
species has not descended from one with such an apparatus,
since the muscular vestiges in wholly lungless forms show how
very slowly the degeneration of such an apparatus occurs.
Neither can we believe that the Nectnnis is a permanent larval

THE YPS1LO1D APPARATUS OF URODELES. 293

form of a lunged salamander, since such a larval form would cer-
tainly show traces of an ypsiloid apparatus. In this connection
it may be noted that H. H. Wilder in a footnote to Miss Emer-
son's recent work on Typldomoge (1905) stated that Kingsbury's
(1905) suggestion that Nectums maybe a permanent larva of one
of the Plethoiiontidte is untenable, since all of the Pleth&dontidce
are lungless. It now seems that in view of its lack of ypsiloid
apparatus, Nectums is ruled out of all possible claim as a sala-
mander larva.

My observations of Ainphiuma give evidence of greater hydro-
static powers than in the case of Nectums. While this animal,
like Nccturus, spends its time largely at the bottom of the aqua-
rium, burrowing if the mud. is sufficiently soft, it occasionally
comes to the surface of the water to breath air. To accomplish
this, the tip of the snout is thrust out of the water, the body
being sustained at the surface by its own active serpentine move-
ments. Air is taken into the lungs by the process of pulmonary
respiration already described for lunged salamanders. As the
air enters, the buoyancy of the body increases perceptibly, often
until the body actually floats. I have seldom, however, observed
a specimen to retain this buoyant condition for more than a few
minutes. It will swim down, allowing bubbles of air to escape as
it goes, until it rests with its usual stability upon the bottom. I
have sometimes observed, however, during these few minutes,
marked constrictions of anterior or posterior body regions with
corresponding changes of buoyancy of these regions, such changes
apparently aiding somewhat in directing the eel-like motion of
the animal. From my own somewhat limited observations I
should conclude, however, that the lungs of the Amphiuma sub-
serve mainly the respiratory function although there is a possi-
bility of use for hydrostatic purposes. In any case it is very
evident that an ypsiloid apparatus affecting as it does, only two
or three body somites, would be practically useless as an acces-
sory hydrostatic apparatus in the case of an animal with a long
eel-like body comprising a very large number of somites like that
of Amphiuma. This fact would in itself account for the lack of
such an apparatus in these slender, eel-like forms, Amphiuma3&&
Siren. The lack of all vestige of both ypsiloid cartilage and

294 INEZ L. WHIPPLE.

muscles does not however as in the case of Necturus preclude
the possibility of descent by degeneration from some higher
lunged form, since the entire pelvic region shows numerous signs
of degeneracy.

With regard to Typhlomolge, Miss Emerson (1905) has already
shown conclusively the probability that this form is a permanent
larva of a lungless salamander. Unfortunately I have not at
hand the means for ascertaining whether in this form, as in the
known Plethodontidte, vestiges of ypsiloid muscles occur, but
Miss Emerson mentions the failure of the cartilage as one of the
characteristics of Typhlomolge. My proof that the use of the
ypsiloid cartilage is correlated with the hydrostatic function of the
lungs, therefore merely strengthens Miss Emerson's argument
that Typhlomolge is the larva of a lungless form.

In conclusion, I wish to acknowledge my indebtedness to Dr.
Harris H. Wilder for much practical assistance in the preparation
of this paper.

SUMMARY.

1. The ypsiloid apparatus is, with the exception of Crypto-
branchus, confined to the suborder Salamandrida. It has arisen
in response to the physical need of controlling the direction of the
body in water through the adjustment of the relative buoyancy
of the anterior and posterior ends. Its function is therefore
closely correlated with the hydrostatic function of the lungs.

2. In origin the ypsiloid cartilage is independent of the pelvic
girdle. Its stem arises as a chondrification of the linea alba of

o

the somite immediately anterior to the pelvic girdle. The arms
are more complex in origin since the process of chondrification
involves not only the myocomma anterior to the above named
somite but also the outer edge of the aponeurosis of the trans-
versalis muscle.

The Cartilage ypsiloides is therefore not homologous either
with the Proccssns epipubicits of the lower Urodelcs or with the
marsupial bones of certain mammals.

3. In the Plctliodontidce and Desmognathidce, in which the
lungs have wholly degenerated, a correspondingly complete
degeneration of the ypsiloid cartilage has occurred, although

THE VPSILOID APPARATUS OF URODELES. 295

vestiges of the ypsiloid musculature remain to indicate the former
possession of the apparatus.

4. This interpretation of the function of the ypsiloid apparatus
throws some light upon the systematic position of certain of the
lower Urodclcs. The more obvious conclusions are :

a. That forms with lungs but without vestiges of an ypsiloid
apparatus, and with no evidence of degeneration in the pelvic
region (e. g., Ncctunts) are neither degenerate forms, nor perma-
nent larvae of any of the Salamandrida.

b. That the absence of the ypsiloid cartilage considered in con-
nection with the absence of lungs in the case of Typhlomolge is
in full accord with the conclusion [Emerson, 1905] that TypJilo-
tnolge is the permanent larva of some lungless salamander.

c. That the presence of a functional ypsiloid apparatus in
Cryptobranchus indicates that Cryptobranchus lies near the line of
descent of the Salamandrida.

SMITH COLLEGE, NORTHAMPTON, MASS.
February 1 , 1906.

POSTSCRIPT.

Since the above article was written, a paper on the Anatomy
of Cryptobranchus allegheniensis by Reese has appeared in the
American Naturalist, Vol. XL., No. 472. In this article the
following statement is made :

" Anteriorly the pubis is prolonged into a long, cartilaginous
epipiibis, which, instead of being forked as in the Japanese sala-
mander and some other Amphibia, is a straight rod, slightly
broadened and flattened at its distal end and somewhat enlarged,
both laterally and dorso-ventrally at its attached end. The union
of the pubis and epipubis is a close one, but allows considerable
freedom of motion."

The results of my own dissections (p. 264) are so completely
at variance with this description of Reese's that I can but feel
that he was mistaken in the form and character of the part in
question. I have, however, based my description upon three
specimens only, and it is possible that we have here to do with
a case of marked individual differences ; but that all of my speci-
mens should have the typical Y-form, while all of Reese's were
rod-shaped, does not seem probable. I. L. W.

296 INEZ L. WHIPPLE.

BIBLIOGRAPHY.
Barrows, Anne Ide.

'oo Respiration of Desmognathus. Anat. Anz. , Bd. 15, Nos. 18, 19.
Baur, G.

'91 The pelvis of the Testudinata ; with notes on the evolution of the pelvis in

general. Jour, of Morph., Vol. 4.
Bruner, H. L.

'96 Ein neuer Muskelapparat zum Schliessen und Offnen der Nasenlocken bei der

Salamandriden. Archiv. ftir Anat. u. Phisiol., Anat. Abteil.
'01 The smooth facial muscles of Anura and Salamandrina. Morph. Jahrb., Bd.

19, Hft. 3.
Carlsson, Albertina.

'03 Beitrage zur Anatomic der Marsupial region bei den Beuteltieren. Zool.

Jahrb., Bd. 18.
Camerano, L.

'94 Ricerche anatomo-fisiologiche intorno ai Salamandridi normalmente apneu-

moni. Libraio della R. accad. della Scienzi. Torino.

'96 Nuove ricerche intorno ai Salamandridi normalmente apneumoni, e intorno
alia respirazione negli Amfibi Urodeli, Libraio della R. accad. della Scienzi.
Torino.
Duges, Anton.

'55 Recherches sur 1'osteologie et la myologie des Batraciens a leurs differents

ages. 4to. Paris.
Emerson, Ellen T.

'05 General Anatomy of Typhlomolge rathbuni. Proc. Boston Soc. Nat. Hist.,

Feb., 1905.
Gadow, H.

'01 Amphibia and Reptiles, MacMillan. London. (In Cambridge Nat. Hist.

Series. )
Gage, S. H ., and S. P.

'86a Combined aerial and aquatic respiration. Science, Vol. 7, No. 16 .

'86b Pharyngeal respiratory movements of adult Amphibia under water. Science,

Vol. 7, No. 169.
'86C Aquatic respiration in soft-shelled turtles ; a contribution to the physiology of

respiration in Vertebrates. Amer. Nat., Vol. 20, March, 1886.
Gage, S H.

'91 Life History of the Vermilion-spotted Newt; Diemyctylus viridescens.
Raf. Amer. Nat., Vol. 25, Dec., 1891.

Hay, 0. P.

'89 Notes on the habits of some Amblystomas. Amer. Nat., Vol. 23.
Hoffmann, C. K.

'73~'?8 Amphibien ; in Bronn's Klassen und Ordungen des Tierreiches. Leipzig
und Heidelberg.

Hyrtl, J.

'65 Cryptobranchus Japonicus ; schediasma anatomicum. 410, Vindobonre.

Kingsbury, B. F.

'05 The rank of Necturus among tailed Batrachia. Biol. Bulletin, Vol. 8, No. 2.
Lonnberg, E.

'96 Notes on tailed Batrachians without lungs. Zool. Anz., Bd. 19, No. 494
'99 Salamanders with and without lungs. Zool. Anz., Bd. 22, No. 604.

THE VPSILOID APPARATUS OF URODELES. 297

*

Maurer, F.

'gi Der Aufbau nnd die Entwickelung der ventralen Rumpfmuskulatur der uro-
delen Amphibien und deren Beziehung zu den gleichen Muskeln der Selachier
und Teleostier. Morph. Jahrb. , Bd. 18.
'94 Die ventrale Rumpfmuskulatur der anuren Amphibien. Morph. Jahrb.,

Bd. 22.
'99 Die Rumpfmuskulatur der Wirbeltiere und die Phylogenese der Muskelfaser.

Anat. Ergeben., Bd. 9.
Oppel, A.

'99 Athmungs-Apparat. Anat. Ergeben., Bd. 9.
Ritter, Wm. E., and Miller, L. H.

'99 A contribution to the life-history of Audorax lugubris Hallow, a Californian

Salamander. Amer. Nat., Vol. 33, Sept., 1899.
Ritter, Wm. E.

'03 Further notes on the habits of Autodax lugubris. Amer. Nat., Vol. 37, Dec.,

1903.
Wiedersheim, R.

'75 Salamandrina perspicillata und Geotriton fuscus ; versuch einer vergl. Anab.
der Salamandrinen mit besond. Berucksichtegung der Skelet-verhaltnisse.
Small 410, Genua.

'88 Zur Urgeschichte des Beckens Berichte der naturforsch. Gesellsch. zur Frei-
burg, i. B., Bd. 4.
'89 Ueber die Entwicklung des Schulter und Beckengiirtels. Anat. Anz., Bd.

4, No. 14.
'go Weitere Mittheilungen uber die Entvvicklungsgeschichte des Schulter und

Beckengiirtels. Anat. Anz., Bd. 5, No. I.
'92 Das Gliedmassenskelet der Wirbelthiere. Jena.
'92 Die Phylogenie der Beutelknochen. Zeitschr. f. wissen. Zool., Bd. L., 3,

Suppl.
Wilder, B. G.

'82 The habits of Cryptobranchus. Amer. Nat., Vol. 16.
Wilder, H. H.

'94 Lungenlosen Salamandriden. Anat. Anz., Bd. 9, No. 7-

'94 Lungless Salamanders : second paper. Anat., Bd. 12, No. 7.

THE OSMOTIC RELATIONS BETWEEN FISHES AND
THEIR SURROUNDING MEDIUM (PRELIMI-
NARY NOTE).1

FRANCIS B. SUMNER.

The effects upon fishes of changes in the density and salinity
of the surrounding medium involve numerous problems of great
physiological importance. Why is an extreme change of density
so fatal in some cases and so harmless in others ? And is it the
change of density which is responsible for the harmful effects
after all ? May not salt water be toxic in a narrower sense to
fresh-water fishes and vice versa? In any case, what is the im-
mediate cause of death ? Are the limiting membranes of a fish
permeable to both water and salts, or are they only semi-perme-
able ? Or are they perhaps impermeable to both ? And are all
of the limiting membranes alike in this regard ? Likewise is
their condition the same for all species and under all circum-
stances ? These are closely related questions. They have re-
ceived many and quite contradictory answers. It is hoped that
the experiments here discussed have contributed something toward
their solution.

The first of these experiments were chiefly concerned in deter-
mining whether a given change in water density was harmful to
a given species of fish, records being kept of the rate of death.
In a second series, weight determinations were made with a view
to ascertaining whether such changes in the density of the sur-
rounding medium were accompanied by appreciable osmotic
effects upon the fishes. Third, it was sought to discover whether
the membranes were permeable to water only or to salts as well.
The passage of salts from the fishes into the surrounding water
was tested chemically, and likewise the salt content of the tissues
of fishes of several species under different conditions was deter-
mined. Finally a series of experiments was performed with a
view to discovering whether such osmotic changes were confined

1 A more complete account of these experiments is in course of publication by the
Bureau of Fisheries and will, before long, appear in the Bulletin of that bureau.

298

FISHES AND THEIR SURROUNDING MEDIUM. 299

to the gills or whether the general integument of the body were
likewise concerned.

Altogether the results of about 1 50 experiments have been
taken into consideration in arriving at the conclusions here pre-
sented. These experiments were carried on during the summers
of 1904 and 1905 at the biological laboratory of the Bureau of
Fisheries at Woods Hole, Mass., and during the spring of 1905
at the New York Aquarium. Acknowledgments are due to
the officials of the Bureau of Fisheries for facilitating the prog-
ress of this work ; and to the director of the New York Aqua-
rium, Mr. Chas. H. Townsend, who placed at my disposal a
room equipped for research, and provided me with abundant
material throughout. My thanks are likewise due to Prof. W. C.
Sabine, of the department of physics of Harvard University, for
valuable criticism, and to Mr. D. W. Davis, assistant at the Fish-
eries Laboratory, for help during the earlier portion of the work.

Full details of these experiments, including the methods em-
ployed, and the precautions taken, must be deferred to the more
extended paper which will soon appear. In the meantime, the
principal results may be summarized as follows :

i . Certain brackish and salt-water fishes were unable to survive
even a gradual transfer to pure fresh zvater, though enduring an
abrupt transfer to water of a very lotv degree of salinity. Tints
fresh vvater, as suc/i, proved fatal to these fis/ies, the degree of ab-
ruptness of t/ie change being of secondary importance.

These conclusions are drawn from experiments upon the three
local species of killifishes (Fundulus lieteroclitus, majalis and dia-
pJianus}, together with the allied species, Cyprinodon variegatus ;
likewise the white perch (Morons americana), cunner (Tantogo-
labnts adspersns), tautog (Tantoga onitis), sculpin (Myoxocephalus
octodecimspinosus}, and winter flounder (Pseudopleuronectes ameri-
camis^. The death of a varying (often a large) proportion of
specimens of F. diaphamts, when transferred from mildly brackish
(sp. gr. 1.002—1.006) to pure fresh water was certainly unex-
pected, since this species in nature is not confined to brackish
waters, but is indigenous to lakes and streams far from the coast.
F. heteroclitus, likewise, is known to occur at times in fresh water ;
but the writer has found (contra Garrey J) that in nearly every

1 BIOLOGICAL BULLETIN, Mar., 1905.

3OO FRANCIS B. SUMNER.

experiment the entire lot died throughout a period of from less
than a day to several weeks.

In addition to a simple reversal of the normal medium (salt-
water fishes in fresh water and vice versa*) experiments were con-
ducted upon acclimatization, which was found to retard, but not
to prevent, the fatal effects of fresh water ; likewise with ^Mater of
a very low degree of salinity, which gave some of the most strik-
ing results to be recorded ; with fresh and salt ivater in alterna-
tion, in which case the fatal effects of the former were diminished
or annulled ; and with distilled water, which soon proved fatal to
F. lietcroclitus (the only species used). Surface abrasions (exten-
sive removal of scales) hastened the death of F. heteroclitus in
fresh water, but only exceptionally led to the death . of fishes
returned to full-strength sea water (again, contra Garrey) \ and
wrought no harm to fishes placed in very dilute brackish water
(three to four per cent, sea water), despite the fact that the latter
was without doubt strongly hypotonic to the fish.

2. Considerable changes of weight ivere found to result, in many
cases, from changes in the salinity (Jicncc the osmotic pressure) of
the surrounding medium.

The weighing operations were conducted with the following
species : Fundulus heteroclitns, F. majalis, F. diaplianus, Myoxo-
ceplialus octodecimspinosus, Microgadus tomcod, Pseudoplenronectes
americanus, Stenotomus clirysops, Ameiurus nebulosus, Lcuciscus
erytliropJitlialmus, Morone americana and OncorJiyncJius tsc/unvyt-
scha. During these experiments, the fishes were kept unfed, and
it is needless to add that abundant control experiments were per-
formed in order to determine the normal rate of loss through
waste. The changes of weight following changes in the density
of the surrounding medium were frequently surprisingly great, at
times as much as five per cent, or more in a single day. In many
cases, moreover, they were not accompanied by any apparent
harmful effect upon the fishes. With a very few exceptions, the
changes were such as to indicate that they were the result of
osmotic action. Thus, as a rule, the fishes gained in weight only
in solutions known to be decidedly hypotonic to their body fluids,
while with few exceptions a significant decrease only occurred in

1 Op. df.

FISHES AND THEIR SURROUNDING MEDIUM. 301

those cases in which they were transferred from a hypotonic or
isotonic medium to one which was strongly hypertonic. Nega-
tive results were indeed encountered at times, but very few which
could be regarded as contradictory.

Extensive surface abrasions did not facilitate the influx or
efflux of water. The changes of weight in dead fishes were such
as to show that factors other than osmosis were concerned.
Dead fishes of most of the species used were found to gain in
weight in water of any degree of salinity up to the strength of
normal sea water.

3 . Considerable changes in the salt [chlorine] content of the body
were likewise found to result, in many cases, from changes in the
salinity of the water.

The problem here involved was attacked from both sides. In
the first place, the passage of salts (strictly speaking, of chlorides) *
into fresh water from fishes taken from salt or brackish water was
tested chemically. In the second place, the salt content of the
tissues of various fishes which had lived in water of various de-
grees of salinity was likewise determined.2 It was found that the
results from these two methods presented some striking points of
agreement, though the latter proved, on the whole, to be much
more satisfactory. These changes in the chlorine content of the
body were frequently astonishing in their magnitude. F. Jietero-
clitus by both methods was found to part with about twenty-five
per cent, of its chlorine in the course of a single day. The loss
of chlorides from the body was, however, found to occur at a
steadily diminishing rate.

The following table indicates the percentages of chlorine found
in specimens of F. diafhanns from brackish water (thejr habitat
locally) and in those kept for varying periods in fresh and in salt
water. The series is certainly suggestive.

in days (8 fishes employed) 0.085
3 days (4 fishes employed) o. 108
I day (4 fishes employed) O.II2

1 Mohr's silver nitrate titration method was here employed. The gain or loss in the
proportion of chlorine was held to be indicative of the behavior of the various saline
ingredients of sea water.

2 It is needless to state that all fishes used in these two series of tests were previ-
ously thoroughly rinsed in fresh water.

3O2 FRANCIS B. SUMNER.

Brackish water (sp. gr., 1.002) (5 lots of 4 fishes each), 0.134 [or 0.142]'
5 days (2 fishes employed) °-I43

Salt water (1.023) i

| 10 days (3 fishes employed) 0.151

It will be seen that the last of these figures is about 78 per
cent, greater than the first ! And yet the fishes were all alive
and apparently well at the time they were killed for analysis. It
will be likewise seen that whichever figure be regarded as the
more correct one for the brackish-water fishes, the latter agree
much more closely with the salt-water than with the fresh-water
individuals (the comparison being of course between the extreme
members of the series). It must be added in strict fairness, how-
ever, that in two different tests fishes kept for only one day in sea-
water gave a much higher percentage of chlorine than those kept
for five or ten days. For this apparent anomaly I believe that a
satisfactory explanation can be given, but this has been deferred
to my longer work.

A series of figures somewhat similar to the above was obtained
from experiments with the white perch.

Experiments with both F. lictcroclitus and F. majalis agreed in
showing a great difference between the effects upon the chlorine
content of the body of pure fresh water and water having a certain
small percentage of salt. This difference is extremely significant
in view of the difference, already mentioned, in their effects upon
the life of the fishes. Moreover, it was further found that the
average percentage of chlorine contained in the salt-water fishes
analyzed was of the same order of magnitude as that of water
containing just enough salt to maintain the fishes in health. That
such water could not have been even approximately isotonic with
the body fluids of these fishes seems evident from the cryoscopic
determinations of other investigators.

4. Careful control experiments excluded the possibility that the
water or sails entered or passed from the body through the alimen-
tary canal, leaving as t/ie only probable alternative an osmotic
exchange through one or more of the external membranes.

The alimentary canal, and indeed the whole abdominal viscera,
together with the washings from the body cavity, were found in

1 This second figure is the mean which results when one very questionable deter-
mination is included. It is inserted for the sake of strict fairness. The other aver-
ages are in each case derived from all of the fishes tested.

FISHES AND THEIR SURROUNDING MEDIUM. 303

some cases to yield less chlorine than passed from the body of a
living fish in the course of a few hours. Likewise those fishes
whose bodies were analyzed gave approximately the same per-
centages of chlorine whether or not the alimentary canal was
included in the analysis. It must be remembered also that the
fishes employed in these experiments had, in all cases, been kept
unfed for some days previously.

5. In certain fishes, at least, it was found that the membranes
cJiicfly concerned in such exchanges were those of the gills.

In the case of certain specimens, salt water was passed through
the gills by means of a rubber tube placed in the mouth, the body
being bathed with fresh water ; while in others the arrangement
was reversed, the gills receiving fresh water and the general in-
tegument salt. In six experiments with the carp it was found
that a considerable loss of weight occurred in all of those cases
in which the former conditions obtained, while the weight re-
mained practically stationary in those cases in which the condi-
tions were reversed.

A complete historical review of previous researches in this field
of physiology would be beyond the scope of the present paper.
The investigations of Fredericq,1 Bottazzi,2 Rodier,3 Garrey4 and
Greene 5 agree in showing that the blood of marine teleosts is far
from being isotonic with the surrounding sea-water, but that it has
an osmotic pressure which is roughly about one half that of the
latter. But it has likewise been shown (Fredericq, Greene, op.
cit.} that the osmotic pressure of the blood of salt-water teleosts
is considerably higher than that of fresh-water ones, though this
fact has been almost lost sight of in the zeal to prove that the
internal medium is not isotonic with the external, and that its
osmotic pressure is relatively constant. Indeed it does not seem
to have been generally appreciated that there is a certain corre-
lation between the inner and outer fluids, both as regards osmotic
pressure and salt content ; and certain authors have been free to
state that the membranes of teleost fishes form an effective
barrier against osmotic changes. Fredericq makes this assertion

1 Archives de Biologic, 1904. 3 Cited by Fredericq, 1904.

2 Archives italiennes de Biologic, 1897. 4 BIOLOGICAL BULLETIN, 1905.
5 Bulletin U. S. Bureau of Fisheries, 1905.

304 FRANCIS B. SUMNER.

broadly, while Garrey says of Fiinduhis heteroclitus : "The in-
tegument and gills are therefore impermeable." Garrey is
cautious enough, however, not to postulate an absolute imperme-
ability, either for Fundulus or for teleosts in general. Greene,
though he finds that the osmotic pressure of the blood of the
Pacific salmon undergoes a decrease of about 17 J^ per cent,
when the fish ascends a river to spawn, is nevertheless doubtful
whether osmotic exchanges with the surrounding water are re-
sponsible for this decrease.

In view of my own experiments, however, we are certainly
not justified in concluding from the absence of osmotic equi-
librium between the fish and its environment that no osmotic
interchanges normally occur. On the contrary, abundant experi-
ments seem to prove that both water and salts may, under certain
conditions, be transmitted in either direction without any harm
resulting to the fish. These conditions seem impossible to state
in advance for a given case. In general we may say that :

1 . Measurable changes in weight result only from considerable
changes in the surrounding water, but —

2. Not all such changes of density suffice to produce changes of
weight, even w/ien the fish is transferred to a medium uhich is
knoivn to be strongly Jiypertonic or hypotonic to its own body fluids.

3. Changes in the salinity of the water may or may not result in
changes in the salt content of the body.

4. Changes in tJie bodily salt content may or may not be accom-
panied by changes in ivcig/it.

5. Neither the changes in zveight nor in salt content are at all
proportional to the changes in the density of the external medium.

It ivould appear that there is normally a tendency on the part
of the fish to resist osmotic changes and to maintain the fluids of
the body at a definite degree of concentration. Under various con-
ditions, however, this resistance is overcome and a certain degree of
permeability is established. This is generally a differential per-
meability, resulting in osmosis and consequent changes of weight.
In such cases, /lozcevcr, the membranes are not strictly semiperme-
able, but transmit salts in some measure. Indeed it ivould seem that
at times the permeability is indiscriminate, in which case the salts
may diffuse freely, but no changes in weight occur. These various

FISHES AND THEIR SURROUNDING MEDIUM. 305

changes continue until a new level of stability is established, after
ivhich the normal resisting power of the fish reasserts itself and no
further alteration occurs as long as the medium is constant. Complete
osmotic equilibrium between the fish and the rvater is probably never
attained except in waters having roughly a medium degree of sa-
linity. The osmotic pressure of the " internal medium" fluctuates
within a much narrower range than that of tJie " external medium.'"

The foregoing conclusions are intended to apply only to nor-
mal fishes. It seems certain that the enfeeblement of the fish
may result in an increased permeability of the membranes,
which, in turn, would doubtless result in a further enfeeblement
of the fish. The death of those fishes which cannot withstand
transfer to a medium very different from that to which they are
accustomed is thus probably in part a cause and in part an effect
of these changes. Death is accompanied (perhaps in some cases
caused) by a giving way in the power to resist an abnormal de-
gree of osmotic exchange. The body becomes water-soaked (if
in fresh water) or dehydrated (if in salt). The difference be-
tween the more hardy and the more delicate species in this re-
gard seems to lie partly in the resisting power of the limiting
membranes (chiefly those of the gills) ; partly, also, in internal
differences, such as composition of blood, etc., which determine
whether or not a given influx or efflux of water or salts shall
prove fatal.

The actual cause of death following a change in the salinity of
the water seems to differ in different cases. With those fishes
which succumb rapidly with but a slight change of weight (e. g.t
scup), it is unlikely that any appreciable alteration occurs in the
tissues at large. Such changes are probably confined to the
blood, perhaps, as Bert ! held, to that in the gill capillaries, in
which case death may result from asphyxiation (Bert,2 Mosso 3).
In those cases, on the contrary, where the fatal effects are not
manifested for some days, it seems likely that the manner of
death is different. In the case of P. heteroclitns, it was found in
most instances that the eudosmotic flow of water had ceased,
and that a secondary decrease in weight had ensued, within one

1 Comptes Rendus de V Academie des Sciences, 1871, 1883. 2 Of. cit.

3 Biologisches Centralblatt, 1890.

306 FRANCIS B. SUMNER.

or two days after transfer to fresh water. On the other hand, it
will be remembered that fishes of this species commonly did not
die for a considerable number of days, while many survived for
a week and some even for several weeks. Again, it will be
recalled that the fatal effects of fresh water upon this and some
other species were nullified by the admixture of a very small
percentage of salt water. Analyses showed that in this latter
case there was little or no decrease in the salt content of the
body. A rough approximation was pointed out between the per-
centage of salts in this faintly saline water and that in the fishes
themselves. All of these facts point to the conclusion that one factor
in t/ie deatli of salt-water fishes in fresh water is the extraction
from their tissues of an amount of salts sufficient to reduce the per-
centage below a certain necessary minimum.

If the question be asked : Why are not fresh-water fishes thus
affected in their own medium ? it is replied that their mem-
branes have been adapted to resisting such an extraction of salts.
It is perhaps also true that the irreducible minimum of salts in
these species is lower than in the case of salt-water ones. In any
case, the percentage actually present is, on the average, less
(Atwater,1 Katz,2 Quinton3 and several others).

Whether or not salt water ever has a toxic effect, in the nar-
rower sense, upon fresh-water fishes cannot be stated definitely.
Bert denied that such was the case, but he is not entirely con-
sistent in this position. In view of the fatal effects upon salt-
water fishes of some of the individual components of sea salt,
when taken separately (Loeb,4 Siedlecki 5), it seems quite possi-
ble that sea-water may act as a poison to fresh-water organisms,
independently of any osmotic effects. Indeed both of the last-
named writers have shown that it is the chemical nature of the
solutions used rather than their osmotic pressures which deter-
mines, in many cases, whether or not they shall prove fatal.

COLLEGE OF THE CITY OF NEW YORK,
March 2J, 1906.

1 Report U. S. Com. Fish and Fisheries for 1888 (1891).

2 Archiv filr die gesammte Physiologie, 1896.

3" L'eau de mer, milieu organique," Paris, 1904.

4 American Journal of Physiology, 1900.

5 Comptes Rendus de r Academic des Sciences, 1903,

^ o f* r~" i c*i r~
3 J J J i U 0