7 aaa

Sh re memati

JOURNAL

OF

MORPHOLOGY

FOUNDED By C. O. WHITMAN

EDITED BY

JS; KENG SE RY:

Tufts College, Mass.

WITH THE COLLABORATION OF

Gary N. CALKINS WILLIAM PATTEN
Columbia University Dartmouth College

W. M. WHEELER Epwin G. ConkKLIN
Bussey Institution, Harvard University Princeton University

VOLUME 23
1912

THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

PHILADELPHIA

0 ee Pe eg ee ae
ne? seo ae a 3 aes
x “! « :

a ; 3

¢

| =o" near oon ly ame
ae Pao 3 : PS veo H arya.
| } biomidod —
i ws moMl vf wou

piqnin add «=
! : scons
| / ) 7 ol == : a4 ana nol LA

Tier HEL e

tg hae nuit Ak

eee
x H asaanv devia Le.
Bocas ikea

CONTENTS

1912

No. 1. MARCH

Jacques Lors. Heredity in heterogeneous hybrids. Nineteen figures......
Davin H. Tennent. The behavior of the chromosomes in cross-fertilized
Hchimoidiec gst dawientiylCUnessres qe see eee ie ets cia. eae
Roy L. Mooprn. The skull structure of Diplocaulus magnicornis Cope and
the amphibian order Diplocaulia. Sevep figures......................
MricHaret F. Guyer. Modifications in the testes of hybrids from the guinea and
the common fowl. Twenty-three figures.....+............:..........5-
Bertram G. SmitH. The embryology of Cryptobranchus allegheniensis, in-
cluding comparisons with some other vertebrates. I. Introduction: The
history of the egg before cleavage. Fifty-six figures...................
Epwin G. Conkuin. Body size and cell size. Twelve figures...............

Nios 25 SUING

Henry Lesuie Osporn. On the structure of Clinostomum marginatum, a
trematode parasite of the frog, bass, and heron. Seventeen figures......
B. F. Kinaspury AND PAULINE Hirsu. The degenerations in the secondary
spermatogonia of Desmognathus fusca. Twenty-one figures...........
R. T. Young. The epithelium of Turbellaria. Six figures..................
GrorceE W. BartTetmMeEz. The bilaterality of the pigeon’segg. A study in egg
organization from the first growth period of the oocyte to the beginning
OScleavareaeehOryoseVveM I 2UTES:% a4. 2 005:% 25+ 2 ee Hee sc heen
FERNANDUS Payne. J. A further study of the chromosomes of the Redu-
viidae. II. The nucleolus in the young oocytes and origin of the ova in
Gelastocoris. Ten figures...... NW I er Petra Bere SES oak

61
159

iv CONTENTS

No. 3. SEPTEMBER

Jutta E. Moopy. Observations on the life-history of two rare ciliates,

Spathidium spathula and Actinobolus radians. Sixty-six figures........ 349
C.H. Danrortu. The heart and arteries of Polyodon. Nineteen figures.... 409

Bertram G. Smira. The embryology of Cryptobranchus allegheniensis,
including comparisons with some other vertebrates. II. General embry-
onie and larval development, with special reference to external features.
Two, hundred ‘and twenty-three figures... .. ... 0.2.05 tie se one 455

No. 4. DECEMBER

Ropert Marurson. The structure and metamorphosis of the fore-gut of
Corydalis'cornutys Li Twentyenime figures: -..25.3.: fe se ee ee 581
W. M. SMatuwoop ANpD EuizaBetu G. CLARK. Chromodoris zebra Heilprin,

a (distinct, Species. . Six: figuresmiys<:. «95.2 nt (oe eee ee 25
SaWe Winniston.. Primitive reptiles. “Onelieumess..-%.-90-ce ene 637
Gary N. Cauxins. The paedogamous conjugation of Blepharisma undulans

Sty Uwentyemve figures: « ...Bypacer case lcictseicieet eegitee ite eset tol een 667
B. W. Konxen. The development of the skull cf Fmys lutaria. Thirty-one

ATOEUAT CSS £7, Ma Tig) gos seul sun voces Pais vac oc CR IR ge NNO Rg cite RMR 693

ide
&
*
| ey
an
a
>
rs "
3,

~ San g

HEREDITY IN HETEROGENEOUS HYBRIDS'

JACQUES LOEB
The Rockefeller Institute, New York

NINETEEN FIGURES

1. The study of heredity in embryos offers in one respect a
wider field than that in adults inasmuch as heterogeneous hybrids
rarely reach the adult stage. Eight years ago I found a method
by which the eggs of the sea-urchin can be fertilized by the
sperm of starfish, ophiurians and holothurians. The larvae
are purely maternal, namely plutei. The results were confirmed
by Godlewski for the fertilization of the egg of the sea-urchin
by the sperm of the crinoid. It is well known that if we cross
two homogeneous forms, e.g., two forms of sea-urchins, the pater-
nal influence can be clearly seen in the pluteus stage. Since I
have never published the figures of my experiments on hetero-
geneous hybridization, I may supplement my former statements
with a few drawings. Figs. 1 to 6 are camera drawings of plutei
of Strongylocentrotus purpuratus produced by artificial parthe-
nogenesis. The plutei are, of course, in every detail identical
with the plutei obtained if these eggs be fertilized with sperm of
their own species. Figs. 7 to 9 are drawings of five days old
plutei of Strongylocentrotus purpuratus ¢ and Strongylocentrotus
franciscanus ~. They differ from the pure breeds of S. purpuratus
in several characters of the skeleton which exist in the pluteus
of franciscanus but are absent from purpuratus, namely the
greater roughness of the skeleton, the presence of cross bars and
the greater length of the arms.? In figs. 10 to 13 are shown the
five days old plutei of the egg of S. purpuratus fertilized with
the sperm of the starfish (Asterias). It is obvious that the latter

1Prepared for the Whitman Memorial Volume but received too late to be
included.
2 Loeb, King and Moore, Arch. f. Entwicklungsmechanik, vol. 29, 1910.

1

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1 &
MARCH 20, 1912

2 JACQUES LOEB

10

Figs. 1 to 6 Five days old plutei of Strongylocentrotus purpuratus produced
by artificial parthenogenesis.

Figs. 7 to 9. Five days old plutei of Strongylocentrotus purpuratus? and
Strongylocentrotus franciscanus o’.

Fig. 10 to 13 Five days old plutei from Strongylocentrotus puipuratus 2 and
Asterias <’.

HEREDITY IN HETEROGENEOUS HYBRIDS 3

plutei are purely maternal. It should, however, be borne in mind,
that the objection might be raised that the presence of the skel-
eton in the sea-urchin pluteus might be dominant over its absence
in the starfish larva. I have thus far vainly tried to produce a
starfish larva with a pluteus skeleton by the hybridization of the
two species.

We are therefore compelled to state that the hybrids between
the sea-urchin egg and the starfish sperm represent more closely
the purely maternal form than do the hybrids between two sea-
urchins, which always show paternal characters.

2. It is well. known that Herbst and Tennent have made
experiments in which the paternal influence in the hybrid embryo
was diminished. ‘Tennent states that in the cross between Hip-
ponoe and Toxopneustes, Hipponoe characters become dominant
in sea water of a high OH concentration and Toxopneustes char-
acters in sea water of a low OH concentration.? The amount of
acid or alkali Tennent needed to accomplish his result was very
small; namely about 2 ec. * acetic or hydrochloric acid to 500
ec. of sea water. A. R. Moore, W. R. King and myself made a
number of experiments in which the hybrids between S. purpura-
tus and 8. franciscanus were raised in sea water to which varying
quantities of HCl or acetic acid or NaHO were added (from 0
to 0.4 cc. 7 acid or NaHO to 50 ce. of sea water). We were
able to retard or accelerate the rate of development, but the char-
acter of the hybrid remained absolutely unaltered.

I wish to call attention to the necessity of sterilizing the pipettes
by boiling them after each experiment, instead of sterilizing them
by rinsing in distilled or fresh water as is often done.

3. Moenkhaus measured the rate of segmentation in hybrid fish
eggs and found that the rate for the first five cleavages is deter-
mined by the egg.t. The egg of Ctenolabrus segments about forty
minutes after impregnation with sperm of its own kind, while
the egg of Batrachus tau, if fertilized with the sperm of the same
species, segments after about eight hours. If the egg of Batra-
chus be fertilized with the sperm of Ctenolabrus it also does not

3 Tennent, Publication 132, Carnegie Institution, 1910.
* Moenkhaus, Am. Jour. of Anat., vol. 3, p. 29, 1904.

4 JACQUES LOEB

segment until after eight hours.® I have repeated these experi-
ments in a number of fish hybrids and confirmed Moenkhaus’
results. The same facts are observed in the rate of development
of hybrid embryos of echinoderms. What is the meaning of this
fact? I believe that it finds its explanation through. artificial
parthenogenesis. These latter experiments have shown that the
spermatozoon does not cause the development by carrying an
enzyme or katalyzer into the egg, which the latter needs in order
to develop, but causes the development by altering the surface
layer of the egg. If the development of the egg were caused by
an enzyme carried into the egg by a spermatozoon, the rate of
cleavage of slowly developing eggs should be accelerated by a
spermatozoon of a species developing at a faster rate. The egg
however behaves exactly as we should expect from the fact that the
spermatozoon removes only certain obstacles for the development
of the egg but does not cause its development by carrying an acti-
vating enzyme. In order to cause the parthenogenetic develop-
ment of the sea-urchin egg two different agencies are required and
I have been able to show that the spermatozoon also causes the
development of the sea-urchin egg by two agencies which act
analogously to the agencies employed in my method of artificial
parthenogenesis.’ FF. Lillie has found the same in Nereis.’?. It was
therefore to be expected that the rate of development of embryos
was determined by the egg. This view meets with a difficulty
in the fact that, with a few exceptions, the later development of
hybrids is as arule retarded. But this difficulty is only an appar-
ent one since the retardation is due to an entirely secondary con-
dition: namely that most hybrids are sickly and not able to hatch
or reach an adult state.

This is most strikingly the case in the heterogeneous hybrid,
e.g., in the cross between the sea-urchin and the starfish. As I
pointed out long ago the larvae die mostly in the gastrula stage,

5 The acceleration of segmentation which Newman observed in the egg of Fun-
dulus majalis fertilized by the sperm of Fundulus heteroclitus is too small to influ-
ence our conclusions.

® Loeb, Die chemische Entwicklungserregung des thierisches EHies. Berlin,

1909, and Das Wesen der formativen Reizung, Berlin, 1909.
7F. R. Lillie, Jour. of Morph., vol. 22, 1911.

HEREDITY IN HETEROGENEOUS HYBRIDS 5

and possibly one egg in a million reaches the pluteus stage. The
development of the pluteus is in such cases always retarded.

We find such a retardation not only in the case of heterogeneous
hybridization but occasionally also in the case of crosses between
closely related forms. While the hybrid purpuratus¢, francis-
canus ¢ is vigorous, the hybrid franciscanus ¢, purpuratus 2, is
sickly and reaches the pluteus stage only rarely and slowly.

The rate of development of the embryo is a function of the
velocity of certain chemical processes which are linked together
like the wheels in a mechanical machine. If one of such processes
be retarded the rate of development of the whole embryo is likely
to be retarded; and if the linkage of the various chemical proc-
esses become disturbed the embryo is likely to be sickly. The fur-
ther apart the species are from which the two sex cells originate
the greater the likelihood that the rate of development is retarded
and that the hybrid embryo is sickly. Why is the development
not retarded from the beginning? Possibly for the reason that
it requires some time before the spermatozoon can cause the
formation of a sufficient amount of harmful substances to cause a
retardation of the development of the egg.

4. Moenkhaus found that the eggs of bony fishes can be easily
impregnated with foreign sperm but that they do not develop
very far. Thus he states, that the hybrids between Menidia
and Fundulus heteroclitus ‘“‘never go beyond the closure of the
blastopore.” I have been able to raise the hybrid between
Fundulus heteroclitus ¢ and Menidia, Ctenolabrus and Stenotomus
+ in large numbers beyond this stage. These hybrids lived a
month or longer, formed hearts, blood vessels, eyes, and fins but
never hatched. With a few exceptions no circulation was ever
established although the heart beat for weeks.

Figs. 14 to 16 show three different hybrids of Fundulus hetero-
clituse and Menidia., from a lot fertilized the 12th of June.
The camera drawings were made from the living material the 2d,
3d and 12th of July. At that time the pure breed of Fundulus
heteroclitus fertilized on the same date were already hatching.
The hybrid embryos had formed the pigment characteristic for
the pure breed of Fundulus heteroclitus. But the anomalies of

6 JACQUES LOEB

the embryos are very obvious. The embryos are rather small,
owing to the slowness with which they digest the yolk. Their
eyes are abnormal and approach the cyclopean condition. In
many specimens only irregular masses of pigment indicate where
the eyes should be. The head is comparatively small and not
bent as is characteristic for the pure breed. The heart is devel-
oped but corresponds to an early stage in the development. It
beats regularly and at an almost normal rate. The main blood
vessels exist and haemoglobin is formed but the creeping of the
pigment cells upon the blood vessels does not take place.

Years ago I found that the marking of the yolk sac of Fundulus
and of the embryo is caused by the creeping of the chromatophores
upon the blood vessels. I showed that this phenomenon is
due to a tropism which depends upon the circulation. When the
circulation was suppressed pigment was developed but the chro-
matophores did not creep upon the blood vessels. At that time
I had succeeded in suppressing the circulation for a few days.’
In the new experiments the hybrid embryos lived for a month
or more with pigment but without a circulation. They demon-
strate the correctness of my former statement, inasmuch as the
creeping of the chromatophores upon the blood vessels did not
take place. They also confirm the statement, that the forma-
tion of pigment cells is independent of the circulation. Newman
seems to hold the opposite view, but he evidently did not test
his assertion experimentally.

These hybrids were also smaller than the pure breeds of the
same age, owing to the fact that the yolk is less rapidly digested
in the hybrids than in the pure breeds. This is a very important
link in our conclusions on heredity. The development of heredi-
tary characters is the result of the nature and the velocity of chem-
ical reactions between the mass of yolk on the one hand and the
substances in the nucleus, especially the chromosomes, on the
other. If two closely related forms be crossed, the chemical
reactions. are not materially different in quality and velocity from
those of the pure breeds. But when distant forms are crossed it
is to be expected that greater differences in the nature and the

8 Loeb, Jour. of Morph., vol. 8, 1893; Woods Hole Lectures.

HEREDITY IN HETEROGENEOUS HYBRIDS fa
rate of chemical reactions will be found and the outcome will be
pathological embryos and very likely a suppression of the paternal
influence. ‘The disturbance is the same in practically all the heter-
ogeneous hybrids. I have also produced the crosses between
Ctenolabrus ~ and Fundulus heteroclitus ¢ and between Stenoto-
mus¢ and Fundulus heteroclitus¢. In all cases the result was
about similar to the one described here. In all cases there was a
consumption of yolk, development of an embryo, of pigment, of a
heart beat, of eyes, lenses, ears, fins; but, with rare exceptions
about which we are to speak later, there was no circulation. The
number of relatively good embryos was very large in the cross be-
tween Fundulus heteroclitus ¢ and Menidia< (where about 90 per
cent formed embryos that lived for about a month); it was much
smaller in the cross between Fundulus heteroclitus ¢ with Cteno-
labrus ~. One word should be said in regard to the development
of. the head in these embryos. In later stages it is often abnor-
mally small in comparison with the body. The reason for this is
that, although at first the head of these heterogeneous hybrids
develops normally, sooner or later its development stops and
often phenomena of degeneration set in, especially in the eyes.
The body of such larvae however continues to grow. -

5. From what was said before, I reached the conclusion, that
these hybrid larvae between Fundulus ¢ and Menidia ~ were
in reality pure breeds, namely Fundulus heteroclitus larvae whose
development was retarded through some interference with the’
normal chemical reactions in the egg; and that the abnormalities
described were in no way hybrid characters. It occurred to me
that it might be possible to produce similar larvae from pure breeds
of heteroclitus eggs, if the latter were compelled to develop under
an abnormal chemical condition. For this purpose the following
experiment was made. Eggs of Fundulus heteroclitus were put
immediately after fertilization into closed Erlenmayer flasks, each
‘of which contained 50 cc. of sea water to which various amounts
of a 0.01 per cent solution of NaCN were added, from 0 to 10 ce.
The eggs remained in these solutions about a month. In the
mixture of 2 cc. 0.01 per cent NaCN and 50 cc. of sea water, em-
bryos were found which in every way resembled the hybrids

8 JACQUES LOEB

between Menidia ~ and Fundulus heteroclitus 9. (See figs. 17
and 18.) Their eyes were poorly developed, they had a tendency
to form cyclopean eyes, the yolk was incompletely digested and
the embryo too small. The heart was formed and beat, but no
circulation was established. Pigment was formed (in the draw-
ing most of the black pigment cells on the yolk sac were not indi-
cated, since the drawing was intended for another purpose).
The head is not bent against the body, and so on. In all respects
these sickly embryos of pure breed resemble the hybrids between
Fundulus heteroclitus ¢ and Menidia..

Since NaCN acts through a retardation in the rate of oxidation,
the idea might be expressed, that in heterogeneous hybrids oxida-
tions are interfered with. It is not safe to accept such an idea
until it has been tested experimentally.

6. The idea that the heterogeneous hybrids in fish are purely
or at least essentially maternal finds support in the fact that a
small number of these hybrids develop more normally than those
thus far mentioned. In a small number of such hybrids circula-
tion is established, though as a rule rather late and often for a few
days only. But these embryos develop rather normally and are,
as far as any present observations go, purely maternal. I have
had many such hybrids but I will give only one camera drawing
(fig. 19), representing a twenty-five day old hybrid between a male
scup and a female Fundulus heteroclitus. The yolk is in this case
pretty well digested.

7. While it is the rule that in the case of heterogeneous hybrid-
ization heredity is purely maternal, it is possibly not without
exception. I have however thus far found only one paternal
character which is possibly transmitted to a heterogeneous hybrid.
The yolk sac of the egg of Fundulus heteroclitus forms branched
red chromatophores which are not found on the yolk sac of Meni-
dia. In two eggs of Menidia fertilized by sperm of Fundulus
heteroclitus a few red chromatophores were observed. It is diffi-’
cult to get this cross and I give this observation with some
hesitancy.

8. Kupelwieser and Baltzer account for the fact that hetero-
geneous hybrids are purely maternal by the assumption that the

HEREDITY IN HETEROGENEOUS HYBRIDS 8)

chromosomes of the sperm are thrown out of the egg or dis-
integrate. This is not ‘1 harmony with the observations of
Moenkhaus for the hybrids between Menidia 7 and Fundulus
heteroclitus¢, and with those of Godlewski for the hybrids
between sea-urchin and crinoid. I am not in a position to decide
the differences in the observations of these authors. The obser-
vations mentioned in the preceding paragraph are more in
harmony with the observations of Moenkhaus and Godlewski.

CONCLUSIONS

The spermatozoon has two distinct effects upon the egg: namely,
it causes its development and it transmits certain parental hered-
itary characters to the offspring. The experiments in heter-
ogeneous hybridization confirm the idea supported by the exper-
iments on artificial parthenogenesis, that the formation of the
embryo is purely a matter of the egg and that the main function
of the spermatozoon is the causation of the development of the
egg. If we may express this statement in the form of a paradox
we may say that fertilization is primarily and essentially artifi-
cial parthenogenesis. The transmission of hereditary characters
through the sperm is in many cases merely an accessory function.
It becomes of vital importance only in those forms where the male
is heterozygous for sex and where the species can only be propa-
cated through sexual reproduction.

If the sperm nucleus be chemically almost identical with the
egg nucleus it is possible for it to force one or a few characters
upon the developing embryo. If the difference between sperm and
ege nucleus exceed a certain limit—which structural chemistry
may one day be able to define—the hereditary influence of the
spermatozoon is as a rule completely or almost completely oblit-
erated; and the result is a purely maternal larva, rendered more_
or less sickly through the presence or formation of foreign or
faulty substances.

The camera drawings of the sea-urchin larvae were made by
Mr. W. O. R. King, those of the fish embryos by Mr. Bagg. To
both gentlemen I wish to express my thanks.

10 JACQUES LOEB

Fig. 14 Hybrid between Fundulus heteroclitus 2 and Menidia o’, three weeks
old.

HEREDITY IN HETEROGENEOUS HYBRIDS et

Fig.15 Hybrid between Fundulus heteroclitus 2? and Menidia o’, twenty days
old.

12 JACQUES LOEB

Fig. 16 Hybrid between Fundulus heteroclitus 2? and Menidia <’, thirty days
old.

HEREDITY IN HETEROGENEOUS HYBRIDS 13

Fig. 17 Embryo of Fundulus heteroclitus thirty-one days old, raised in 50
cc. sea water + 2 cc. 0.01 per cent NaCN.

14 JACQUES LOEB

Heart
Beating

Fig. 18 Embryo of Fundulus heteroclitus thirty-one days old, raised in 50 ce.
sea water + 2 cc. 0.01 per cent NaCN.

HEREDITY IN HETEROGENEOUS HYBRIDS ' 15

Heart
Beating

Fig. 19 Hybrid between Fundulus heteroclitus 9 and scup <, twenty-five
days old.

THE BEHAVIOR OF THE CHROMOSOMES IN CROSS
FERTILIZED ECHINOID EGGS!

DAVID H. TENNENT
Bryn Mawr College

TWENTY FIGURES IN TWO PLATES

It is my purpose to present in this paper some of the facts
gained by a cytological study of Echinoid crosses and to consider
critically the results of Baltzer and Herbst in similar studies.

In my first work on the chromosomes of cross-fertilized Echi-
noid eggs (08) I found that beyond the mere determination
of species differences in chromosomes, little could be done until
a careful study of the chromosomes of the parent forms, in all
phases of mitosis, had been made.

Since that time important investigations on the straight-
fertilized eggs of some of the parent species have been completed
by two workers in my laboratory, Dr. Heffner (10) and Miss
Pinney (’11) so that we are now prepared to identify somewhat
adequately the chromosomes in Toxopneustes < Huipponoé,
Toxopneustes < Moira, Arbacia < Moira and Arbacia X
Toxopneustes crosses. In addition we now have the important
studies of Baltzer (’09, 710) and Herbst (’09) on European forms,
for comparison. .

During the past year I have investigated reciprocal Toxo-
pneustes < Hipponoé crosses made in normal sea water and the
same crosses made in sea water whose alkalinity had been reduced
as described in my papers of 1910-1911.

Heffner (10) showed for Toxopneustes long rods and two Vs
or three Vs. Pinney (’11) showed for Hipponoé four Vs in all
eggs and in addition a long armed V, or hook shaped chromosome,
in half of the eggs. My own observations on the Toxopneustes

1 Prepared for The Whitman Memorial Volume, but received too late to be
included.

JOURNAL OF MORPHOLOGY, VOL. 23, No. 1

17

18 DAVID H. TENNENT

@ & Hipponoé ¢ cross (’11) showed that a hook-shaped chromo-

some was present in 50 per cent of the eggs and that it must
have had its origin in the paternal, i.e., Hipponoé, nucleus. I
shall proceed immediately to a consideration of the crosses.

TOXOPNEUSTES ° x HIPPONOE &

As a general comparison between the chromosomes of straight-
fertilized Toxopneustes eggs and those of straight-fertilized
Hipponoé eggs, it may be stated that in Toxopneustes the chro-
mosomes are more slender and less closely massed on the spindle
than in Hipponoé. The massing of Hipponoé chromosomes is
due, in part, to their larger size and the fact that the Hipponoé
egg and amphiaster are smaller than those of Toxopneustes. But
in spite of the fact that the Hipponoé chromosomes are larger,
when studied in straight-fertilized eggs, I have found it impossible
to distinguish in general between the chromosomes of the two
species in cross-fertilized eggs; that is, I cannot find one group of
slender chromosomes and another group of thicker chromosomes,
which I can identify as of Toxopneustes and Hipponoé respec-
tively. In the cross-fertilized eggs all seem to be of about
equal thickness.

The chromosomes of the straight-fertilized eggs are in general
rod-like in form, some long, some short; in particular, a few have
individual peculiarity of form. These latter in Toxopneustes
(Heffner, ’10) are either two Vs or three Vs and two long rods; in
Hipponoé (Pinney, ’11) four Vs, or four Vs and a hook-shaped
chromosome.

When we examine the mitotic figures of Toxopneustes 9? X
Hipponoé ¢ material (figs. 1-11) we may identify these elements
readily. In fig. 1, four Vs, a long rod, a hook and, in addition,
two other bent elements (fig. 1, C) which I have never found in
other division figures. Fig. 2, three Vs anda hook. Fig. 3, three
Vs and a hook. Fig. 4 three Vs and a hook. Fig. 5 three Vs a
long rod and no hook. Fig. 6, two Vs in early anaphase. Fig. 7,
two Vsandno hook. Fig. 8, three Vs and no hook. Fig. 9, three
Vs and no hook. Fig. 10, two Vs. Fig. 11, three Vs and a hook.

CHROMOSOMES IN ECHINOID EGGS 19

In a careful study of fifty-two amphiasters in which all of the
elements might be seen with clearness sufficient to enable me
to determine the presence or absence of the hook-shaped chromo-
some, it was found to be present in twenty-eight instances and
not present in twenty-four, from which I have concluded that it
occurs in half of the Toxopneustes eggs which have been ferti-
lized with Hipponoé sperm.

The point which I wish to emphasize here is that an unpaired
element seen in straight-fertilized Hipponoé eggs is thus shown to
be carried by the Hipponoé sperm.

HIPPONOE 9° X TOXOPNEUSTES &

If we turn now to the figures of the reciprocal cross (figs. 12
to 20), we will find that in all instances three Vs are present while
a hook-shaped chromosome is never found. The identification
of the Vs is sometimes made with difficulty since the arms of the
V may be brought closely together during anaphase, the result
being that the V acquires the appearance of a very thick rod. In
figs. 18 and 20 the Vs may be seen with separated arms; in the
other figures one or more of the Vs are seen with the arms in con-

wach. é
CONSIDERATION OF THE CROSSES

This cross-fertilization has given us a conclusive means of
solving the problem of the origin of the hook-shaped chromosome
in Hipponoé. Let me restate our facts.

1. The hook-shaped chromosome is found in half of the straight-
fertilized Hipponoé eggs. It is an unpaired element. From a
study of these eggs we have no means of determining whether it is
an element peculiar to the egg or to the sperm nucleus.

2. The hook-shaped chromosome is found in one half of the
Toxopneustes eggs which have been fertilized by Hipponoé
sperm. An element of this form does not occur in Toxopneustes,
therefore it has been brought in by the Hipponoé spermatozoan
and occurs in half of the spermatozoa.

3. The hook-shaped chromosome is not present in the Hipponoé

° < Toxopneustes «cross: This indicates that it is not present

20 DAVID H. TENNENT

in the Hipponoé egg and furnishes corroborative evidence that our
conclusion reached in the second statement is correct.

Our analysis has given conclusive evidence that in Hipponoé
there is a heterochromosome which is of paternal origin. ‘This
evidence is of value since it indicates that the conclusion drawn
from Baltzer’s (’09) investigation, which is supported by the work
of Heffner (10), i.e., that in Echinoids the female is the hetero-
gametic sex, while the male is homogametic, is not of general
application, and that in one Echinoid at least, Hipponoé, we have
conditions which are similar to those found in insects.

THE BEHAVIOR OF THE CHROMOSOMES IN CROSS-FERTILIZED
ECHINOID EGGS

A very different phase of my work is concerned with the idea
of the fate of the chromosomes in these crosses and the correla-
tion of this behavior with the environment.

Baltzer (’09, ’10) and Herbst (’09) have shown the fact of chro-
mosome elimination in various crosses. Baltzer (’10, pp. 608-609)
has given a valuable tabular summary of facts in Echinoid
crosses, with the character of the resulting pluteus. In most
instances chromosome elimination is followed by a maternal
pluteus and chromosome retention is followed by an inter-
mediate pluteus. In two crosses, however, Echinus ¢ x Antedon
3 and Strongylocentrotus @ x Antedon 2, there is no elimination,
and the skeleton of the pluteus is maternal in character.

In other crosses, Strongylocentrotus ¢ X Echinus < and Sphae-
rechinus ¢ Xx Arbacia ¢, there is no elimination and a ‘maternal
intermediate’ pluteus results.

For the crosses under consideration in this paper I have pre-
viously shown (’10, 711) that the plutei resulting from the cross,
no matter in which direction the cross has been made, havea
skeleton which resembles that of the Hipponoé pluteus more
nearly than it does that of Toxopneustes, and I have stated that
as a result of this cross we have a Hipponoé dominance with
respect to the character of the skeleton. This dominance is
not complete and in many instances the skeleton is of a ‘Hipponoé
intermediate’ type.

CHROMOSOMES IN ECHINOID EGGS Pail

We may get at the matter before us most quickly by seeking
to ascertain whether in either or both of the Toxopneustes x
Hipponoé crosses there is an elimination of chromosomes.

I am not able at this time to make a final statement as to the
exact number of chromosomes in either Toxopneustes or Hipponoé.
In Toxopneustes it is either 36, 37 or 38. In Hipponoé 32, 33 or
34. Thisis sufficiently near to enable us to determine a noticeable
elimination. In my illustrations the numbers found (and prepara-
tions were selected in which I believe that I was able to count all
of the chromosomes that were present) were fig. 1, upper 34;
lower 34; fig. 2, 30 and 31; fig. 3, 34 and 32; fig. 4, 30 and 31; fig. 5,
aang 33: fig, 7,35 and 34hig. 8, 38 (?) and 35.(7).

Similar counts hold for older segmentation stages up to the 32
to 64 cell stages, so that we have in this cross no chromosome
elimination. The Hipponoé chromosomes are retained and we
have in correlation Hipponoé dominance.

With the reciprocal cross, Hipponoé ¢ x Toxopneustes < the
counts range; fig. 12, 22 and 29; fig. 13, 30 and 36; fig. 14, 28 and
24; fig. 15, 27 and 24; fig. 18, 31 and 32 (those at center not
counted); fig. 17, 30 and 25 (anaphase four cell stage, one at
center not counted); fig. 20, sixteen cell stage, 16 and 16 (two
at center not counted).

It will be seen from these examples that in this cross the
behavior of the chromosomes is irregular from the beginning and
that in some instances there is an elimination of fully half of the
chromosomes, presumably Toxopneustes chromosomes, by the
time the sixteen cell stage has been reached.

The facts regarding the chromosomes and the correlation with
dominance in these crosses may be stated in general terms,
Toxopneustes @ < Hipponoé 2

Without chromosome elimination, Hipponoé dominance.
Hipponoé ¢ xX Toxopneustes 7
With chromosome elimination, Hipponoé dominance.

In the Hipponoé ¢ xX Toxopneustes ¢ cross there is a con-
stant elimination of chromosomes up to the sixteen and immedi-
ately succeeding cell stages until only (?) Hipponoé chromosomes
remain. This elimination is brought about by lagging and failure

-22 DAVID H. TENNENT

to be included in the reconstructed daughter nuclei. In fig. 13,
-A and B, and fig. 14, A and B, many chromosomes which are
lagging in the first division may be seen. In fig. 17, A and B,
a continuation of this elimination by a similar process during the
third division is shown.

A further part of my investigation is concerned with the study
of the eggs fertilized in sea water of decreased alkalinity. Unfor-
tunately I did not obtain eggs in sufficiently late stages of segmen-
tation to enable me to give a final statement of the results. Figs.
9 and 10 are representative of the anaphases of the first division
in Toxopneustes eggs fertilized by Hipponoé sperm in sea water
whose alkalinity had been reduced by the addition of acetic acid.

In all of this material more lagging chromosomes are found
and the average number of chromosomes present in the daughter
plates is smaller than in eggs fertilized in normal sea water.

I had hoped to show that a behavior of the chromosomes cor-
related with the results of the artificial control of dominance
might be demonstrated. Such a correlation would be indicated
by the elimination of Hipponoé chromosomes in the Toxopneustes

9 X Hipponoé ¢ cross.

As the figures show, there is some evidence that such an elimi-
nation takes place but the evidence is not sufficient.

The study of late segmentation stages should determine the
question at once.

DISCUSSION AND COMPARISON WITH RESULTS OF OTHER
INVESTIGATORS

The time has not yet come when we may give a satisfactory
discussion of the meaning of our facts and make a trustworthy
correlation of these with those determined for insects.

Herbst’s (90) work was valuable in showing the elimination
of chromosomes, or rather the failure of the paternal chromo-
somes to take part in the activities of mitosis. But Herbst’s
experiments do not show that a changed environment results in
a change in the character of the pluteus. Chemical fertilization
would have given, as Herbst shows, maternal plutei. Delayed
fertilization of these chemically treated eggs gives plutei of the

CHROMOSOMES IN ECHINOID EGGS 23

maternal type simply because the male nucleus has entered too
late to take a normal part in the process. <A partial union may
give rise to a partially thelykaryotic pluteus. In some cases an
apparently complete union with a subsequent elimination of
Strongylocentrotus chromatin occurred. Baltzer (710) in hisstudy
of the same cross, found that there was no elimination of chro-
mosomes under normal conditions, and that plutei with an
intermediate type of skeleton were found. .

All that can be claimed for the result of Herbst’s treatment
is that the sperm was added so late that, when the modified fusion
of the pronuclei did take place, the effects of the chemical fertili-
zation had attained too great an impetus to be overcome by the
materials of the paternal nucleus. He obtained varying degrees
of modification. All were not modified; in all, the result depended
on the amount of development that had been attained in the
thelykaryotic activities.

This is precisely the result that I obtained with the starfish
egg (’06) by the superposition of fertilization on artificial partheno-
genesis. If the sperm were added before the egg nucleus had
gone too far, a union of the pronuclei and a normal division fol-
lowed. If the addition were made later, the paternal chromatin
entered the mitosis irregularly and was in part rejected.

Baltzer’s important results lie in his accumulation of facts
regarding individual chromosomes and in his determination of
elimination and non elimination.

I have shown that his idea of the female Echinoidas the hetero-
gametic sex must be restricted to the cases in which it has been
observed and cannot be used as a general interpretation.

I realize the impossibility of interpreting the results of another
investigator from drawings, without having seen the sections
from which the illustrations were made. It is therefore with
hesitation that I suggest that Baltzer’s fig. 23a (10) shows for
the Sphaerechinus ¢ X Strongylocentrotus < cross exactly what
I have shown for Toxopneustes ° xX Hipponoé @.

Baltzer (09) has shown that the particular chromosomes of
Strongylocentrotus are two long hooks or two long hooks and a
short.hook. (See Baltzer, ’09, figs. la, 1b, 5a, 5b and 11, 12 a-d.)

24 DAVID H. TENNENT

The short hook is definitely localized, according to Baltzer, in the
ege (text and diagrams, p.579). The fact that it does not occur
in his figs. 16a and 6 nor in 17a, b and c harmonizes with this idea;
but when we turn to fig. 23a (Baltzer, ’10)—Sphaerechinus. ¢ xX
Strongylocentrotus #— we find what might be interpreted from
the illustration as a short hook. The author definitely states
(10, p. 509), that he has never found a hook-shaped chromosome
in Sphaerechinus. If both of these apparent hooks be such, one
of them must be the unpaired element seen in half of the straight-
fertilized Strongylocentrotus eggs. It is also difficult to under-
stand why the elongated rods in the anaphase plates (Baltzer’s
metaphase) in figs. 66, 7b and 11 (’09) should not be regarded as
long rods, as in Echinus (Baltzer 2a and 6, 3a and b). My hesi-
tation in speaking of these possible interpretations is overcome
only by the fact that the conditions in the illustrations are similar
to division figures in my own material.

The investigations of Heffner (’10) on Toxopneustes show that
this is in some respects like Strongylocentrotus. Heffner shows
for straight fertilized Toxopneustes eggs two Vs three Vs and
two long rods. The Vs may be regarded as comparable to the
hooks. The long rods in Toxopneustes are like the long rods shown
in Baltzer’s figures of Strongylocentrotus.. We need facts con-
cerning the chromosomes in many species of Echinoids. They
should be studied not only in straight-fertilized eggs, but in
crosses, in chemically fertilized material and in fertilized enucle-
ated egg fragments.

bo
Or

CHROMOSOMES IN ECHINOID EGGS

LITERATURE CITED
BautzeR, F. 1909 Die Chromosomen von Strongylocentrotus lividus und
Echinus microtuberculatus. Arch. f. Zellforsch., bd. 2.

1910 Uber die Beziehung zwischen dem Chromatin und der Entwick-
lung. Arch. f. Zellforsch., bd. 5.

‘Herrner, B. 1910 A study of chromosomes of Toxopneustes variegatus which
show individual peculiarities of form. Biol. Bull., vol. 19.

Hergst, C. 1909 Vererbungsstudien VI. Arch. Entwicklungsmech. bd. 27.

Pinney, M.E. 1911 A study of the chromosomes of Hipponoé esculenta. Biol.
Bull., vol. 21.

TENNENT AND Hocure 1906 Studies on the development of the starfish egg.
Jour. Exp. Zodl., vol. 3.

TENNENT, D.H. 1908 The chromosomes in cross-fertilized echinoid eggs. Biol.
Bull., vol. 15.

1910 Dominance of maternal or of paternal characters in echinoderm
hybrids. Arch. f. Entwicklungsmech., vol. 29.

191la Echinoderm hybridization. Carnegie Institution pub. 132.

1911b A heterochromosome of male origin in echinoids. Biol. Bull,
WO, Pil.

PLATE 1
EXPLANATION OF FIGURES

All figures are drawn to a magnification of 1500 diameters.

1 A, B, C, D. Toxopneustes 2 X Hipponoé &@. Normal sea water. Three
longitudinal sections of anaphase of first division. Vs shown in B were lying
beneath Vsin A. Chromosomes: upper plate, 34; lower, 34.

2 A, B, C. Toxopneustes 9 XX Hipponoéc’. Normal sea water. Three
longitudinal sections of anaphase of first division; chromosomes: 30 and 31.

3 A, B,C. Same; chromosomes: 34 and 32.

4 A,B. Same; three sections combined in two; chromosomes: 39 and 31.

5 A, B, C. Same; chromosomes: 33 and 33.

6 Same; early anaphase; part of one section.

7 A,B. Same; three sections combined in two; chromosomes: 33 and 34.

8 A, B. Same; three sections combined in two; chromosomes: 38 (?) and
35 (7). Some of chromosomes probably sectioned.

9 A,B. Same; sea water and acetic acid; chromosomes 28 and 29; those at
center not counted.

26

CHROMOSOMES IN ECHINOID EG(¢ PLATE 1

an ee f at

ein, | a we J

iD
aT ri \ : |
f a WA aN 6
ee eee
i A ar J\, is NW iy NY fy Ara
2C

Aine CTA a
v .

due Nf ; INOS ny

ee I i \
Pony diy ) lay
oe as
ind AA NA dein
amet 4B 8A ° 8B

Nu! Neary yl Wty
i \ 2° \t Ny \ lity it (hy
oe

Wave MCh ty
Ma AN () “gen
aN

10

11
12

PLATE 2
EXPLANATION OF FIGURES

Same; sea water and acetic acid; one section of spindle showing lagging.
A, B,C. Same};sea water and acetic acid; chromosomes: 30 and 35.
A,B. Hipponoé 9 X Toxopneustes o. Longitudinal sections of anaphase

of first division; chromosomes: 22 and 29.

18 Same; chromosomes: 39 and 36.

14 Same; chromosomes: 28 and 24.

15 A,B. Same; chromosomes: 27 and 24.

16 Samal: single section showing lagging.

17 A,B. Same; anaphase of third division; chvomosonen 30 and 25.

18 as B. Same; chromosomes: 31 and 32; cee at center not counted.

19 Same; single eertion showing lagging.

20 A,B. Same; anaphase in one cell of sixteen cell stage; chromosomes: 16
and 16; two at center not counted. °

CHROMOSOMES IN ee EGGS PLATE 2
DAVID H. TEN

CH Ne My Vth) ut
“3
(yd
10
Nie aa cide Al
See he

A

oW\ \Y
‘Kit a
13B V

CGIAR
KS ee Li Sib
\ Y i

\ | : Kaine \ Ni
: \ :
| Willy ; | v oe |
_- 7 \s, \y Wr iyMy iy
17A 178
je.
aus Lac ae : Ath
yy \/ He oeiete
Ne Sy
i 18B 20A /| s
DAT, £8.

29

THE SKULL STRUCTURE OF DIPLOCAULUS MAG-
NICORNIS COPE AND THE AMPHIBIAN
ORDER DIPLOCAULIA

ROY L. MOODIE

From the Zoological Laboratory, the University of Kansas
SEVEN FIGURES

The Permian vertebrate known as Diplocaulus magnicornis
Cope is one of the most aberrant and specialized of all the extinct
Amphibia. The species was first described by Cope from frag-
ments of several crania and portions of the vertebral column;
material which had been collected in the Permian of Texas prior
to 1882. The genus had, however, been established previously
on fragmentary material which had been discovered by Dr. J.C.
Winslow and Mr. W. F. E. Gurley in the Pennsylvanian of Ver-
milion County, Illinois. The genus Diplocaulus was first located
by Cope in 1881 among the Pelycosauria, but later researches
inclined him to place the form among the stegocephalous Am-
phibia with relationships to the Microsauria.

The skull of Diplocaulus magnicornis is very peculiar in the
elongation of the posterior elements of the upper surface of the
cranium. The anterior elements do not take part in the posterior
prolongations which give the skull such a bizarre appearance.
This specialization is due to the extreme elongation of the supra-
temporal, the squamosal, the parietal, the epiotic and the supra-
occipital.

The pineal foramen is apparently absent. I was unable to dis-
cover it on a well preserved skull in the collection of the Univer-
sity of Chicago (No. 2, U. of C. Collections). Cope figured it
on the skull of this species which he studied in 1895. Broili
says nothing of the presence of the opening and does not figure
it in the restoration of the skull which he gave in 1902. Williston

Ba

32 ROY L. MOODIE

was unable to locate it on the skull of Diplocaulus hmbatus Cope
which he figured in 1909.

The nostrils are located far forward on the anterior edge of the
skull which curves slightly downwards so that they look forward.
The orbits are small, almost circular, and are situated far forward
as is the case in many of the carboniferous Microsauria. Other
than these there are no openings on the dorsum of the skull.

The arrangement of the cranial elements is found to be approxi-
mately as Cope gave them in 1895 although there were discovered
in the complete specimen evidences of the postorbital which had
not been previously detected. Williston was unable to locate this
element in the excellent skull of Diplocaulus limbatus Cope and
concluded that the element behind the orbit is the postorbitofrontal
(Trans. Kans. Acad. Science, 1908, pl. 1a) an interpretation which
is open to much question. We have yet to have a definite proof
of the union of these two elements, the postorbital and postfrontal,
in any vertebrate skull. If there be only one present it is either
one element or the other and not both. Further discussion of
this will be postponed for a paper on the development of the alli-
gator skull. The outlines of the jugal and quadratojugal were
also determined. The suture separating the parietal and the
supraoccipital was not found to be so erratic as Cope figured it
but it continues directly across the skull. This may easily have
been an individual variation. The suture separating the frontal
into two equal parts was not detected although careful search was
made for it. This seems to be the only case on record among the
Stegocephala in which there has been an actual fusion of two
paired elements of the cranium. - Maggi has made some interest-
ing suggestions as to the origin of the interparietal of mammals
and its correlation with the fused epiotics of the Stegocephala.
In discussing Maggi’s paper it was stated that there was no case of
a fusion of any of the cranial elements of the Stegocephala known.
This I believe to be an error as is demonstrated in the present
skull. This would not, however, change the decision in regard to
Magei’s conclusions.

On the dorsum of the skull there were detected in several places
evidences of the lateral line canals which occur as shallow grooves.

DIPLOCAULUS MAGNICORNIS oo

These are almost universally known as the ‘slime canals of the
Stegocephala,’ a decided misnomer since the lateral line canals
have nothing to do with the production of slime. The canals
were also detected on the mandible which is associated with the
present skull. On the right of the skull (fig. 1) there will be
seen a distinct groove which runs along the edge of the skull.
This canal had been called by Allis in Amia “the anterior portion
of the infraorbital.’”’ There were also detected portions of the
supraorbital canals but the skull is so badly crushed that it is
impossible to follow the complete course of the canals.

The palate of the skull, as here given, is an advance over any-
thing heretofore known. There still remains much to be deter-
mined in regard to its structure, but the following is offered
as a contribution to the more complete knowledge of the subject.
Cope, in 1895, gave a figure of the anterior portion of the palate
and Broili, in 1902, gave further notes on its structure. Broil,
however, had but a small portion of the skull of the animal on
which to base his conclusions. Unless there be an extreme varia-
tion in the shape assumed by the skull of this species Broili’s
restoration is at fault in regard to the posterior curve of the skull,
there being no indication of such a condition in the present speci-
men, nor in the species D. limbatus Cope. Cope gave very accu-
rately the positions of the teeth on the palate but was unable
to determine the elements which bore the teeth. In the present
skull the tooth-bearing elements are found to be the premaxillae,
the maxillae, the vomers, the palatines and the transverse bones.

There are five pairs of openings and depressions on the palate
of the skull. These are: the internal nares, correctly repre-
sented by Cope and Broili; the palatine vacuities; the depressions
which Broili calls the ‘Ohrenschlitzgruben’ or auditory fossae:
the infratemporal foramina and depressions along the posterior
lateral border of the palate which appear to be partly due to a
folding over of the skull elements and doubtless gave place for
the attachment of the masseter and temporalis muscles. The
arrangement of the palatal openings of Diplocaulus does not
differ in any essential respect from the condition found in the
larger Stereospondylia.

JOURNAL OF MORPHOLOGY VOL. 23, NO. 1

34 ROY L. MOODIE

It is worthy of note that the palatal elements in Diplocaulus
do not take part in the formation of the prolongation of the
dorsum of the skull and we get an idea of the primitive condition
of the skull of the Diplocaulidae from the circumscribed area of
the palatal elements which are restricted by the quadrates to the
anterior portion of the skull so that all of the elongation and expan-
sion has taken place in the dorsum.

The palatal openings are all lateral in position. The internal
nares are the most anterior. They are small, oval and transverse
in position. They are bounded by the premaxillae, vomers
and palatines. The palatine foramina are large, oval openings
situated below the orbits on either side of the median line. Their
long axis is parallel to the axis of the skull. They are bounded
by the parasphenoid, transverse, vomers and palatines. The
infratemporal foramina lie anterior and somewhat medial to the
quadrates. The openings have a rounded triangular form and
are bordered by the transverse, the maxilla, the quadratojugal,
the quadrate and the pterygoids. Posterior to the quadrate there
is an elongate groove which is possibly homologous with the
quadrate foramen of the Pelycosaurian genus Dimetrodon,
of Sphenodon, Anaschisma and other stereospondylous forms.
Its function here seems to be for the attachment of the masseter
and temporalis muscles. It certainly has the position of the
quadrate foramen in other forms. Its elongation is due to
the backward growth of the epiotic horns. The other opening
marked es in fig. 6, is undoubtedly the external auditory meatus.
It represents in part the otic notch or ear slit of other Stegocephala
so well shown in Metoposaurus, Mastodonsaurus, and Archego-
saurus. Broili has called them the ‘Ohrenschlitzgruben’ and he
is undoubtedly right.

The mandible of Diplocaulus magnicornis Cope is moderately
heavy, though comparatively slight when compared to the size
of the skull. The sutures on the mandible have been impossible
to determine, with the exception of those bounding the articular.
They show the articular to have been a triangular element. The
teeth of the mandible consist of about thirty-five to forty short
blunt cones. The form of tooth appears to be well adapted to

DIPLOCAULUS MAGNICORNIS aa

crushing shell fish, as Case suggests, and Diplocaulus may have
fed on some of the smaller Mollusca of the Permian rivers and
lakes. On the lateral face of the mandible, there is a distinct
groove, the operculo-mandibular canal of the lateral line system.
It would be interesting matter to determine if this canal extended
entirely around the mandible.

Dr. Case in 1908 published a restoration of the entire animal, as
the structure seemed to him to demand. However, Case neg-
lected the insertion of the clavicular girdle which was already
known and which would seem to indicate the presence of limbs.
As a matter of fact limbs are still unknown in this species although
Williston has recorded the discovery of small limbs in the closely
related species D. limbatus Cope. That limbs will ultimately
be discovered in the present species can not be doubted. The
habits of the animal were undoubtedly as Dr. Case has suggested
for them (Pop. Sci. Monthly, December, ’08).
~ Paleontology teaches us nothing as yet of the ancestry of this
peculiar genus of amphibians nor have we any record of its de-
scendants. It is one of those peculiar forms which stands alone.
It shows, however, characters which are more nearly those of
the Branchiosauria than of the Microsauria in which order it is
usually placed. The characters separating the early orders of
Amphibia are essentially those of the ribs and vertebrae. The
structure of the skull is essentially similar in all of the groups.
We are not able, from the structure and composition of the skull,
to distinguish a branchiosaurian from a microsaurian. The
characters of the ribs and vertebrae are, however, perfectly
constant and distinctive. Of course there are certain superficial
characters of the skull which hold true for all branchiosaurs and
microsaurs such as the absence or presence of sculpture of the
cranial elements and the absence of the lateral line grooves from
the skulls of the Branchiosauria. Other characters such as
the presence of external branchiae in the Branchiosauria, the lack
of endochondral ossification in the long bones and absence of
clawed digits would seem to be of considerable importance.

Except for superficial characters the skull of Diplocaulus mag-
nicornis Cope is essentially similar to those of the Branchiosauria

36 ROY L. MOODIE

and Microsauria in structure and composition. The elongation
of the epiotic regions to form the wide, fan-shaped, horns, the
fusion of the frontals and the absence of a parietal foramen are
individual or ordinal characters the importance of which is open
to debate. That the fan-shaped horns were developed for the
protection of gills would seem most absurd. [ the creatures
had gills the horns probably served to protect them but there
is no evidence whatever that these forms were branchiate. Horns
of a similar character are developed in many of the Microsauria
in genera which are otherwise and structurally unrelated. Just
what the development of these horns may mean is a difficult
problem. The solution offered by Beecher of the significance of
spines and horny excrescences indicating decadence may be a
good one here but we know so very little about these creatures
that conclusions would be premature.

When we consider the characters broadly we perceive that
they indicate a group separation of the species of Diplocaulus
as I have already indicated (Geol. Mag., May, ’09, p. 220). The
characters which have been discussed ally the present genus with
the Branchiosauria rather than with the Microsauria. The only
character of the microsaurs which the species of Diplocaulus
possess is the sculptured nature of the elements of the clavicular
girdle and cranial elements. The characters of the ribs and ver-
tebrae are essentially those of the Branchiosauria but they differ
from these in the specialization of the zygosphene and zygantrum,
which have not been detected in the branchiosaurs, and in the
structure of the ribs. The fact that the ribs are borne on the
middle of the centrum on an elongate transverse process would be
sufficient to indicate its complete separation from the Micro-
sauria in which the ribs are universally intercentral. The pres-
ence of an epicondylar foramen in the humerus is another distine-
tive character of Diplocaulus and entirely lacking in both Bran-
chiosauria and Microsauria. In short the characters presented
by Diplocaulus are so confusing and contradictory that they compel
us to perceive that after all a final classification is impossible.
There will always be new classifications so long as there are new
forms and new intellects at work upon the material. But if we

DIPLOCAULUS MAGNICORNIS oil

must have a classification for convenience why not have it consist-
ent at least? If we use a character to distinguish two orders of
Amphibia like the Branchiosauria and Microsauria, and the char-
acter has been accepted for nearly half a century, then why
not apply the same rule to another group of amphibians, and on
the structure of the vertebrae, ribs and limb bones establish the
new order Diplocaulia? It seems consistent at least if nothing
more.

In pursuance of this I give here the ordinal characters of this
new group of Amphibia-Diplocaulia: Skull proportionately very
large with epiotic angles drawn out into fan-shaped horns, the
expansion being due to the supratemporal, epiotic, parietal,
squamosal and supraoccipital. Frontals fused into a single
plate. Lachrymal probably absent. Pineal foramen absent.
Orbits small, circular and anteriorly placed. Sclerotic plates
unknown. Nostrils on or near the anterior edge of cranium.
Teeth borne on mandible, premaxillae, maxillae, palatines, vomers
and ectopterygoids. Teeth rounded, acrodont, denticles, abun-
dantly present and apparently fitted for crushing hard substances.
Palatal region restricted to the anterior portion of the skull.
It does not take part in the epiotic prolongation. The palatal
aspect of the cranium interrupted by five paired openings which
are: the internal nares, the palatine foramina, the infratemporal
foramina, the quadrate foramina for the attachment of the masse-
ter and temporalis muscles and the auditory slits or external
auditory meatus. The occipital condyles occur under the pro-
jecting shelf of the supraoccipital plates. Basioccipital partly
cartilaginous and condyles borne by exoccipitals. Lateral line
grooves present on skull and mandible.

Atlas ribless and essentially urodelous in structure. The ribs
bicipital and borne on large transverse processes springing from
the arch and centrum. The zygosphenal articulation not so
well developed as the zygopophysial one. Vertebral formula
unknown. Vertebrae elongate with low spine. Notochord but
partly persistent and absent in the middle portion of the centrum,
persisting as a double cone in the intervertebral regions. Clavie-
ular girdle composed of interclavicle, clavicles and coracoids (?)

38 ROY L. MOODIE

the first two of which are sculptured. These are quite large and
indicate the presence of limbs in species where actual limbs have
not been found. Humerus with endochondral ossifications resem-
bling the Branchiosauria. Epicondylar foramen and muscular
expansions present. Carpus and tarsus unossified. Femur elon-
gate and somewhat twisted.

There are three species of this order known. They are:

Diplocaulus salamandroides Cope, described from fragmentary
material collected in the upper carboniferous beds of Salt Creek,
Vermilion County, [linois, in 1877.

Diplocaulus limbatus Cope, described from fragmentary
material from the Permian of Texas. Further descriptions and
figures of the skull, girdles and limb bones given by Williston in
1909.

Diplocaulus magnicornis Cope, described from a nearly com-
plete cranium from the Permian of Texas.

Closely related forms of this group are possibly to be found in
the Crossotelidae from the Permian of Oklahoma, but the group
is as yet very imperfectly known.

BIBLIOGRAPHY
Broom, R. 1910 Comparison of the Permian reptiles of North America with
those of South Africa. Bull. Amer. Mus. Nat. Hist., vol. 28, p. 214.

Corr, E. D. 1882 Third contribution to the history of the vertebrata of the
Permian formation of Texas. Proc. Amer. Phil. Soc., vol. 20, p. 453.

1881 Catalogue of vertebrata of the Permian formation of the United
States. Amer. Nat., vol. 15, p. 162.

1882 Permian vertebrata, Amer. Nat., vol. 16, p. 925.

1888 Systematic catalogue of vertebrata from the Permian. ‘Trans.
Amer. Phil. Soc., vol. 16, p. 286.

1896 The reptilian order Cotylosauria. Proce. Amer. Phil. Soc., vol.
34, p. 455. Pl. 9.

Miter, 8. A. 1889 N. A. Geology and paleontology. p. 621.

Casz, E. C. 1900 Vertebrates from Permian bone bed, Illinois. Journ. Geol.,
vol. 8, p. 710.

1908 A great Permian delta and its vertebrate life. Pop. Sci. Monthly,
vol. 73, p. 567, figs. 12, 18.

DIPLOCAULUS MAGNICORNIS 39

Broiui, F. 1902 Beitrage zur Kenntniss von Diplocaulus Cope. Centralblatt

fiir Mineralogie, p. 536.

1904 Permische Stegocephalen und Reptilien. Paleontographica, Bd.
Hil, Joo ey folie Ik IY, We

JAEKEL, Orro 1903 Uber Ceraterpeton, Diceratosaurus und Diplocaulus.

Neues Jahrb. Mineral., p. 126.

Moopig, Roy L. 1908 The dawn of quadrupeds in North America. Pop. Sci.

Monthly, vol. 72, p. 565, fig. 5.

1908 The lateral line system in extinct amphibia. Jour. Morph., vol.
19, p. 522, figs. 9, 9a, 10.

1908 Carboniferous quadrupeds. Trans. Kans. Acad. Science, vol.
22, p. 243.

1909 The Microsauria. Geol. Mag., Dec., vol. 6, p. 220.

Wiuuiston, 8. W. 1908 The skull and extremities of Diplocaulus. Trans.

Kans. Acad. Science, vol. 22, p. 122, pls. 1-6. Describes more fully
Diplocaulus limbatus Cope, locates Diplocaulus in Microsauria.

1910 Dissorophus. Journ. Geol., vol. 18, p.534. Gives list of Permian
amphibia of North America.

PLATE 1

EXPLANATION OF FIGURES

1 Dorsum of skull of Diplocaulus magnicornis Cope. Infraorbital canal at
point of arrow. X 3.

2 Ventral surface of the mandible. The operculo-mandibular canal at the
point of the arrow. X 3.

3 Oblique view of one ramus of the mandible to show the entire course of the
operculo-mandibular canal. Nearly natural size.

40

1

=,
u

PLAT

NICORNIS

x
MOODIB

DIPLOCAULUS MAC

L.

ROY

1

JOURNAL OF MORPHOLOGY, VOL. 23, NO.

41

PLATE 2
EXPLANATION OF FIGURES

4 Outline of the cranial elements of Diplocaulus magnicornis Cope. £, epiotic;
F, frontal; J, jugal; Max, maxilla; NV, nasal; P, parietal; Pf, postfrontal; Po, post-
orbital; Pr, prefrontal; Px, premaxilla; Qj, quadratojugal; So, supraoccipital;
Sq, squamosal; St, supratemporal.

5 Mandible from the side to show arrangement of the operculo-mandibular
canal,

6 Outline of the openings and elements of the palate of the skull. Af, internal
nares; Co, condyle; Ep, epiotic; Hs, auditory fossa or ear slit; Hro, exoccipital;
Mz, maxilla; Pa, parasphenoid; Paf, palatine foramen; Pal, palatine; Pt, ptery-
goid; Px, premaxilla; Q, quadrate; Qj, quadratojugal; 7’, ectopterygoid or trans-
verse bone.

7 Side view of theskull. J, jugal; Mx, maxilla; Pr, prefrontal; Px, premaxilla;
Qj, quadratojugal; Sy. squamosal.

42

PLATE 2

DIPLOCAULUS MAGNICORNIS

ROY L. MOODIE

MODIFICATIONS IN THE TESTES OF HYBRIDS FROM
THE GUINEA AND THE COMMON FOWL}

MICHAEL F. GUYER

From the Zoological Department, University of Wisconsin

TWENTY-THREE FIGURES IN TWO PLATES

The following study is based on material obtained from four
guinea-chicken hybrids, the offspring of a yearling black lang-
shan cock and a common guinea hen three years old. There were
originally five of the hybrids, all male, but during my absence
from Cincinnati one died and was not preserved. An account of
the general habits and appearance of these fowls together with
an analysis of their peculiar color pattern has already been pub-
lished.”

Three of the hybrids were killed at the age of three years,
one at the age of six, and the last one died at the age of seven.
When five years old the two older ones each developed a pair of
sickle feathers in the tail similar to those so characteristic of
the ordinary domestic cock, although previous to this time all
alike possessed the simpler type of tail feathers seen in the com-
mon hen.

When young the hybrids resembled young guineas in appear-
ance except for the fact that the legs were feathered after the
manner of the langshan breed of. chickens. These feathers dis-
appeared for the most part after a few months, leaving only a
few scattering ones on the legs of the adults. All of the hybrid
fowls were infertile. Two of them were kept at the Cincinnati
Zoological Garden for a number of years in a large inclosure with

1 Prepared for The Whitman Memorial Volume, but received too late to be
included.

2 Guyer, M. F.: Atavism in guinea-chicken hybrids. Jour. Exp. Zool., vol. 7,
1909. Also, La livrée du plumage chez les hybrides de pintade et de poule. Bul.
Mus. d’hist. nat., Paris, 1909.

45

46 MICHAEL F. GUYER

various other gallinaceous birds, such as peafowls, pheasants,
guineas, and bantam chickens. As long as both were alive they
remained together constantly but.after the one was killed the
other attached itself to the guinea contingent of the enclosure.
Whether this was due to an instinct of kinship or whether it was
the result of earlier associations with the guinea mother I am
unable to say.

Upon opening the body cavity the testes in three of the hybrids
were found to be normal in size and external appearance. In
the fourth, while the left testis was somewhat smaller than the
average, the right was greatly hypertrophied, weighing 90 grams
and measuring 85 mm. long by 54 mm. broad, by 30 mm. thick.
The enlarged testis was kidney-shaped and lay diagonally across
the body cavity. Two distinct regions, separated externally
by a shallow furrow and internally by more or less of a connective
tissue septum, were visible. The anterior region, representing
about one-third of the testis, was a white, waxy, fatty mass. The
remaining portion was somewhat more firm and was richly
supplied with small blood vessels which at intervals exhibited
numerous plexuses and varicosities.

Toward the anterior end of this hypertrophied organ and marked
off from it as a distinct body by a constricted band of connective
tissue was what appeared to be a small accessory testis. Subse-
quent microscopic examination showed that the left testis, the
posterior region of the hypertrophied right, and this smaller
accessory body all contained seminiferous tubules, although
they were comparatively scarce in the hypertrophied body. Such
a hypertrophied condition of one testis has been noted by other
students of hybrids such as Suchetet,* for instance, who in speak-
ing of a hybrid duck (C. moschata x A. boschas) states that one
of the two testes had the dimensions greatly exaggerated, and
with its juices devoid of spermatozoa. Although, with the
exception just mentioned, the testes of the hybrids appeared
normal from a macroscopic examination, microscopic investiga-
tion showed them to be markedly abnormal; to such a degree in
fact that no trace of spermatozoa were observable although in

’Suchetet, André: Des hybrides 4l’etat sauvage; Oiseaux, Tome 1, 1896. Lille.

TESTES OF GUINEA-CHICKEN HYBRIDS 47

certain favorable regions spermatogenesis was seen to be in prog-
ress as far as the formation of spermatids. None of the latter,
however, was ever found in process of transformation into the
spermatozoon.

Upon examining sections of the testis the first thing to strike
the attention was the great scarcity of seminiferous tubules.
While a few of these were visible in sectionsfromnearly all regions,
they existed for the most part as narrow scattered tubules (fig. 1)
surrounded by vast areas of intervening tissue, or more frequently
as small island-like groups of a few larger tubules in a more or less
homogeneous field of connective tissue and stroma cells (fig. 2).
At times, however, limited areas (fig. 3) would be encountered
in which the seminiferous tubules were almost as plentiful as in
normal non-hybrid individuals. One hybrid in particular differed
markedly from the others in having a plentiful supply of tubules
throughout the most of the testis. In certain of these tubules divi-
sion stages of spermatogonia and first spermatocytes were plentiful.
Less frequently dividing spermatocytes of the second order were
found. Side by side with such tubules, and often in greater
numbers, would be found other tubules in process of degeneration.

The seminiferous tubules of the guinea have a thicker investing
rind or wall than do those of the ordinary domestic cock, and while
there is considerable fluctuation in the diameter of different
tubules in both species, on the whole, the former has a noticeably
broader, coarser looking type of tubule. There is also slightly
more interstitial tissue in the testis of the guinea. In all of these
respects, as well as in the general appearance of the various mitoses
the hybrids approximate more nearly the condition found in the
guinea,

In the hybrids, as already stated, the seminiferous tubules
were usually few in number and widely separated by intervening
tissue. Under the low power of the microscope this inter-tubular
field ordinarily had a homogeneous cellular appearance, but
under a moderate or high power the cells were seen to be arranged
in irregular cords or strands, although this corded appearance
often graduated indistinguishably into a homogeneous field. The
stroma or interstitial cells occupying most of the inter-tubular

48 MICHAEL F. GUYER

field are for the most part characterized by their round, plump,
slightly eccentric nuclei with a comparatively large, well-marked
sphere of denser appearing cytoplasm at one side (fig. 10). In
places, however, they resemble closely the spermatogonia lining
the inner wall of the tubules and, apart from location, are practi-
cally indistinguishable from them. Not infrequently, moreover,
scattered here and there among them are to be seen typical sper-
matocytes in process of division. Such areas appear to be the
remains of former tubules of which the walls have become entirely
obliterated and the characteristic arrangement of the tubule
cells lost. Such inter-tubular areas with numerous cells exhibit-
ing various features of the maturation phenomena were especially
plentiful in the hybrid already indicated as possessing a great
number of tubules (fig. 4).

The germ-cells in many of the tubules were in process of degen-
eration. In some cases only a peripheral row of spermatogonia
and Sertoli cells remained but more commonly the cells progressed
to the spermatocyte type and then halted at the point of synapsis.
So common was this, indeed, that the characteristic appearance
of the tubules was that of a mass of cells in the contraction phase
of the primary spermatocytes (fig. 9). The deeply staining chro-
matin strands became massed at one side of the nucleus, commonly
lying in more or less of a crescent along the inner surface of the
nuclear membrane. Frequently such nuclei had a vacuolar,
abnormal appearance as if on the point of dissolution. Less
often the nuclei of the spermatocytes had a clear watery looking
center with the chromatin spread around the periphery. Nota
few syncytial masses existed containing numerous degenerating
nuclei, among which there were often evidences of fusion or of
direct division.

As to why so many of the cells should be unable to progress
beyond the beginning of the synaptic phase I can offer no further
suggestion than that made in connection with a similar study‘

4 Guyer, M. F.: Spermatogenesis of normal and of hybrid pigeons. Disser-
tation, University of Chicago, 1900. Later published as Bul. 22, University of
Cincinnati, 1903. In case this paper is inaccessible to any investigator, the author
will gladly supply copies as long as his stock of reprints lasts.

TESTES OF GUINEA-CHICKEN HYBRIDS 49

carried out on hybrid pigeons from 1897 to 1900; namely, that
(p. 46)

in hybrids it may be supposed that in the ordinary cells of the body, the
chromosomes from the paternal and the maternal species lie side by side
ana carry on the customary functions of the cells but when it comes to an
actual fusion of chromosomes to form the bivalent type necessary for
reduction, the incompatibility of the two different plasmas renders the
union incomplete or prevents it entirely.

It may be of some interest in connection with the present mem-
orial volume to record the fact that practically all of this earlier
work on hybrid pigeons was done on material supplied by
Professor Whitman himself. He was particularly curious about
the preponderance of males among his hybrids, and was much
interested in the interpretation expressed in my doctoral thesis
of 1900 to the effect that the refusal of the chromosomes of hybrids
to unite in the first spermatocyte indicated that synapsis normally
was a fusion of maternal with paternal chromosomes, and that,
granting this to be true, in fertile hybrids in the segregation of the
chromosomes after synapsis we find a plausible reason for returns
in the third generation to grandparental characteristics. Since
this paper has had but a limited circulation it may not be amiss
to restate briefly my conclusions at that time regarding this
point. In a paper in 1899> I had already pointed out the fact,
illustrating with a diagram, that when white ring-doves and brown
ring-doves are cross-mated and their offspring interbred, there is
frequently a return in color in the third generation to the grand-
parental types. This phenomenon we now recognize at once as
Mendelian but at that time Mendelism had not yet been redis-
covered. In my thesis! of the next year I restated these facts of
my 1899 paper a little more explicitly as follows (p. 35-36):

Offspring of the common ring dove when crossed with the white ring
dove are brown in color. One member of the resulting pair is frequently
a few shades lighter in color than the other. In the next or third genera-
tion there is generally a return to the original colors of the grandparents;
one of the young is white, the other brown. Occasionally both of the

young are brown or, less frequently, both white. There is a marked
tendency for the white ones to be female and the brown ones male.

5 Zoological Bulletin, vol. 2, no. 5; 1899.
* Loc. cit.

50 MICHAEL F. GUYER

. . So far as the writer has carried his experiments, the indi-
cations are that on the whole there are more brown than white birds in
the third generation, and this points to the conclusion that in the brown
birds we may have both intermediate forms like the hybrids of the second
generation and forms which have reverted to the brown grandparent,
as the white doves have seemingly returned to the white grandparent.

: The birds of this generation, then, might mate in such a way
that the offspring could exhibit the ancestral white while yet remaining
intermediate in other characters. As we shall see in the conclusions
from the study of the germ-cells of hybrids, there are certain phenomena
in the germ-cells which apparently afford us a definite physical basis
for the production of intermediate forms and for returns to pure ancestral
species. From this basis there must necessarily be a greater number of
intermediate forms in the offspring of hybrids than there are reversions
to the respective ancestral species.

I had interpreted the irregular phenomena occurring in hybrids
at the time of synapsis as due to the tendency of the chromatin
of each parent to retain its own individuality. And while I had
attributed some importance to these irregularities of division in
accounting for variations and reversions in the third generation,
I did not regard them as the chief factors in such returns, as is
evident from the following excerpt (p. 47):

In discussing irregular divisions, however, it must not be forgotten
that many apparently normal divisions of the spermatocytes also occur
in hybrids, and constitute by far the predominant kind of division in
hybrids from closely related forms. Unequal distribution of chromatin
can not therefore play the most important part in variation or reversion.
There seems to be no other interpretation, indeed, than that in the many
normal mitoses of the bivalent chromosomes which occur, the chromatin
of the father and of the mother is set apart so that the ultimate germ-
cells are what might be termed ‘pure’ cells; that is a given egg or sperm-
cell contains exclusively or at least predominantly qualities from one
parent. The offspring from fertile hybrids of the same parentage might
then be similar to the mixed type of the original hybrid, or revert to one
of the grandparent types, dependent upon the chances of the various
cells for union at fertilization. If a spermatozoon and an egg containing
characteristics of the same species unite, then the reversion will be to
that species; if a sperm-cell containing the characteristics of one species
happens to unite with an ovum containing characteristics of the other
species, then the offspring will be of the mixed type again. By the
law of probability the latter will be the more prevalent occurrence,
because there are four combinations possible, and two of the four would
result in the production of mixed offspring, while only one combination
could result in a return to one of the ancestral species.

TESTES OF GUINEA-CHICKEN HYBRIDS 51

This, it will be seen, is a close approximation to the Mendelian
statement of germinal segregation and combination. The quali-
fications necessary to bring it more strictly within the pale of Men-
delism as then known were made in a brief paper® early in 1903.

As previously stated, in many of the seminiferous tubules of
the guinea-chicken hybrid synapsis occurred and occasionally
some cells progressed to the completion of the second division
and the formation of spermatids. But even where synapsis
was effected there was more or less of a tendency for the union to
be incomplete or partial. This was evidenced by the unusual
bipartite appearance of the conjugated chromosomes and in an
occasional excess of chromosomes over the number (nine) charac-
teristic of the corresponding stage in the chicken or the guinea.
In cases of such excess the extra chromosomes had the smaller
size of the univalent type.

Even in normal spermatogenesis one not infrequently encoun-
ters fluctuations in the number of chromosomes. This is true
to such an extent indeed that, judging from my own studies on
the chromosomes of various birds and man, one is led strongly to
the opinion that in these instances at least we are dealing with
compound chromosomes which may occasionally resolve them-
selves into simpler components and in consequence exist in
greater than the typical number when ready for division. On the
other hand, instead of an increase there may be a reduction below
the recognized haploid number, as I have shown to be true in
man where the spermatocytes of the second order typically appear
as five (or seven with the accessories) instead of the expected
ten (or twelve); or in the case of-the guinea, the chicken, and the
pigeon, four (five with the accessory) instead of eight.

While in the guinea-chicken hybrids the secondary spermato-
cytes tend to exhibit four (or five) chromosomes in division, there
is more irregularity than in the normal fowl. If my interpreta-
tion that the fusion between guinea and chicken chromosomes
in the primary spermatocytes is inhibited in some way because of
their inherent dissimilarities be correct, the same fact might
account likewise for the occasional increase of numbers in the

®° The germ-cell and the results of Mendel. Cincinnati Lancet-Clinic, May 9,
1903.

JOURNAL OF MORPHOLOGY, VOL. 23, No. 1

52 MICHAEL F. GUYER

second spermatocytes, inasmuch as they also would likely each
contain chromosomes of different parentage.

By far the greatest number of division stages to be seen in the
hybrid testis were those of the first spermatocytes. These were
comparatively plentiful and to my surprise were much more
normal in appearance than corresponding division stages in
pigeon hybrids from more closely related species. The multi-
polar spindles so characteristic of the first spermatocytes in hybrid
pigeons were very seldom encountered in the guinea-chicken fowls.

Perhaps the most interesting feature of this first maiotie divi-
sion was the appearance of the accessory chromosome or X-ele-
ment. This chromosome, whenever favorably located for obser-
vation, was found invariably to be of the guinea and not of the
chicken type.

It will be recalled that in various species of invertebrates the
X-element of the male is now known to be represented in the female
by two such elements in all cells with the diploid number of
chromosomes. In such females, after the reduction divisions
each egg thus comes to have a single X-element whereas, since
there was only one such element in the somatic and early germ-
cells of the male, and inasmuch as this body does not divide in
one of the maturation divisions of the cell but goes entire to one
of the daughter cells, the X-element is lacking in half of the sper-
matozoa.

In all known cases where an X-element exists in the male,
the eggs fertilized by a spermatozoon without the X-element are
the ones that give rise to the new males, hence the subsequent
X-element of one of these new individuals can only be one which
was originally in the egg. The fact that the male zygote must
always reeeive the X-element from the mother was pointed out
by Wilson’ in 1906. In the present instance, since it was the
mother of the hybrid that was the guinea, the X-element of
the hybrid should be of the guinea type, and such, in fact, was
found to be the case.

The chicken and the guinea types of X-element are readily |
distinguishable, that of the chicken being typically of larger

7 Wilson, E. B.: Studies on chromosomes, III. Jour. Exp. Zool., vi, 1906.

im

TESTES OF GUINEA-CHICKEN HYBRIDS 53
size, of stouter build, and of different shape. While each usually
appears as a curved body, the chicken X-element (figs. 5, 11, 12,
13, 14) is more curved than the other, having a U-shape with
both ends of the loop of the same size, while the guinea X-element
(figs. 6, 15, 16, 17, 18) is more comma- or pistol-shaped with
one end noticeably narrower than the other. While there may
be greater or less deviation from these types, generally in the
nature of unusual elongation or compression, on the whole, after
the observer’s eye has become accustomed to the elements in ques-
tion, he has little difficulty in readily identifying the two types.
The X-element of the hybrid (figs. 7, 19, 20, 21, 22, 23) is clearly
of the guinea type. In both the chicken and the guinea the X-ele-
ment, instead of having its more chacteristic appeararance, may
oecur occasionally as two closely apposed spherical chromosomes.
Under such circumstances the two components are of approxi-
mately the same size in the chicken, whereas one is always notice-
ably smaller than the other in the guinea. The same modification
may obtain in the hybrid (fig. 22), but here again the double
element is of the guinea type.

While it is very difficult to secure representative appearances of
the X-element of these fowl by photography because of differ-
ences in the focal plane of its different parts, still the pictures
obtained’ are sufficiently clear to demonstrate the point in ques-
tion. Fig. 5 (5a magnified 750, and 5b, 1500 diameters) is a
photograph of the X-element of the langshan cock. In the photo-
graph the typical U-shaped element appears to be more of a V,
but this is due to the fact that when the extremities of the chro-
mosome were in focus as they are in the picture, the bend of the U
was below the plane of focus and thus made to appear sharp-angled.
Fig. 6 shows in the guinea the metaphase of a dividing first sper-
matocyte viewed from one pole. The X-element at the top of
the field, is plainly seen to be narrower at one end. The focus
was such that its curved condition is not visible in the photograph.
Fig. 7 shows a dividing first spermatocyte of the hybrid, viewed
from one pole. What appears to be a long curved body at the

8 The writer makes grateful acknowledgment to Dr. Charles Goosmann for the
microphotographs of Plate I.

54 MICHAEL F. GUYER

top consists really of two chromosomes; one, at the base, a deeply-
staining, rounded, ordinary chromosome, and the other the
curved X-element. In order to show the curve of the latter the
camera had to be so focussed as to blend the two images. The
X-element of the hybrid, like that of the guinea, is seen to be
narrower at one end.

In my paper on the spermatogenesis of the chicken,’ I noted
the fact that it was not uncommon to find what appeared to be a
tripartite accessory, but I am now inclined to believe that when
such a body exists it is the X-element plus one member of a char-
acteristic pair of small chromosomes, the other member of which
passes to the opposite pole, either slightly in advance or at the
time of the regular division of the ordinary chromosomes. Figs.
7, 9, 11 and 12 of that paper? show evidence of this condition.
In fig. 11 the small element is completely detached. Fig. 10
probably represents a condition in which the opposite member of
this small pair stands apart from the equatorial plate towards
one pole, the accessory being seen on the other side of the equa-
torial plate. In the guinea the corresponding pair of small chro-
mosomes (see figs. 11 and 12 of my paper?’ on the spermatogenesis
of the guinea) is considerably smaller than in the chicken. The
one which passes to the pole also reached by the X-element shows
less tendency to unite with the latter than in the chicken, hence
itis only rarely that a tripartite condition of these bodies is ob-
served in the guinea.

In a paper! written in 1909 before I had discovered the pres-
ence of an X-element in the common fowl, I suggested that if
we assume, as some investigators have done, that increased nutri-
tion favors the production of females, diminished nutrition, the
production of males, then the excess of males among hybrid birds
might be due to the fact that in hybrids, ‘‘there would in all proba-
bility be more or less default in the metabolic processes because of
the incompatibilities which must necessarily exist between two

® Guyer, M. F.: The spermatogenesis of the domestic chicken (Gallus gallus
dom.). Anat. Anz., xxxiv, 22-24, 1909.

10 Guyer, M. F.: The spermatogenesis of the domestic guinea (Numida mele-
agrisdom.). Anat. Anz., xxxiv, 20-21, 1909.

4 Guyer, M. F.: On the sex of hybrid birds. Biol. Bul., xvi, 4; 1909.

TESTES OF GUINEA-CHICKEN HYBRIDS HY)

germ-plasms so dissimilar.”” The discovery of an X-element, how-
ever, together with the knowledge that it is of maternal instead of
paternal origin possibly gives us a simpler and more plausible
explanation. It is the spermatozoon without the large X-element
which unites with the egg in the production of the new male, and
since such a spermatozoon is much smaller than one of the other
type, the whole question may resolve itself into a mere matter
of the relative sizes of the spermatozoa. For inasmuch as such
hybrids are obtained with difficulty even under the most favor-
able conditions we may reasonably suppose that the egg-plasm
is more or less resistant or antagonistic to the entrance of a foreign
sperm, and that because of this the smaller type of spermatozoon
enters more readily, with the result that a male is produced.

SUMMARY

1. With one exception, where one testis was greatly hyper-
trophied, the testes of the four guinea-chicken hybrids examined
were of normal size.

2. Microscopic examination showed them to be abnormal,
however. No spermatozoa were developed and the seminiferous
tubules were few in number in most regions of the testis and often
contained disintegrating and defective cells.

3. As in hybrid pigeons the critical point seemed to be the
synaptic phase, the chromosomes of different parentage seemingly
being unable to unite normally in many instances.

4. In spite of this difficulty, however, not a few first sperma-
tocytes succeeded in passing through synapsis and subsequent
division with more or less of an appearance of normality.

5. An accessory chromosome or X-element of the guinea
(maternal species) type is present.

6. The X-element is of large size in the common fowl and con-
sequently the mature spermatozoa without it are much smaller
than the ones which bear it. It is suggested that inasmuch as
males are produced only from eggs fertilized by a spermatozoon
without the X-element, the great preponderance of males among
such hybrid offspring may be due to the simple fact that the
smaller type of spermatozoon can more readily penetrate a foreign,
and hence more or less incompatible, egg-plasm.

PLATE 1
EXPLANATION OF FIGURES

1 Section of a testis of a guinea-chicken hybrid showing the narrow, atrophied
seminiferous tubules separated by wide areas of interstitial substance. x 75.

2 Section of a testis of a guinea-chicken hybrid showing an island-like mass of
seminiferous tubules, normal in size, in a broad field of connective tissue and
stroma cells. > 75.

3 Section showing an area of the testis in one of the guinea-chicken hybrids in
which, in places, the seminiferous tubules were almost as plentiful as in normal
non-hybrid individuals. X 75.

4 Section of a testis of a guinea-chicken hybrid showing aninter-tubular area
in which numerous cells were in various stages of maturation. Primary spermato-
cytes in process of division were not uncommon. X 75.

5 Metaphase of a dividing first spermatocyte of the langshan cock, fig. 54
being at a magnification of 750, and 5b, 1500 diameters. The typical U-shaped
element appears to be more of a V in the photograph, but this is due to the fact
that when the ends of the chromosome were in focus as they are in the photograph,
the bend of the U was depressed below the plane of focus, giving to the picture a
sharp-angled appearance which the real object did not possess.

6 Metaphase of a dividing first spermatocyte of the guinea, viewed from one
pole. The X-element is at the top and is plainly seen to be narrower at one end
than at the other. It lay in such a position that its curved condition could not be
shown by photography. 1500. ‘*

7 Metaphase of a dividing first spermatocyte of the hybrid, viewed from one
pole. What appears to be along curved body at the top consists really of two chro-
mosomes; one, at the base, a deeply-staining, rounded, ordinary chromosome, and
the other the curved X-element. To show the curve of the latter the camera had
to be so focussed as to blend the two images. Like that of the guinea, the X-ele-
ment of the hybrid is seen to be narrower at oneend. X 1500.

8 Side view of a first spermatocyte in the hybrid showing the cell ready for divi-
sion with the X-element lying just above the level of the regular equatorial plate
of chromosomes. The photograph does not reveal the curved shape of the ele-
ment although this could readily be detected by manipulation of the fine adjust-
ment of the microscope. 1500.

56

1

PLATE

TESTES OF GUINEA-CHICKEN HYBRIDS

MICHAEL F, GUYER

JOURNAL OF MORPHOLOGY. VOL. 23, NO. 1

57

PLATE 2
EXPLANATION OF FIGURES

All of the drawings in this plate were made with the aid of a camera lucida
although in most cases, in order to show anything of the curved nature of the X-
elements or the full field of chromosomes, the focal plane had to be shifted, so that
most of the drawings from fig. 11 to fig. 22 represent composites of two planes of
focus. The magnification is in every case approximately 1800 diameters.

9 First spermatocytes of one of the hybrids showing the prevalent contraction
phase at which the maturation process commonly came to a halt.

10 A stroma cell from the testis of one of the hybrids.

11, 12, 18, 14 Drawings showing the characteristic U-shape of the X-element
of the langshan cock. Figs. 11, 12 and 13 represent division stages of first sper-
matocytes. Fig. 14 is from a smear preparation and shows the X-element at the
equator of the spindle in a secondary spermatocyte ready for division. {t divides
lengthwise at this time.

15, 16,17, 18 Drawings showing the characteristic comma- or pistol-shaped X-
element of the guinea. Fig. 17 is viewed from one pole.

19, 20, 21, 22,23 Drawings to show the X-element of the guinea-chicken hybrid.
it will be observed that the X-element is of the guinea type. Fig. 20 is viewed
from one pole.

58

PESTES OF GUINEA-CHICKEN HYBRIDS PLATE 2
MICHAEL IF. GUYER

| a
19
4
15 &.
“Se & we 4g S

le 16 by

1 3 ee

18

I4

59

THE EMBRYOLOGY OF CRYPTOBRANCHUS ALLE-
GHENIENSIS, INCLUDING COMPARISONS
WITH SOME OTHER VERTEBRATES

I. INTRODUCTION, THE HISTORY OF THE EGG BEFORE
CLEAVAGE

BERTRAM G. SMITH

From the Zoological Laboratory of Columbia University

FIFTY-SIX FIGURES

CONTENTS
ch, LARALIN@TG NEY CTHIGYD SER. hea Rg ee rr ee at Aa Sy 62
II. The adults.
ke LELSU OWES RRAN ES Sg) Sn 8 9 ee ed an ae 66
By General characteristics: size, form, coloration!...:............ fe 0G
C. Breeding habits
ep rced ia, SCAR ONS es. Ha. ls 4 wns oe Ae AE ace 68
Ze xternal sexual characteristics. ... ii:t45 <2 aossnot.. 2 aod ask 69
SeOeK Taio and sex segreration:..:.; csc eaewn ccs cade. ce 69

4. The eggs
(a). General history of the eggs and their envelopes before

vhetimeoispawning...:......°., cae msasmaticl cola Sehralue 71
(nh) Ovipositien, and nesting, habifiseeges..°:..4 ... 640.00 73
(c). The newly-laid egg and its envelopes................. 74
Sem ihen se rrMmtbOZOOBE 6 tf 00... sis SRE ate ee. 81
Ge neancthod af fertilization... ...).. Gee eee. eee 82
PR CREO O CEM OUST 3 6 a .e tle «x atv See ee wade kad hd. ie 83
ID), SUR ee ees (lke ee 88
III. Methods and technique.

Act heeollcction and’care of living material... 4.7... <0)s... 0.22508: 89
Barr Geta OEGSOnVAGlON:) ...\. .... «ise < cae ioe, lelatic sc oui ee 90
Ge eT SANTI, ew. «acs: Sis se Ro ke eS, scans A 94

IV. The external history of the egg before cleavage.
A. External changes preceding and accompanying maturation........ 96
B. Capacity of uterine eggs for fertilization................0......-. 100
C. Changes visible from the surface during fertilization.............. 101
1D YSU IS Ce oo er ee et Se a i rr 106

62 BERTRAM G. SMITH

V. The internal history of the egg before cleavage.

Are OVOP CTESIS 6 {0 5:235.2 Sah ei eee eee ERC Se Lae ee eee: 107
1. The formation of the follicle and the egg membranes.......... 108

2. The establishment of polarity, and the progress of axial differ-
@WCLALL OW re temp ee ke ger fos hoo ns ci oA acghsked cee 118

3. Resorption of ovocytes; the follicle cells in a phagocytic réle.. 126
4. Organization of the ovocyte shortly before the appearance of

the germinal vesicle at the surface.............ces.00.e008 128
B. Maturation.
1, Eheygerminalivesicleat theisurfaceseess.ch. sees ee eee oe 129
2. The dissolution of the germinal vesicle, and the formation of
fhesirst polar spindles o2n.2..e tian. eee oak at ae eee eae ee 130
a. Dhemecona polar spindles: ees ee ee ee eee 134

4. The organization of the egg immediately before fertilization... 137
C. Fertilization.

i eeistory Of the €eg-nucleus's = .tat am cote a ate o ke ee 138
2. History of the sperm-nucleus.

(a). Penetration of the egg by the spermatozoon........... 140

(b). Polyspermy, and the fate of the supernumerary sper-
IMATOZOM RS Scioto sien cts +o we aun eochies Bee ete Ar eae 145

3. Union of the germ-nuclei, and the formation of the first polar
SPU es Ne SMe lec eee oo cay Otc seit a set ee eek 146
4. (Changes imtthe’ blastodisc 7, soccer ee ete toe eee 147
Ds SS ani yin er ye ee eer on hs Ps oe aa eu ey ee 2.2 inte eerie ge gee ere Ste 148
BTM ORTAD DY: oko nesses sehen cic sic Ra een Ce ae ee = cree 151

I. INTRODUCTION

For more than a generation zoologists have eagerly sought
for the embryological material of the hellbender, Cryptobranchus
allegheniensis Daudin. Until quite recently these efforts have
been conspicuously lacking in success. It seems remarkable that
the life history of an animal so large, so abundant in localities
easy of access, and so important from a phylogenetic point of
view, should so long remain shrouded in mystery. But the same
difficulty has been encountered in attempts to work out the

natural history of several nearly related forms. Eycleshymer
(06) says:

After years of persistent and patient effort Professor Whitman finally
discovered the nests and eggs of Necturus. Only those who have for
years been baffled in their attempts to obtain the embryological material
of other North American urodeles, such as Siren, Amphiuma, and Cryp-
tobranchus can properly appreciate the enormity of the task.

EMBRYOLOGY OF CRYPTOBRANCHUS 63

In the case of Cryptobranchus the difficulty in finding embryo-
logical material seems to have been-enhanced by the unusual
breeding season of the animal; the eggs are laid in the fall, while
most amphibia spawn in the spring. Townsend (’82) published
a general description of some fertilized eggs which he states were
deposited in August. McGregor (’96) described very briefly an
embryo 16 mm. in length, and (’99) stated that the eggs are depos-
ited in August andSeptember. Yet the information thus acquired
in regard to the time of spawning seems not to have become gener-
ally known to others who were searching for the eggs. A sugges-
tion might have been obtained from Sasaki’s (’87) observation
that the Japanese ‘giant salamander,’ Cryptobranchus japoni-
cus (Megalobatrachus maximus Schlegel), deposits its eggs in
August; but this also seems to have been overlooked. Reese
(04) sueceeded in obtaining some unfertilized eggs, of which he
gave the first detailed description.

The embryological record for Cryptobranchus allegheniensis
remained almost a blank until 1906, when I published a prelimi-
nary report containing, besides a description of the sexual ele-
ments, a brief account of the external development during the
cleavage stages. A later contribution (Smith, ’07),devoted chiefly
to the habits, more particularly the breeding habits, included a
very general account of the life history.

From a phylogenetic point of view great interest attaches to
the amphibia; there is no doubt that they lie close to the extinct
ancestral stock of the highest forms of vertebrate life. Concern-
ing the origin of the amphibia themselves Kingsley (’99) says:
‘All the facts of structure and development go to show that the
amphibia have arisen from the crossopterygian ganoids, and that
existing groups have descended from the stegocephali, each by
its own line of ancestry.” But when we inquire further, and
attempt to trace more particularly the origin of any group of exist-
ing amphibia from an extinct form exhibiting affinities to the
crossopterygii, we are landed at once in the midst of uncertain-
ties. Confining our attention to the urodeles, we are confronted
with the difficult question of the phylogenetic relationships of
the different members of this group. The problem will be more

64 BERTRAM G. SMITH

fully discussed in a later section; for the present it will be suffi-
cient to call attention to one of its leading aspects. From exist-
ing urodeles we may select a series of forms illustrating all stages
in a transition from an aquatic to a terrestrial mode of life, or
vice versa. In which direction should the series be read? Or
have we stated the question incorrectly, and have the urodeles
reached their present condition, some from an aquatic, some from
a terrestrial ancestry?

In studying this aspect of the phylogenetic problem our atten-
tion cannot fail to be attracted by Cryptobranchus. For here
we have a urodele whose entire life is spent in the water, charac-
terized by persistent gill slits, the most primitive brain (Osborn,
88), and external fertilization (Smith, ’07). On the other hand
Cryptobranchus is known to possess. deciduous external gills,
functional lungs, and a method of locomotion by crawling on the
bottom which suggests a former terrestrial habit. Is Crypto-
branchus primitively aquatic, or does it come down to us bearing
evidence of a former land-living existence? An answer to this
question would go far in advancing our knowledge of the phylog-
eny of the entire group.

In the solution of our phylogenetic problem comparative anat-
omy, paleontology and embryology must work together. It is
the embryological evidence that has hitherto been most conspicu-
ously lacking. Notwithstanding the important position of the
aquatic urodeles, it is here that we find one of the widest gaps in
our knowledge of comparative embryology. Not only has the
development of Cryptobranchus allegheniensis remained unde-
scribed, but little or nothing is known concerning the embryology
of most of its near relatives. Very recently, it is true, consider-
able progress has been made in the study of the embryology of
Cryptobranchus japonicus, but part of this work was done on very
scanty material, and the field is by no means exhausted. Of the
development of Amphiuma and Siren practically nothingis known.
Some results have been obtained with special problems in the
development of Necturus, but the life history has not been covered
in a comprehensive manner. For a study of the phylogenetic
relations of these forms a knowledge of the development in its

EMBRYOLOGY OF CRYPTOBRANCHUS 65

general aspects as well as along special lines is imperative; and in
no other form do the embryological data promise to shed greater
light on phylogenetic problems than in the case of Cryptobranchus.

For the analysis of developmental processes from a morpho-
genetic point of view the eggs of Cryptobranchus present certain
favorable features. One of these is the presence of a larger
amount of yolk than is known in the egg of any other amphibian;
they are thus favorable objects for the study of the influence of
yolk on development. The eggs, moreover, are lacking in pig-
ment, and the early segregation of the yolk gives a translucency
to parts of the embryo even in the gastrula stage, enabling one to
study satisfactorily in the living egg some of the internal features
of development. ‘The embryo is found to respond admirably to
the influence of chemicals modifying the course of development;
for certain experiments of this sort it gives results decidedly
more definite than have been obtained with the embryo of the
frog. :

The present contribution is to be followed by other parts deal-
ing with the embryonic and larval development.

This investigation has been pursued under a great variety of
circumstances, and with many protracted interruptions due to
the pressure of other work. Field work on the habits of Crypto-
branchus, the collection and preservation of material, and the
study of the living egg, have been carried on each autumn (’05—
’11 inclusive) in northwestern Pennsylvania. For comparison
with Cryptobranchus, I have collected embryological material
of Necturus during the seasons of 1910 and 1911, from Lake
Monona, Wisconsin. Laboratory work, principally on preserved
material, begun in the Zoological Laboratory of the University
of Michigan (’05-07), has been continued in the zoological
laboratories of Lake Forest College (’07), Syracuse University
(08-09), the Bureau of Fisheries at Woods Hole (summer of
1908), the University of Wisconsin (’09-11), and Columbia Uni-
versity (11-12). To the directors of these respective laboratories
I wish to express my sincere thanks for uniform courtesy in
placing the resources of each institution at my disposal. To
Professor Bashford Dean, under. whose direction the work is being

66 BERTRAM G. SMITH

carried on during the present year, I am profoundly indebted for
his constant encouragement, kindly criticism, and valuable advice;
for this I desire to record my grateful appreciation.

Il. THE ADULTS
A. HABITAT

Cryptobranchus allegheniensis was found abundantly in the
Brokenstraw Creek, a tributary to the Allegheny River, in north-
western Pennsylvania. The most favorable locality extends
from the confluence with the Allegheny five or six miles up-
stream. The stream has a rather rapid descent, and a gravelly
or rocky bottom. Shallow and rocky rapids make up the greater
part of its course, alternating with areas of deeper and more quiet
water.

As a rule, Cryptobranchus is found more abundantly in rather
shallow and rapid water, where large flat rocks afford suitable
cover. Usually the animals lie concealed in cavities under these
rocks. As more than one individual is seldom found under a
single rock, we conclude that its life is in general a solitary one.
Cryptobranchus rarely comes out of its hiding place in the day-
time, except in the spring and early summer and during the breed-
ing season (the first two weeks of September). At night they
venture abroad in large numbers; they are then seen by fishermen
spearing by torchlight, who commonly report them in the deeper
and more quiet water.

The cavity or cavern used for a more orless permanent dwelling-
place has a rock for its roof and the gravelly bed of the stream for
its floor. In perhaps the majority of cases, ready-made cavities
are chosen as homes, and these are reached by a natural opening.
But the cavity sometimes bears evidence of having been in part
hollowed out by the animal, and is occasionally reached by a single
tunnel-like entrance on the down-stream side of the rock; this is
more often the case in cavities used for spawning purposes.

There is a striking similarity between the habitat of Crypto-
branchus allegheniensis and that of the ‘giant salmander’ of
Japan as described and figured by Ishikawa (’04).

EMBRYOLOGY OF CRYPTOBRANCHUS 67

B. GENERAL CHARACTERISTICS
1. Size

Out of the many hundreds of adults captured, the largest male
(September 3, 08) measured 60 em. (234 inches) long and weighed
21 pounds. The largest female captured (September 3, ’09)
weighed exactly 3 pounds. The latter animal unfortunately
escaped from the aquarium in which it was confined and was not
measured; probably it was no longer than the longest male, but
heavier because distended with eggs. Professor McGregor
reports a specimen 25 inches long, taken from the Scioto River.

The great majority of specimens captured by me were much
smaller; specimens of about 30 to 50 em. were most frequently
taken. The smallest sexually mature male measured 30 cm.;
the smallest mature female 35 em.

2. Form

As compared with the young, the adult is more flattened dorso-
ventrally—an adaptation to life in shallow crevices. The head
particularly shows this flattening: it is wedge-shaped as viewed
from the side, a form which enables the animal to force its soft
body into very shallow crevices.

Moreover, as compared with the young, the adult is distin-
guished by a general looseness and wrinkling of the skin at the
sides of the body, forming broad lateral horizontal folds; and by
similar flaps of skin on the posterior sides of the limbs. During
locomotion these folds and flaps undulate in the water, contrib-
uting to the uncouth appearance of the animal.

3. Coloration

Young sexually mature individuals vary little in color or color
pattern. The ground color is dull brown, with conspicuous black
spots and less conspicuous yellow spots scattered over the dorsal
and lateral surfaces. Both kinds of spots are irregular in size and
form. The coloration of young adults is practically that of

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1

68 BERTRAM G. SMITH

immature specimens from 16 em. body length upwards; in these
stages the spots are more conspicuous than in the larvae or the
older adults. In the older, full-grown specimens the general color
effect may vary in two ways: it may become either greenish-brown
or decidedly reddish brown. As stated by Reese (’03) these vari-
ations in color occur in both sexes.

C. BREEDING HABITS

1. Breeding season

The following data indicate the beginning of the breeding
season, as shown by the deposition of eggs, in northwestern
Pennsylvania during a series of years:

IQS ood Scola hon coce aeree: August 30 UG OG) cas comeste Pree Peete he August 29
IGE) =. 3 Sistod ete: okcreaeeeoee September 8 ADO): Sei ea eee September 1
GO SMrrrerter tetas ane ss August 28 Mi beat Coctcse eke eee September 4

The summer of 1907 was an unusually ‘late season’ as regards
vegetation as well as the breeding season of Cryptobranchus.
This indicates the probability that climatic conditions influence
the time of spawning.

Egg-laying continues for a period of about two weeks. At
the end of this time females have in a few instances been taken
with the full complement of ripe eggs still in the ovary and show-
ing signs of degeneration.

The occurrence of the breeding season of Cryptobranchus in
the fall is in marked contrast to the habits of nearly all other uro-
deles. Some other urodeles which have a late breeding season are
C. japonicus, which according to several authors (Sasaki, °87;
Kerbert, ’04; Ishikawa, ’04; de Bussy ’04 and ’05) lays its eggs
during the latter part of August and the early part of September;
and Amphiuma, which according to McGregor (’99) breeds in
midsummer. Among the anura, Scaphiopus holbrooki spawns
during the summer, the time varying from June to August (Pike,
86; Hargitt, ’88).

EMBRYOLOGY OF CRYPTOBRANCHUS 69

2. External sexual characteristics

The adult male may be recognized (Reese, ’04) by the presence
of a swollen ring about the cloaca, due to glands beneath the skin.
This swelling is quite prominent during, and for a few weeks before
the breeding season. I found it difficult to distinguish by exter-
nal characteristics the sexes of a few specimens taken during the
first week of July; during the latter part of July the males could
easily be distinguished by the presence of the cloacal swelling.
In a few males obtained and examined during the early part of
November, the swelling was less pronounced than is usually the
case during the breeding season. Females are characterized by
the entire absence of the cloacal protuberance found in the male;
also, the abdomen of the gravid female is slightly swollen.

3. Sex ratio and sex segregation

As a general rule, fewer females than males have been captured.
The record of the sex of the great number of adults captured dur-
ing the progress of the work is not complete, but the conclusion
reached by later work is that the original ratio of 1:8 determined
(Smith ’07) during the fall of 1906 is much too large. In a series
of years the proportion of females to males captured is about 1:2
or 1:3. These results are of course not conclusive as to the actual
sex ratio; as will presently be explained, the sex ratio in the speci-
mens captured varies for different times and places, and the true
ratio may be disguised by the occurrence of seasonal segregation
of the females from the more accessible localities.

In studying the distribution of the sexes throughout the year
a distinction must be made between localities which experience
has shown are chosen as breeding grounds, and other localities
unsuited for breeding purposes. The breeding grounds are char-
acterized by shallow water, a moderate current, and the pres-
ence of large flat rocks affording cover for cavities protected from
the current. Elsewhere a swifter current, smaller rocks barely
large enough to serve as cover, or deeper pools of quiet water,
afford conditions in which Cryptobranchus can live, but which
are not adapted for purposes of reproduction.

70 BERTRAM G. SMITH

Studies of the sex ratio indicate a more or less perfect segrega-
tion of the sexes at certain seasons of the year. A dozen adults
captured in June on the breeding grounds, by an assistant,
proved to be all males. During the summer, search of the breed-
ing grounds results in the capture of a few females and a much
larger number of males; in localities unsuited for breeding one
is more likely to find females, and males are seldom found in
their immediate vicinity. Just before the breeding season one
is more likely to find females on the breeding grounds, but the
males are still considerably in excess, and there is apparently
a tendency for the sexes to occur in groups: within a restricted
area one may find only males, while within another area a short
distance away one may find only females. At the height of the
breeding season, both sexes are found on the breeding grounds in
more nearly equal numbers.

For some days or weeks after the close of the breeding season
the male remains in possession of the nest; females have never
been found in nests containing eggs. At this time females have
been found in considerable numbers in localities unsuitable for
breeding, with no males in their immediate vicinity.

The general results of the studies of the sex ratio and the dis-
tribution of the sexes indicate that the males abound in localities
suitable for breeding, throughout the year, and that they are less
numerous elsewhere; it is positively established that the males
alone are in possession of the nests after spawning takes place;
and it is probable that there is a more or less perfect segregation
of the females from the breeding grounds during a period extend-
ing from the close of the breeding season until the middle of the
following summer.

In Necturus, segregation of the sexes at a certain season of the
year seems to be more complete than is ever the case with Crypto-
branchus allegheniensis. Eycleshymer (’06) says:

In the autumn they are found in pairs or small groups. From this
fact and others to be recorded later it is inferred that this is the mating
season... . During egg-laying [in the spring] the males
are never found with the females, and where they remain is unknown.

EMBRYOLOGY OF CRYPTOBRANCHUS (igi

4. The eggs

(a). General history of the eggs and their envelopes before the time
of spawning. In an adult female of average size about 450
eggs are matured each season—225 from each ovary. In general
the number is greater in the larger and presumably older females
than in the smaller ones. At the approach of the breeding sea-
son the eggs which are about to become mature are readily dis-
tinguishable from the others by their much greater size and yolk
content. The liberation of these eggs from the ovary and their
passage down the oviduct takes place shortly before spawning.
The exact date varies considerably in different individuals; for
a week or ten days after the first cases of spawning, females may
be found with mature eggs all still in place in the ovaries. The
process of liberation of the eggs and their passage down the ovi-
duct, once begun, must be accomplished with considerable rapid-
ity; for out of more than a hundred females opened and examined
during the breeding season in the course of several years, only
four have been found in which the process was actually taking
place. This state of affairs is in marked contrast to the condi-
tion in Bufo, where according to King (’05) the great majority
of specimens collected soon after they had emerged from their
hibernation contained eggs free in the body cavity. In three
out of the four cases above mentioned for Cryptobranchus, eggs
were found along the entire route: some still in place in the ovary;
some free in the body cavity, for the most part collected at its
anterior end, near the opening of the oviduct; others forming a
procession down the oviduct; the remainder aggregated in the
uterus. In the fourth case, the ripening eggs were found only
in the body cavity, oviduct and uterus. The process takes place
on the two sides of the body simultaneously.

During their passage down the oviduct the eggs receive their
gelatinous outer envelopes, the product of the oviduct. At the
upper end of the oviduct, the eggs collect in masses; a little further
down, they are arranged in a solid row. In these parts of the
oviduct the covering is absent or just beginning; the eggs are very
soft, and elongated by pressure of the walls of the oviduct. In

ee BERTRAM G. SMITH

the middle and lower portions of the oviduct the eggs are distrib-
uted at fairly equal intervals; here the envelope is well formed,
and consists of a capsule about each egg, and a slender connect-
ing cord, giving a general resemblance to a string of beads.

After their descent through the oviduct, the eggs of each side
of the body form a single string aggregated in a much twisted and
tangled mass in the uterus. Considered as individuals without
regard to their sequence in the string, the eggs display a striking
regularity in their arrangment in the uterus, being packed in par-
allel spiral rows; but thisismerely the result of mechanical pressure,
as the string pursues a very sinuous and complicated course
throughout the mass.

The egg capsules at the end of the uterus nearest the cloaca,
hence those first formed, do not contain eggs; those nearest the
oviduct, hence the last formed, are likewise devoid of eggs. These
empty egg capsules are in general smaller than those that contain
eggs, with a regular gradation in size from those at the extremity
of the cord, which are scarcely more than a millimeter in diameter,
up to those nearest the egg-containing capsules, where the diam-
eter is only slightly less than normal. However small the size,
these capsules are always perfectly formed, with a central spherical
space; they are never solid. The ‘empty’ capsules contain a
small amount of coelomic fluid in which are distinguishable under
the microscope leucocytes, erythrocytes and yolk corpuscles. A
cloudy mass of fluid with the same constituents occurs in the upper
part of the uterus, outside of the egg envelopes. Similar cap-
sules devoid of eggs are the ‘wind eggs,’ known in various verte-
brates: birds, reptiles, sharks, chimaeroids.

As a result of experimental studies on the nature of the stimulus
which causes the shell to be formed about the hen’s egg, Pearl
(09) reached the following conclusions: (a) the stimulus which
sets the shell-secreting glands of the fowl’s oviduct into activity
is mechanical rather than chemical in its nature; (b) the for-
mation of a shell on the hen’s egg is brought about by a strictly
local reflex, and is not immediately dependent upon the activity
of other portions of the reproductive system (nervous impulse or
hormone formation). In this connection it is interesting to note

EMBRYOLOGY OF CRYPTOBRANCHUS ie

that in Cryptobranchus the mechanical stimulus can hardly be
the true cause of the formation of the capsule, since capsules are
formed when only a small drop of coelomic fluid is present. More-
over it is here observed that coelomic fluid may pass down the
oviduct without becoming enclosed in such capsules; on the other
hand, every egg is provided with a capsule.

When distended with eggs, the uterus is spindle-shaped, about
10 em. long, with a transverse diameter of about 4 em. at its wid-
est part. Its thin walls have a rich blood supply.

Apparently the eggs do not, as a rule, remain long in the uterus
before spawning takes place. During the breeding season com-
paratively few females are found having eggs in the uteri; the
majority of the females captured are either spent or with eggs
still in place in the ovaries. Eggs taken from the uteri are, in the
great majority of cases, capable of artificial fertilization; this
subject will be more fully discussed later.

(b). Oviposition, and nesting habits. Under strictly natural
conditions egg-laying takes place under cover of rocks in the bed
of the stream; but in creek aquaria, arranged to afford conditions
as natural as possible without too much cover, the process has
been repeatedly observed.

Egg-laying begins slowly, a short string of eggs sometimes
protruding from the cloaca for several hours before spawning
begins in earnest. In the natural habitat, such short strings of
eggs are often found in the open. Later, two long strings of eggs
proceed slowly from the cloaca, one from each uterus; the majority
of the eggs are then deposited more rapidly, in multiple strands,
the process requiring less than five minutes. When egg-laying
is completed, the strings are usually twisted together in a single
tangled mass.

The ‘nest’ of Cryptobranchus allegheniensis has already been
described as either a burrow or a natural cavity under a rock
which is wholly or partially submerged. The eggs are not fas-
tened in any way, but are protected by this sheltered position
from being swept away by the current.

The nests of Cryptobranchus japonicus have been described
by Ishikawa (’04), and closely resemble those of C. allegheniensis.

74 BERTRAM G. SMITH

Das Tier legte seine Hier in tiefe horizontal verlaufende Lécher, in
denen das Wasser sehr ruhig ist. Manchmal is solch ein Loch 10 oder
mehr Fuss tief und kaum fiir das Licht zugiinglich. Die Brutstellen
fiir die Eier sind aber nicht immer so tief. Oft fand ich Eier in einem
Loch nicht tiefer als 3 oder 4 Fuss. Oeffnet man ein solehes Loch, so
findet man eine abgerundete Stelle, deren Boden ganz rein gehalten ist.

The nesting habits of Necturus have been described by Eycle-
shymer (’06), and the writer (’'11). The eggs are attached singly
by their gelatinous envelopes to the under side of a rock, board,
or other object lying at the bottom of the water (figs. 55 and 56).

The eggs of Amphiuma found by Hay (’88 and ’90) in an Arkan-
sas Swamp were in a comparatively dry situation, in a small
excavation under a log several rods from the nearest water.

Brief reference to the nesting habits of some other amphibia has
been made in previous papers (Smith, ’06 and ’07). Very remark-
able are the nesting habits of the anuran Phyllomedusa, described
by Budgett (99); the eggs are deposited in a pocket made by
bringing together the edges of a leaf overhanging the water.

Amongst the dipnoi, the nest of Protopterus (Budgett, ’01 a
and ’01b) is an oval hole filled with water and surrounded by
swampy ground. The nest is at first entirely submerged, but by
the partial drying up of the swamp it is left as an isolated pool.
Lepidosiren (Kerr, ’00) nests in a veritable burrow excavated in
the black peaty soil of the swamp.

Nesting habits are well known in many teleosts, and in Amia
(Dean, ’96; Reighard’03). According to Budgett (’01 a) the crossop-
terygian Polypterus probably makes no nest, and certainly lays
but few eggs at a time, these being scattered broadcast through
the thick vegetation of the flooded grass lands. Comparison
with Cryptobranchus suggests the possibility that these scattered
eggs are but preliminary attempts at egg-laying.

(c). The newly-laid egg and its envelopes. (Figs. 54, 1 and 2.)
In eggs taken from the uterus, the outer egg envelope or capsule
fits closely about the egg proper; the envelopes are flaccid and
much wrinkled. The capsule of the newly-laid egg takes up water
rapidly; in the course of one or two hours a space, filled with
fluid, appears between the egg and its capsule, sufficient to enable
the egg to orient itself with the animal pole uppermost.

EMBRYOLOGY OF CRYPTOBRANCHUS 75

The egg proper is perfectly spherical when fresh, but it gradu-
ally becomes slightly oblate from the effects of gravity. It is
about the size of a pea, and bright yellow in color—a rather deep
yellow at the lower pole, grading to a very pale yellow at the upper.
The general intensity of the yellow color varies considerably
in eggs of different spawnings, but is quite uniform in eggs from
the same female. The absence of black pigment is probably
correlated with the fact that the eggs are laid in darkness: the

Fig. 1 Eggs and egg envelopes of Cryptobranchus allegheniensis, natural size.
op. b., opaque body; lam. z., lamellar zone of envelope.
! Fig. 2 Optical longitudinal section through the lamellar zone of the envelope
in the region of junction of the egg capsule with the connecting cord. X 13.
Fig. 3 Spermatozoon. X 500. m., middle piece.

76 BERTRAM G. SMITH

eggs of Necturus, Plethodon, Spelerpes and Desmognathus, which
are also laid under cover, are likewise unpigmented.

A very thin and transparent’ ‘vitelline membrane’—the zona
pellucida of the ovocyte—closely invests the egg; it is quite
inconspicuous in fresh material. This is not the true cell wall of
the egg, which, as described in detail on page 112 lies immediately
within the vitelline membrane and represents in a modified form
the zona radiata of the ovocyte.

Proceeding from within outward, the coverings of the egg may
be enumerated as follows: (a) the cell wall; (b) the vitelline mem-
brane lying in close contact with the preceding; and (c) the cap-
sule or thick gelatinous outer envelope, which is separated from
the vitelline membrane by a space filled with fluid.

During the first few hours after fertilization the capsule gradu-
ally becomes turgid by osmosis, becoming in this way a much
more efficient protection to the egg; the space between the egg
and its capsule is increased by the absorption of water and in
this the egg almost floats, resting lightly on the lower inner sur-
face of the capsule. When the eggs are removed from water the
egg proper looks much larger than it really is, because magnified
by the spherical capsule.

For a day or two the envelopes are quite soft and somewhat vis-
cous, making it rather difficult to cut them with scissors in order
to remove the eggs. Gradually the material of the envelopes
becomes firmer. The connecting cord is at first quite elastic, but
it loses this quality to a considerable extent after prolonged immer-
sion in water.

Until after the eggs have been in water for several days, the
outer layers of the envelopes are still cast into wavy folds or
wrinkles, usually extending spirally about the capsules and the
connecting cord. As a rule the spiral is constant in the direction
in which it extends about the axis of the string in all portions of
the cord and capsule. These spiral folds are usually most strongly
marked at the ends of the cord adjacent to the capsule, and here
they often persist (fig. 1), suggesting the chalazae of the hen’s

egg.

EMBRYOLOGY OF CRYPTOBRANCHUS 77

The envelope is perfectly transparent when fresh, except that
wherever viewed tangentially its inner layers have a misty appear-
ance, represented by the shaded zone in fig. 1, and due to a fine
Jamellar structure sketched in optical section in fig. 2. The misty
appearance is caused by the diffusion of light passing through
these concentric layers in a direction tangential to their surfaces.
The core or axis of the connecting cord has the same misty appear-
ance, due to a continuation of the lamellar structure. The various
layers or lamellae of the gelatinous envelope are in intimate
contact; there is no fluid-filled space between them such as occurs
between the capsule and the vitelline membrane.

The inner layer of the lamellar core of the cord in some cases
exhibits a marked twisted or spiral arrangement, like that of the
inner portion of the cord connecting the eggs of Ichthyophis as
described by the Sarasins (’87—’93).

The eggs of a given spawning are fairly uniform in size, but there
is considerable variation in the size of eggs from different parents.
The average dimensions of the living egg and its envelopes, after
two days’ immersion in water, are as follows:

Drameter ober POPE wor. teks oe... c case» 0 one aa ee ee hie 6.2 mm.
Drametenokerawaitmenvelopeeasc.s: ++... 050s daeee oe ene 18 mm.
Wirametverm or connectimeCord 2595 60h is sisi 2 Re ee 5 mm.
Distance of one egg from another, measured from center to center

ALON Pe CNCM ORC AD OUTS Bee ts 5 «<i css osaceer wee oer aera 30° mm.

A few egg capsules, particularly among the empty ones, are
double, formed by the union of two capsules without a connect-
ing cord. In such cases the cavities of the two capsules are usually
separated only by a thin gelatinous septum; but all gradations
occur between this condition and that in which two capsules are
connected by an unusually short cord. Rarely, three capsules
are closely approximated.

I have found a few instances in which two eggs occurred in
the cavity of one simple capsule, without any separation by a gela-
tinous membrane. It would seem possible that double embryos
might be formed in this way, by the fusion of the yolk masses of
two such eggs; but this could not account for the only double

78 BERTRAM G. SMITH

embryo that I have found in nature, for in this case, to be
described later, each embryo is half the normal size.

After fertilization, numerous spermatozoa are found imbedded
in the egg capsule, and floating in the liquid between the capsule
and the egg; they also occur in capsules that do not contain eggs.
The spermatozoa occur singly, not in masses, and they are entirely
absent from eggs taken from:the uterus. Fertilization occurs
only after the eggs have been deposited in the water (Smith, ’07).

An envelope so tough and thick as that of Cryptobranchus
must exert a decided selective power with regard to the spermato-
zoa; of a considerable number of spermatozoa simultaneously
coming in contact with the envelope, the most vigorous, as well as
the ones structurally best adapted, would succeed in first entering
the egg.

Floating in the liquid between each egg and its envelope, there
occurs a fairly large irregular and slightly opaque mass, in appear-
ance like a faint white cloud (see fig. 1; this mass is also faintly
shown in the photograph, fig. 54). Under the microscope it is
found to consist of a clear viscous matrix in which are imbedded
numerous leucocytes and occasionally a few erythrocytes. In
fertilized eggs, this mass, which I have called (’07) the ‘opaque
body’ sometimes contains spermatozoa, but they are not restricted
to it, nor especially numerous init. The opaque body is uniformly
present in eggs that do not contain spermatozoa.

A mass similar in general appearance and location to that de-
scribed above as the opaque body, is figured by Ishikawa (’04)
within the egg capsule of Cryptobranchus japonicus. In the text
he refers to these masses as ‘Samenhaufen,’ and speaks of the
presence of spermatozoa within the egg capsules as evidence of
internal fertilization. He considers it improbable that the sper-
matozoa are able to penetrate the egg capsule, and supposes that
they are taken up into the oviduct before the egg capsules are
formed.

DeBussy (’04, p. 11) found a mass (‘vlokje’) of similar appear-
ance within the capsules of unfertilized eggs of C. japonicus.
Under the microscope he found the mass to consist of a slimy sub-
stance containing red blood corpuscles and yolk granules, but

EMBRYOLOGY OF CRYPTOBRANCHUS 79

no spermatozoa. He concludes that the presence of spermato-
zoa is not essential to the formation of the mass, but that they may
merely form an element of it; hence that the name ‘Samenhaufen’
is scarcely justified. This conclusion is in essential agreement
with my results on C. allegheniensis; it seems therefore that the
masses called ‘Samenhaufen’ in C. japonicus by Ishikawa are of
the same nature as the ‘opaque bodies’ of C. allegheniensis, and
like them of no significance in fertilization. The opaque body
apparently consists of coelomic fluid that has escaped into the
oviduct.

The egg strings of Cryptobranchus japonicus as described by
Ishikawa (’04) and deBussy (’04) closely resemble in structure
those of C. allegheniensis. Both eggs and capsules of the Japan-
ese form are slightly larger; according to Ishikawa the egg proper
is about 7 mm. in diameter, and the capsule varies from 20 to 25
mm. in diameter in different spawnings.

The egg capsules of Amphiuma as described by Hay (’88 and
90) have the same general structure as those of Cryptobranchus.
For an opportunity to examine one of Hay’s specimens of the
embryological material of Amphiuma, I am indebted to Prof.
C. W. Hargitt, to whom the specimen had been presented by the
finder. The egg capsule has a glistening surface like isinglass;
it is thinner and apparently tougher, and the connecting cord
more slender, than in Cryptobranchus. These peculiarities may
be due in part to preservation in alcohol, which tends to produce
the same condition in the envelopes of Cryptobranchus; but my
impression is that the ege capsules of Amphiuma are better
adapted to retain moisture when exposed to the air.

Other amphibians whose egg capsules are fastened together
like a string of beads are Alytes, Ichthyophis, and Hypogeophis
(Brauer 797).

The general appearance of the egg capsules of Necturus is
shown in figs. 55 and 56; some further details of structure are
shown in fig. 4. There are three layers to the gelatinous envelope:
(a) a comparatively thin but very dense inner layer, consisting
of several lamellae; (b) a thicker median layer of moderate density,
consisting of many lamellae; and (c) a very thick outer layer of

80 BERTRAM G. SMITH

homogeneous material, much less dense than either of the pre-
ceding. ‘This outer layer is produced to form the stalk by which
the capsule is attached to some solid object. As seen in optical
section, the lamellae of the two inner layers have a somewhat
sinuous or wavy outline. Leaving the stalk out of account, the
entire structure bears a close resemblance to the gelatinous envel-
opes of the frog’s egg. In the early stages of development of the
embryo, the dense inner layer of the capsule fits so closely that it
is not clearly differentiated from the embryo; this layer is best

Fig. 4 Optical section through an egg capsule, and surface view of an embryo,
of Necturus. The embryo is shown in a stage with neural folds, when the capsule
is slightly separated from it by a space filled with water. X 4.

studied after the embryo has passed the gastrula stage, when a
narrow space, filled with liquid, appears between the embryo and
its capsule (see fig. 4). This space is not strictly homologous with
the similar space that appears much earlier in the egg of Cryp-
tobranchus; for in Cryptobranchus the space appears between
the gelatinous envelope and the vitelline membrane (zona pellu-
cida of the ovarian egg) which remains in close contact with the
egg, while in Necturus the vitelline membrane apparently func-

EMBRYOLOGY OF CRYPTOBRANCHUS 81

tions as the innermost lamella of the capsule. The entire inner
layer of the capsule of Necturus has a tough consistency similar
to that of the vitelline membrane of Cryptobranchus; this per-
haps is the reason why it is so slow in enlarging.

In Necturus the egg proper is slightly smaller than that of
Cryptobranchus; in the early cleavage stages it measures about
5.8 mm. in diameter.

A general description of the gelatinous envelopes of several
species of Amblystoma has been given in a previous paper (Smith
11).

5. The sperm

The spermatozoon: (fig. 3) has been figured by Reese (’04),
who fails, however, to picture the middle-piece. The sperma-
tozoon is about 225u long, and stout in structure as compared
with the spermatozoa of Amblystoma and Diemyctylus. The
head, excepting the acrosome, stains deeply with Delafield’s
haematoxylin. The acrosome appears to be uniformly tapering,
not spear-shaped as in Amphiuma (described by McGregor,
99). As stated by MeGregor (99) the middle-piece in Crypto-
. branchus is very short*in comparison with that of other urodeles.
The tail-piece is provided with an undulating membrane, bordered
with a convoluted filament.

The ripe spermatozoon is motile, as regards both shaft and
filament; but the spermatozoon as a whole is not so flexible as the
more slender spermatozoa of Amblystoma and Diemyctylus.
This greater rigidity of the spermatozoon is perhaps correlated
with the method of fertilization: the spermatozoon must pene-
trate the tough and thick egg capsule after a brief exposure of the
latter to the hardening effects of water.

In seminal fluid obtained from occasional individuals, an oval
mass of granular protoplasm, about 134 by 16y, surrounds the
posterior part of the head of the spermatozoon. The long axis
of this bead-like mass coincides with that of the spermatozoon.
I have found a similar mass of protoplasm present in spermatozoa
from some individuals of Amblystoma punctatum, but here the
oval mass usually occurs about the junction of the head and mid-

82 BERTRAM G. SMITH

dle-piece. Probably the condition observed in the two species
represents a late developmental stage of the spermatozoon—the
metamorphosis of the spermatid into the spermatozoon is not
quite complete.

The amount of seminal fluid present at one time in the vasa
deferentia of a ripe male is very great in proportion to the size
of the animal—a condition correlated, doubtless, with external
fertilization. In one instance 20 cc. of seminal fluid was readily
stripped from a single male.

6. The method of fertilization

The method of fertilization (external) has already been de-
scribed (Smith, ’07); subsequent observations have supplemented
this account only in the fact, discussed later, that a single male
may spawn with more than one female.

The method of external fertilization is well adapted to the nor-
mal breeding conditions. The ‘nest’ of Cryptobranchus consists
of a hollow under rocks, a confined space protected from the cur-
rent, and filled with very quiet water. As has been shown, the
amount of milt that may be discharged at one time by a single
male is considerable; in the case of a pair that spawned while
being carried in a pail of water, it was sufficient to turn several
quarts of water milky white. Such a quantity of sperm set free
in the confined space of the nest would become diffused, especi-
ally when stirred about by the movements of the animals, so that
every egg would be quickly reached and fertilized. As a matter
of fact, few unfertilized eggs are found.

So far as I have been able to learn, this is the only case of exter-
nal fertilization recorded for the urodeles. The inconclusive
observations of Kerbert (04) on Cryptobranchus japonicus,
and Kunitomo (’10) on Hynobius, suggests to me that external
fertilization may take place in these forms.

In Necturus, from the observations of Kingsbury (’95) it seems
certain that internal fertilization takes place. A compound
receptaculum seminis or spermatheca is present in the female;
spermatozoa have been found in these spermathecae during the

EMBRYOLOGY OF CRYPTOBRANCHUS 83

fall and winter, which suggests an autumnal fertilization, though
it is possible that spermatozoa are left over from a spring fertili-
zation. The method of transference of the seminal fluid from the
male to the seminal receptacle of the female is unknown.

The breeding habits of some urodeles in which internal fertili-
zation takes place by means of spermatophores (e.g., Amblystoma
punctatum and Diemyctylus) have been considered by the writer
in former papers (Smith, ’07, 710 and ’11). Fertilization is exter-
nal in the anura, internal in the apoda.

In the elasmobranchs and holocephali, fertilization is internal.
In the crossopterygian Polypterus (Harrington, ’99; Kerr, ’07b),
during the breeding season the anal fin of the male is modified in
such a manner as to suggest internal fertilization; or possibly it
serves to direct the sperm against the stream of eggs issuing from
the female. Nothing conclusive is known regarding the method
of fertilization in the dipnoi. In teleostean fishes, with a few
exceptions, fertilization is external; e.g., as in Chrosomus (Smith,
’08).

The question whether external fertilization in Cryptobranchus
is primitive or secondarily acquired will be discussed under laste
genetic considerations in a later section.

7. The brooding habit

In a previous paper (Smith, ’07) a paternal brooding habit was
described for Cryptobranchus. This was observed in aquaria,
and more extensively under natural conditions.

The data on the existence of a paternal brooding habit under
natural conditions, while necessarily incomplete, are quite conclu-
sive. In one case, a male occupying a nest containing eggs was
observed to fight and drive away several males and a spent
female which were attempting to enter the nest (Smith, ’07);
in another case, a male occupying a nest containing eggs was
observed to oppose the attempt of a single male to enter the nest.
It is not always possible to tell whether an adult is present in the
nest; the rock may be too large to overturn, and while the eggs
may be obtained by tilting the rock with a crow-bar, this method

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1

84. BERTRAM G. SMITH

is not always successful in dislodging the adult. In cases where
the rock is lifted and overturned, the water is discolored and the
hellbender, aided by its protective coloration and the swift
current, may escape. When seen, however, it may usually be
captured. A record kept for six years (’06—11) shows that from
twenty-nine nests containing eggs a male was captured in ten
cases, a female never.

The duration of the brooding habit has not been definitely
determined, and perhaps varies greatly. In different nestsin which
a male was present, eggs were found in various stages of develop-
ment up to about three weeks old; unfortunately I was obliged
to discontinue field work at a date varying from two to four weéks
after the beginning of the breeding season. In no case where the
eggs were in an advanced stage of development can it be recorded
that the male had been continuously present, or even that he was
the same male that fertilized the eggs; but the entire absence
of females from nests containing eggs is significant.

With regard to the origin of this paternal brooding instinct
two suggestions (Smith, ’07) were made: (a) the brooding instinct
may have originated in connection with the feeding habit; or (6)
in holding the nest the male may be primarily concerned in await-
ing the coming of another ripe female. Both views assumed that
in Cryptobranchus we have an example of the brooding habit in an
incipient state. Further observations indicate that the brooding
habit is well established and manifested as a distinct impulse
from the moment of fertilization; its origin is thus thrown back
into the remote past, and concerning it we can only speculate.

The evidence for the first interpretation may be briefly stated
as follows: Both sexes are voracious eaters of the newly-laid eggs;
during the spawning season the majority of the adults taken have
the stomach filled with eggs. There is evidence that the females,
when opportunity is afforded, gorge themselves with eggs more
freely than the males. ‘The number of eggs found in the stomach
of a single adult usually ranges from fifteen to twenty-five, a
number sometimes greatly exceeded in the stomachs of spent
females. In one case, in which the body of a spent female ap-
peared greatly swollen, the stomach was found to be greatly dis-

EMBRYOLOGY OF CRYPTOBRANCHUS 85

tended with eggs. When removed and measured by displacement
of water, the stomach and its contents were found to have a bulk
of over 200 ec. The mouth also was full of eggs, and strings of
eggs protruded from the pharyngeal openings. The quantity of
eggs present seemed to represent almost an entire spawning.
The feat of swallowing such a quantity of eggs would seem pos-
sible only if they were taken before the swelling of their envelopes.

The digestive processes of the hellbender are extremely slow,
and I have taken undigested eggs from the stomach a week after
they were eaten. Under these conditions the presence of a single
male hellbender in the nest is in the main protective. On account
of the small number of eggs eaten at once, and the slowness of his
digestive processes, fewer eggs are eaten than would be the case
if other hellbenders, and especially the spent females, had free
access to the nest.

As previously noted, a male has been observed to fight and drive
away a spent female and several males that were attempting to
enter the nest. The male in such cases has the advantage over
the female because of the weakened condition of the latter; as
regards the other males, he has the advantage of position.

The facts suggest that the male, in thus driving away others of
his own kind, may be primarily concerned in guarding his own
food supply; this guarding habit may become modified into a true
brooding instinct. But it is difficult to believe that the male,
after having filled his stomach with eggs, would any longer be
concerned with the fate of the remaining eggs on account of their
value as food.

According to the second interpretation, the male may hold the
nest in expectation of the coming of another ripe female; an exten-
sion of this habit may give rise to the brooding instinct. Accord-
ing to this view, the brooding instinct has its origin in the breeding
habit.

The reception of a ripe female by a male guarding eggs has not
been directly observed, but the following data afford sufficient
evidence on this subject: Out of twenty-nine nests examined
during the seasons of 1906-1911 inclusive, eleven were found to
contain eggs of at least two different spawnings, the product of

86 BERTRAM G. SMITH

different females. The eggs were in different stages of develop-
ment, hence fertilized at different times. In two cases the num-
ber of eggs present in a single nest was sufficient to represent at
least three spawnings by different females. It is possible to deter-
mine this with considerable certainty, for the number of eggs
matured by a single female each season is limited, and these
are all laid at one time. Moreover it is known that the intensity
of the yellow color of the eggs is constant for all the eggs of a single
female, but varies considerably in eggs from different individuals.
In view of the observed vigilance and effectiveness of the male in
possession of the nest in driving away other males, it is highly
improbable that successive pairs of adults have occupied the nest;
hence the facts indicate that the same male has spawned with
successive females.

The second hypothesis seems supported by better evidence than
the first; but while it is entirely possible that such may have been
the origin of the habit in the remote past, there is evidence that
at present the eggs are the object of paternal care from the time
of fertilization, and this brooding instinct is only temporarily
overcome by hunger or diverted by the breeding instinct. The
behavior of males breeding in aquaria strongly suggests this:
after fertilizing the eggs the male usually remains close beside
them or crawls under or amongst them.

Concerning the brooding behavior of some specimens of Crypto-
branchus japonicus in captivity Kerbert (’04) says:

Nach Beendigung der Eiablage legte sich das Weibchen offenbar in
grésster Ermattung in eine Ecke des Behalters hin und kiimmerte sich
um das Gelege gar nicht mehr. Das Minnchen hingegen hat seitdem
die Kiermasse nicht verlassen—ja sogar die Brut fortwdhrend bewacht.

Denn sobald das Weibchen die Eiermasse zu nahe
kam, stiirzte das Mannchen in sichtbarer Wut auf die Mutter los und
vertrieb sie. . . . . kriecht der Mannliche Riesensalamander
zwischen den verschiedenen Strangen der Eiermasse hindurch und
bleibt dan von der Hiermasse umbhiillt liegen, oder er legt sich einfach
neben die Hiermasse hin. In beiden Fallen aber halt er, hauptsichlich
durch eine pendelartige Bewegung des ganzen Kérpers, von Zeit zu Zeit
die ganze Hiermasse in Bewegung. Durch diese Bewegung entsteht
eine fir den Atmungsprozess der Bier und Embryonen héchst wichtige

Wasserstromung, wahrend die Lage der Hiermasse hierdurch gleich-
zeitig fortwihrend wechselt.

EMBRYOLOGY OF CRYPTOBRANCHUS 87

It thus appears that the paternal brooding instinct in both
species of Cryptobranchus is manifested from the moment the
eggs are fertilized, though in C. allegheniensis at least it may be
temporarily inhibited or overcome by hunger or by the breeding
instinct. The brooding habit of Cryptobranchus is undoubtedly
very old, and we must look to other forms to find examples of it
in the incipient condition.

According to Whitman (’98) there are three distinct elements in
brooding behavior: (a) the disposition to remain with or over
the eggs; (b) the disposition to resist and to drive away enemies;
and (¢) periodicity. The first of these elements has its origin in
the need for rest, protection to the offspring being at first inci-
dental. The second element, pugnacity, is periodical and a part
of the reproductive cycle. The third element, periodicity, is
apparently an attribute of the two other elements, based on physi-
ological conditions; its adaptiveness lies in correlating the other
two elements with the hatching period of the eggs.

In Cryptobranchus, after spawning, the female is evidently
much the weaker of the two; as a matter of observed fact, she is
driven away by the stronger and more pugnacious male. It
can scarcely be the need for rest that keeps the male in the nest,
since he maintains exclusive possession at the cost of alert watch-
fulness and occasional combat. If the element of weakness were
the important factor in initiating the brooding habit, we should
expect the female rather than the male to remain in the nest.
It may be that primitively the brooding impulse is a phase of the
reproductive cycle that applies to both sexes, the female losing
it on account of her hungry and exhausted condition due to the
accumulation of a large amount of yolk in the egg. Perhaps
(for this suggestion I am indebted to Professor 8. J. Holmes)
on the part of the male there is involved a proprietary interest
in the nest, which he has chosen and in part excavated, and which
he occupies as an advantageous breeding place and as a more or
less permanent home.

To obtain conclusive evidence regarding the origin of the brood-
ing habit one must study a series of closely related forms illus-
trating the habit in the making. In Cryptobranchus it appears

8S BERTRAM G. SMITH

that the habit is well established; it is improbable that the question
can be settled by the study of this form alone, and the data here
given are presented only in the hope that they may contribute
something toward the final solution of the problem.

Concerning the brooding habit of C. japonicus in its natural
habitat Ishikawa (’04) says: ‘‘Fast in jedem Loch, wo man von
Ende August bis zu Anfang October ein weibliches Tier gefunden
hat, findet man einen Eiklumpen. Dieser Umstand lasst schon
vermuthen, dass das Tier eine Brutpflege hat wie Ichthyophis
oder wie so viele andere Amphibien.”’ Kerbert, however, asserts
(04) that it is the male that guards the eggs, and states that the
sex of his specimens was carefully determined.

Other amphibia known to possess brooding habits are the uro-
deles Desmognathus and Plethodon; the caecilians Ichthyophis
and Hypogeophis; Alytes and several other anura (Wiedersheim
700). In the cases of Desmognathus, Plethodon, Ichthyophis
and Hypogeophis the female is said to care for the eggs; in the
case of Alytes, the male.

The brooding habit seems to be lacking in Necturus. According
to Eycleshymer (’06), Necturus sometimes eats the eggs of its
own species.

The brooding habit is well known in many teleosts, and in
Amia (Dean, ’96; Reighard, ’03); it is well developed in the lung-
fishes Protopterus (Budgett, ’01 a and ’01 b), and Lepidosiren
(Kerr, 00). I can find no record of any observations pointing
to the existence of a brooding habit in the crossopterygil.

D. SUMMARY

The breeding season of Cryptobranchus allegheniensis in north-
western Pennsylvania begins about the first of September and lasts
about two weeks.

There is a tendency toward segregation of the females from the
breeding grounds during a period extending from the close of
the breeding season until the middle of the following summer.

About 450 eggs are matured each year by an adult female of
average size. The egg capsules from each oviduct are fastened

EMBRYOLOGY OF CRYPTOBRANCHUS 89

together in a single string. At the ends of each string are formed
some small but perfect capsules which do not contain eggs.

The eggs are unpigmented, heavily yolk-laden and strongly
telolecithal.

The nest consists of a submerged cavity under a rock in the bed
of the stream. The cavity is sometimes in part the work of the
animal.

Fertilization is external.

There is a paternal brooding habit, which is manifested from
the moment of fertilization. The origin of this habit is problem-
atical.

III. METHODS AND TECHNIQUE

A. COLLECTION AND CARE OF LIVING MATERIAL

To insure a convenient supply of adults for various purposes,
these were collected before and during the breeding season and
placed in a large creek aquarium, constructed of wire netting
and placed in shallow water with a gentle current. This arrange-
ment of the aquarium afforded abundant aération; flat stones
placed on the bottom provided cover; in general the conditions
closely resembled those of the natural environment. The aqua-
rium proved of great value as a means of insuring a supply of
adults for use at frequent intervals in securing material for the
study of ovogenesis, maturation, fertilization and the early cleav-
age stages.

Artificial fertilization was often resorted to in order to control
the time of fertilization for the study of fertilization and early
cleavage stages, and occasionally eggs were used that had been
deposited and fertilized by specimens in captivity; but the greater
part of the material used for the study of the development was
obtained from the nests of the animals in their natural environ-
ment.

At first the problem of keeping the eggs alive in a favorable
environment while studying their development promised some
difficulty. Early attempts to keep the eggs in creek aquaria
met with disastrous failure through the attacks of water-mould.
The method finally employed was to keep the eggs in shallow

9() BERTRAM G. SMITH

earthenware dishes containing well water, in a cool cellar; a lim-
ited number of eggs were placed in each dish, and the water
changed daily. Under these conditions they developed normally.
During the early autumn all the laboratory work onthe living egg,
and the preservation of material, were carried on in this cellar,
so that at no time were the eggs subjected to an unfavorable tem-
perature. The eggs were in general shielded from the light;
but for working purposes both direct and diffused sunlight, or
a Welsbach light, were used.

On account of teaching duties observations in the field have
never extended quite to the time of hatching, consequently it
has been necessary to transport the living embryo for consider-
able distances. In the case of embryos taken after the closure of
the neural folds, material shipped in cool weather by express, in a
pail containing shallow water, did quite as well as material which
was given personal care during transportation and for which the
temperature was regulated with ice; in both cases the embryos
developed normally. Younger embryos require much greater
care in transportation; material in cleavage and gastrula stages
shipped by express has usually died or developed abnormally,
perhaps in the main because of untimely warm weather; all such
material was discarded. Material kept in the laboratory thrives
in shallow dishes containing well water, the dishes being partly
immersed in cool running water; no artificial a€ration is neces-
sary. As a check on possible abnormalities in material that has
been transported, I have had a series of late stages preserved from
material kept without transportation.

B. FIXATION AND PRESERVATION OF MATERIAL

The envelopes may be removed in any stage without much
difficulty, by means of scissors. This is very easily done after
the eggs have been in water for several days, since the envelopes
become inflated. For earlier stages, more care is necessary.
Eggs from the uterus, and fertilization stages, may be handled
more rapidly by fixing in Solution B (see below) before the
removal of the envelopes; they may be preserved thus in formalin,
but not in alcohol. After fixation the envelopes become brittle

EMBRYOLOGY OF CRYPTOBRANCHUS 9]

and may readily be removed with needles. Comparison with
eggs fixed after the removal of the envelopes shows no essential
difference in the results.

The fixation of such large and heavily yolk-laden holoblastic
eges presented a problem of considerable difficulty. A great
variety of the usual fixing fluids were tried, but none of them
succeeded without modification. After extensive experimenta-
tion, the mixture described below as Solution B was found to be
very satisfactory for all the yolk-laden stages, for surface study,
photography and for sectioning.

The following fixing solutions were found useful for the purposes
indicated:

Solution A. Formalin, 10 per cent. Useful for preserving
eggs in the envelopes for demonstration purposes, or for the study
of the envelopes, as it leaves the envelopes clear and preserves the
eggs in their natural color. Formalin is of some value for the
surface study of cleavage, as it brings out the faint cleavage fur-
rows of the lower hemisphere with great distinctness, and occa-
sionally gives remarkably good preparations for the surface study
of the cleavage of the upper hemisphere. In general the fixa-
tion of the micromeres is unsatisfactory, both for surface study
and for sectioning. Formalin is unsurpassed for fixing larvae
for museum purposes; for permanent preservation they should
be changed to alcohol.

Solution B. Bichromate-acetic-formalin. The following pro-
portions must be quite strictly adhered to:

OPA SUMMIBLG MROUNAL ess Se dee ses eee acs Lahr aih Megtee a hands cee! 1 gram
Glacialkace tema claters wtare Nee o2 20 tet slg ae apes, eee dies ate 24 ce.
Schering’s formalin, added at the time of using................. 5 &
MAIER? cides 6 oie Wy & CI IONS aE eee arte ree ain 7, @@-

Fix about forty-eight hours in plenty of the solution, at a low
temperature; change the solution once or twice.

Rinse in water and wash in 5 per cent formalin, in the dark,
for at least two weeks, changing the formalin as often asit becomes
discolored; preserve in 5 per cent formalin. Preservation in alco-
hol also gives good results for sectioning, but is not so good for

surface study nor for photography.

92 BERTRAM G. SMITH

During the process of washing in formalin the color changes
from yellow to green. The yolk becomes dark green, while the
blastodise or embryo proper is much lighter in color, giving a
sharp differentiation of the protoplasmic portions of the egg. The
form of the egg is preserved perfectly, and remarkably good defini-
tion for surface study is secured. The eggs are easily sectioned
by the paraffin method.

Not until after the closure of the neural folds is it possible to
alter the proportions in the formula as given above without
injury to the form of the embryo; an increase in the proportion of
potassium bichromate results in the collapse of the embryo when
in melted paraffin if not in an earlier stage of the process of prepa-
rationforimbedding. For later stages the proportion of potassium
bichromate may be slightly increased (e.g., to 13 per cent), with-
out detriment to the surface features and perhaps with some gain
in the histological results.

Solution C. Sublimate-acetic-formalin.

Saturated solution corrosive sublimate in 10 per cent formalin. . .975 parts
MerlcroleacetiCy ACUG:, «5. 5 cn An in Netoete i tcheriatetine brawios bitch s Mite mete 2% parts

Fix for a few hours, then transfer to formalin for a few days to
insure thorough fixation of the yolk. Wash and preserve in either
formalin or alcohol.

This is not so satisfactory a fixing solution as Solution B, but
may be used for comparison. For surface study the results,
especially in the early stages, are decidedly inferior to those
secured with Solution B. For sectioning, good results are se-
cured in the early cleavage stages and after the closure of the
neural folds; the mercury crystals must be removed by prolonged
treatment with iodin. In the blastula and gastrula stages the
embryo usually collapses during the process of preparing for im-
bedding.

Solution D: Lavdowsky’s.

Ao} BOTY SHON eves pb oD CNS CeO EES cto Tatra ia ar.c His & od LD MOIIEME OGG, Ox 10 parts
ATGORG];DOcEE COME ci.. 2 ove ods 4 2 - crake oa tent ete ne EMED sab, coe eked 50 parts
Glacial seetie saciid ish. dns hic ae he. ied eae etaaee eae She ks Rasta 2 parts

Warten. lcci ewes Bons h Avid cis dua. ea A ee eet ors eae 40 parts

EMBRYOLOGY OF CRYPTOBRANCHUS 93

Fix for several days; preserve in 70 per cent or 80 per cent
alcohol.

This mixture is especially useful for the yolk-laden ovarian
eggs, and for maturation stages; it is not very satisfactory for
embryonic stages. Envelopes, if present, must be removed before
the eggs are fixed in this solution. The best results are obtained
by sectioning the material soon after preservation.

Solution EH: Zenker’s. This mixture was found most useful
for the early stages of ovogenesis, before the formation of any
considerable amount of yolk. It is not good for embryonic stages,
unless parts of the embryo are to be dissected off from the yolk
before sectioning. It gives very inferior preparations for sur-
face study in every stage.

For the early stages of ovogenesis, before the formation of yolk,
both Flemming’s and Bouin’s solutions were used with fair results.
For larvae after the disappearance of the yolk sac, Tellyesnicky’s,
Zenker’s, or almost any good fixing solution may be used.

Of the various mixtures experimented with for the yolk-laden
stages, those containing picric acid proved to be the very worst.
The invariable result of the use of a solution containing picric
acid was to cause the egg to disintegrate.

In preserving the embryological material of Necturus, Solution
B was principally used. In the early stages of development,
before the formation of the neural folds, the results are not so
uniformly good as with Cryptobranchus; this is perhaps due to
the fact that in these stages the eggs of Necturus are almost
necessarily preserved before the removal of the very closely-fitting
gelatinous envelopes. In successfully preserved eggs in the
cleavage stages, the furrows of the upper hemisphere are more
conspicuous and the contour of the micromeres more rounded,
than in Cryptobranchus; they are thus, except for difficulties
arising from the character of the envelopes, more favorable objects
for photography. Late gastrula and neural groove stages of
Necturus, preserved by this method, are rarely so favorable for
surface study as the same stages in Cryptobranchus. After the
formation of the neural folds, when a space has appeared between
the envelope and the egg, the embryos of Necturus are preserved

94 BERTRAM G. SMITH

with very uniform success, whether fixed before or after the
removal of their envelopes. In particular, stages after the clo-
sure of the neural folds give a sharpness of detail in the surface
features rarely found in Cryptobranchus; these stages of Nec-
turus are very favorable objects for photography.

C. SECTIONING AND STAINING

Kerr (’01), in describing the technique employed in studying
the egg of Lepidosiren, has well said: ‘‘The investigation of a
holoblastic egg 7 mm. in diameter and packed with yolk involves
great technical difficulties, for the whole of each egg has to be
converted into thin sections. The full extent of these difficulties
will only be appreciated by embryologists who have essayed a
similar task.’”’ In sectioning the heavily yolk-laden stages, Kerr
used the celloidin method, and a combination of the celloidin
and paraffin methods. DeBussy (04) used the celloidin method
in studying the cleavage stages of Cryptobranchus japonicus.

For sectioning the embryological material of Cryptobranchus
allegheniensis and Necturus I have used the paraffin method
exclusively; success with this method was found to be entirely
a matter of careful attention to technique. The most important
considerations are proper fixation and washing, and thorough
infiltration with paraffin. In handling serial sections of large
numbers of these eggs the advantage of the paraffin method is
obvious.

With regard to staining, for general purposes the best results
were obtained by staining in toto with Grenacher’s borax carmine,
and counterstaining on the slide with Lyons blue in absolute
alcohol; to the Lyons blue solution sufficient picric acid was added
to turn it green. By this method the effect of a triple stain,
with excellent differentiation, is obtained. The chromatin is
stained red, cell walls and cytoplasm blue; the yolk is first stained
red by the borax carmine, but turns green in the counterstain.
It is usually best to cut short the action of the counterstain at
a time when the smaller yolk particles are stained green, while
the larger ones are left red. The method has the advantage of

EMBRYOLOGY OF CRYPTOBRANCHUS 95

rapidity, an important consideration when great series of large
sections are to be handled in considerable numbers.

In sectioning and staining the early cleavage stages the exact
mode of procedure is as follows:

From formalin pass the eggs to alcohol, 35 per cent, 50 per cent, two hours each.

Grenacher’s borax carmine in 70 per cent alcohol, about two days.

Acid alcohol (0.25 per cent HCl in 70 per cent alcohol), about two hours.

Ninety-five per cent alcohol, two to twelve hours; 100 per cent alcohol, two to
three hours.

Xylol, four to ten hours.

Paraffin with melting point 52° C. (at a temperature not exceeding 55° C.), two
days. Change the paraffin at least once.

Imbed in a paper box, hardening the block under alcohol.

Cut sections 10u to 15u thick, using a Minot rotary microtome.

Counterstain on the slide with Lyons b!ue and picric acid mixture in absolute
alcohol.

Wash in xylol long enough to destain slightly.

Mount in Canada balsam.

Early stages require longer for fluids (especially paraffin) to
penetrate than do later stages. For the yolk-laden ovarian eggs,
and maturation and fertilization stages, from two to three days
in borax carmine, and about three days in melted paraffin, are
necessary. I have found no serious ill effects in these stages from
this prolonged immersion in paraffin at the temperature given.

Material fixed in Lavdowsky’s solution stains and infiltrates
more rapidly than with the other methods of fixation; also the
yolk is less likely to crumble.

Late cleavage, and gastrula stages, are penetrated by the vari-
ous fluids more rapidly than the early cleavage stages, so that the
time may be reduced to two-thirds or one half. For still later
stages, there is a further gradual reduction in the length of time
required.

During the present year, at the suggestion of Professor Wilson,
I have employed a slight modification of the method described
above. After being cleared in xylol, the objects were left several
days in a mixture of xylol and paraffin at about 38°C. By this
preliminary treatment, the time required for infiltration with
melted paraffin at a high temperature was reduced at least one-
half, with some improvement in the quality of the preparations.

96 BERTRAM G. SMITH

IV. THE EXTERNAL HISTORY OF THE EGG BEFORE CLEAVAGE

Some superficial aspects of the history of the egg before cleavage
have already been considered in connection with the account of
the breeding habits.

A. EXTERNAL CHANGES PRECEDING AND ACCOMPANYING
MATURATION

Except where otherwise mentioned, the observations recorded
under this heading were made on the living egg.

If the ovary of an adult Cryptobranchus be examined at any
time during the summer, the eggs which are about to become ma-
ture are readily distinguishable by their much greater size and
yolk content.

In the living ovaries of adults taken about the middle of August,
the eggs show no positive surface indications of a telolecithal
structure. The same eggs fixed by a variety of methods show a
circular area or ‘calotte’ about 60° in diameter, which is somewhat
lighter in color than the remaining surface of the egg. On account
of its large size the egg now causes the ovarian wall to bulge strongly
outward. In general the pale circular area is situated in the cen-
ter of the more exposed hemisphere of the egg, and is not so pro-
fusely covered with ovarian blood-vessels as the remainder of this
hemisphere, but this relation is not always exact.

Sections show that the calotte is the outward expression of a
peripheral dise-shaped region richer in protoplasm and small
yolk granules than the remainder of the egg; in the center of this
dise lies the germinal vesicle. From its homologue in the teleos-
tean egg I shall call this region the germinal disc or blastodisc;
in surface views it may be referred to by the same names, or more
strictly speaking, as the germinal area. Fixation serves to accen-
tuate the optical differences between the germinal disc and the
remainder of the egg, making the germinal area visible in pre-
served material at an earlier stage than in the living egg. The
center of the germinal area defines the animal pole of the egg. _

Shortly before the egg is ready to leave the ovary, the germinal
vesicle appears at the very surface, at the center of the germinal

EMBRYOLOGY OF CRYPTOBRANCHUS 97

area which is now visible in the living egg. After remaining here
for a length of time that has not been accurately determined, the
germinal vesicle disappears from view leaving only a faint dark
spot to mark its former site.

In most ovaries obtained about the time of the beginning of
the breeding season (the last week in August and the first week in
September), all stages in the emergence of the germinal vesicle
will be found; in some eggs the germinal vesicle has not yet
reached the surface, but in a considerable proportion of cases it
will be found exposed in varying degrees (see fig. 53).

The phenomena concerned with the appearance of the germinal
vesicle at the surface are very striking, owing to the large size
of the germinal vesicle, the sharp contrast between its transpar-
ent fluid contents and the surrounding opaque substance of the
egg, and the distinct appearance of several opaque-white bodies,
presumably nucleoli, within the germinal vesicle. All this may be
seen even with the naked eye.

In order to obtain the sequence of the changes occurring in
a single egg during this stage, many individual eggs were isolated
in normal salt solution, or identified while in position in the ovary,
and kept under observation for several hours. In the case of the
first ovary studied during the fall of 1907, some of these eggs
changed sufficiently before death ensued, to enable me, by combin-
ing several individual histories, to get a fairly complete idea of the
normal course of events in a given egg. But in succeeding years,
although several dozen ovaries containing eggs in this stage have
been studied in females recently killed or anaesthetized with chlo-
retone, no marked changes could be detected. Hence in the
following account dependence is placed chiefly on a comparison
of individual eggs in the same freshly-exposed ovary, and a com-
parison of ovaries in slightly different stages of development.. In
particular, incipient stages in the approach of the germinal vesicle
to the surface could be distinguished from possible later stages in
which it has disappeared from view, through a comparison of
ovaries such as those described above, with others in which nearly
all the eggs had been set free from the ovary.

98 BERTRAM G. SMITH

The first indication of the approach of the germinal vesicle
to the surface is the appearance of a faint dark spot, 1 to 2 mm. in
diameter, in the center of the blastodise. This dark spot grows
more distinct; it is the optical effect of the scarcely-submerged
germinal vesicle. Within this large dark area appears a small
sharply defined much darker area, circular in outline, which
grows at the expense of the larger and fainter dark area. At the
time of its first appearance the small dark area is sometimes seen
to pulsate slowly, quiver and change form, disappear and reap-
pear. This small dark spot is a portion of the germinal vesicle
which is actually in contact with the zona radiata (see section
V). It may increase in size until almost an entire hemisphere of
the germinal vesicle is exposed. In light of moderate intensity
the germinal vesicle appears as a deep, dark well of transparent
substance walled in by the opaque material of the blastodise;
in strong sunlight one may see within the germinal vesicle the
reflection of the bright yellow yolk beneath. Several opaque-
white bodies of various sizes appear within the germinal vesicle;
these are probably nucleoli, though the largest ones are much larger
than the nucleoli shown in sections.

An ovarian egg dissected out and immersed in water at the time
of the appearance of the germinal vesicle at the surface orients
itself with the animal pole upward.

The actual disappearance of the germinal vesicle from the
surface, and the relation of this process to the rupture of the
nuclear wall, have not been satisfactorily observed. Whether the
germinal vesicle recedes slightly from the surface before or during
the rupture of its wall, or disintegrates at the very surface, has
not been positively established; it is possible that all three con-
ditions occur in different eggs. In several cases there seemed
to be a welling-up of material from the germinal vesicle which
spread out to form a broad crater at the surface; in other cases the
appearances favored the impression of a slight subsidence of the
germinal vesicle. It is possible that the egg has never been ob-
served at the exact time of the rupture of the nuclear wall, for
though a large number of eggs from ovaries containing eggs with

EMBRYOLOGY OF CRYPTOBRANCHUS 99

the germinal vesicle at the surface have been sectioned, in none
of these eggs has the nuclear wall been found ruptured.

Several females have been taken in which only a few eggs
remained in the ovary, the others being found in the body cavity,
oviduct and uterus. The ovarian eggs of such specimens were
invariably found to be in a later stage than those just described:
sections showed that the dissolution of the germinal vesicle was
complete, and in surface views these eggs showed a small faint
dark spot or slight depression at the animal pole (see fig. 5).

5 6

Fig. 5 Surface view of the animal hemisphere of an egg of Cryptobranchus
allegheniensis ready to leave the ovary, after the rupture of the germinal vesicle.
The lightly stippled area indicates the blastodisc. Sketched from the living egg.
Deets

Fig.6 Surface view of the animal hemisphere of an egg taken from the uterus,
ready for fertilization, showing pit at the center of the blastodise. Sketched from
preserved material. X 7.

The dark spot is sometimes surrounded by a tumid ring, but this
condition is probably pathological.

At the time of the escape of the egg from the ovary and its
passage through the body cavity and upper oviduct, the egg seems
softer in consistency than at other times. Some fixing solutions,
particularly Lavdowsky’s, which usually preserve perfectly the
spherical form of the egg, now fix it as an irregularly-shaped mass.
This plasticity may be of use to the egg in its escape from the
ovary and passage down the oviduct.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1

100 BERTRAM G. SMITH

In eggs taken from the lower oviduct there is found a slight
extension of the blastodise and a marked increase in the intensity
of its differentiation, both in living and preserved material.
Moreover, outside the rather indefinite limits of the blastodise
proper there seems to be a continuation of the same sort ot mate-
rial asan extremely thin whitish superficial layer extending beyond
the equator and well into the lower hemisphere.

In eggs entering the uterus the blastodisc is well differentiated
throughout an area about 90° in diameter, while the entire remain-
ing surface of the egg shows a slight paleness as compared with
earlier stages. '

The dark spot or shallow depression at the animal pole persists,
though often very faintly, up to about the time of fertilization,
when its site is occupied by a minute but deep and sharply-
defined pit (see fig. 6). The change usually does not take place
until after the eggs have been for some time in the uterus. As
shown by the study of sections, the appearance of this pit usually
coincides with the time of formation of the second polar spindle.

B. CAPACITY OF UTERINE EGGS FOR FERTILIZATION

To test whether eggs newly arrived in the uterus are capable
of fertilization, a female was taken in which only a small portion
of the eggs had reached the uteri, the others being distributed
all along the route from ovary to uterus. The eggs from one
uterus—about 75 in number—were mixed with milt after the
usual manner in artificial fertilization. Of the entire lot, not a
single egg developed.

In another female nearly all the eggs had arrived in the uteri,
a few remaining in the oviducts and body cavity, and none in
the ovaries. All the eggs from the uteri were mixed with milt;
about 5 per cent of them developed.

In a third female all the eggs were in the uteri, but none of them
showed a distinct pit at the animal pole—evidence that they had
only recently entered the uterus. All the eggs were mixed with
milt; none of them developed. It should be noted that this
female was evidently in the first year of sexual maturity and the

EMBRYOLOGY OF CRYPTOBRANCHUS 101

eges may have been slow in undergoing maturation changes, or
defective in some way.

In the great majority of cases of females taken with all the eggs
in the uteri, artificial fertilization has been successfully performed ;
a high percentage of fertilized eggs is reached when all the eggs
show a distinct pit at the animal pole. In every case in which
seminal fluid was examined under the microscope during the breed-
ing season, the spermatozoa were motile; so it is not likely that
any cases of failure in artificial fertilization were due to defective
spermatozoa.

The evidence indicates that the eggs are incapable of fertili-
zation at the time when the first eggs reach the uterus, but that
about the time all the eggs reach the uterus the majority of them
become capable of fertilization. This change in their potentiality
coincides in time with, or slightly precedes, the formation of a
distinct pit at the animal pole; it is probably correlated with the
formation of the second polar spindle (see section V).

C. CHANGES VISIBLE FROM THE SURFACE DURING FERTILIZATION

The appearance of the blastodise shortly after fertilization is
shown in figs. 7 and 8. During the first eight hours after fertili-
zation there is an increase in the extent of the blastodisc from a
diameter of 90° to 130°-160°, with a corresponding increase in
the intensity of its differentiation. From this time up to first
cleavage there is no constant increase in the extent of the blastq-
disc, though the transition from the blastodise to the darker region
surrounding the vegetal pole becomes more gradual. The pit
at the animal pole persists unchanged almost up to the time of
first cleavage; it is sometimes double (see fig. 10). Shortly before
first cleavage it becomes broader and shallower, and usually dis-
appears before the beginning of the first cleavage furrow.

As early as fifteen minutes after artificial fertilization, pits or
scars made by the actual or attempted entrance of a spermatozoon
have been found on the surface of the egg. It seems remarkable
that the spermatozoon can pierce through the thick and tough
gelatinous capsule in so short a time. In living material, the

102 BERTRAM G. SMITH

point of actual or attempted entrance of a spermatozoon is
often visible as a minute but sharply-defined pit, barely visible
to the naked eye; hence the name ‘sperm pit’ will be used to
designate the precise locality where the spermatozoon enters,
though the word ‘pit’ does not always accurately describe the
appearance in preserved material.

The sperm pits are best studied in material killed in the bichro-
mate-acetic-formalin mixture and preserved in formalin. The
‘pits’ are not all alike, but readily fall into the following classes,
which probably represent consecutive stages in the penetration of
the egg by the spermatozoon (see fig. 8):

(a). A simple pit, deep and sharply defined, as observed in
living material.

(b). The pit is surrounded by a very small circular opaque
white spot.

(c). The pit has disappeared, and the white spot remains. This
type is most numerous. (Rarely, the pit persists until much
later—see fig. 10.)

(d). The white spot is surrounded and sharply limited by a
dark circular line.

(e). The white spot is surrounded by two concentric circular
lines separated by a narrow space which is darker than the general
surface of the egg (best shown in fig. 7).

It is not always possible to tell from surface views whether the
spermatozoon has actually entered the egg, but from the study of

* Fig. 7 Equatorial view of an egg of Cryptobranchus allegheniensis, 15 minutes
after fertilization, showing a single sperm pit. The lightly stippled area in the
upper part of the figure indicates the extent of the blastodisc.

Fig. 8 Equatorial view of an egg 45 minutes after fertilization, showing numer-
ous sperm pits.

Fig. 9 View of the animal hemisphere of an egg 3{¢ hours after fertilization,
showing a sperm area near the edge of the blastodisc.

Fig. 10 View of the animal hemisphere of an egg 3} hours after fertilization,
showing a later stage in the history of the sperm area. Theboundary of thesperm
area is a trifle too conspicuous in the figure.

Fig. 11 Equatorial view of an egg 63 hours after fertilization, showing further
extension of the sperm area.

Fig. 12 View of the animal hemisphere of an egg 7} hours after fertilization,
showing two sperm areas, on opposite sides of the blastodisc.

All the figures are drawn from preserved material. 7.

EMBRYOLOGY OF CRYPTOBRANCHUS 103

11 12

———_———

104 BERTRAM G. SMITH

sections it appears that the first three types of sperm pits indicate
that the spermatozoon has barely penetrated through the cell
wall, or that the attempt is an abortive one; the last two types
indicate with considerable certainty that the spermatozoon has
penetrated well into the egg.

Polyspermy is the rule. Cases of penetration by more than one
spermatozoon have been found fifteen minutes after fertilization,
while the surface of the egg may be scarred by a dozen or more
wounds presumably made by other spermatozoa. An hour later,
the majority of the eggs have been penetrated each by from one
to ten spermatozoa, and sometimes scarred by as many as fifty
more. In one case observed the entire number of sperm pits
reached nearly a hundred.

About three hours after fertilization the small white spot
representing the sperm pit is surrounded by a circular area about
10° to 15° in diameter, slightly darker than the general surface
(see fig. 9). An hour later this area has increased in size, is whiter
throughout its central portion, and is sharply bounded by a dark
line which forms a perfect circle (see fig. 10). This dark line is,
partly at least, due to a slight depression in the general surface of
the egg. For convenience the area enclosed by this circle will be
called the ‘sperm area.’

During the next few hours the sperm area increases in size
until it covers almost an entire hemisphere (figs. 11 and 12). Its
surface is now in general a trifle paler than the remainder of the
egg outside the blastodise; its boundary may pass the animal pole
without interruption. Two. or even three sperm areas in this
advanced stage may be present, their boundaries usually over-
lapping. Fig. 12 shows an egg fertilized from two opposite sides,
the spermatozoa entering near the margin of the blastodisc;
the two sperm areas meet at the animal pole, but remain widely
separated in the lower hemisphere.

In monospermic eggs preserved and dissected in this stage,
the sperm area is found to overlie a lenticular or disc-shaped
mass, of firmer consistency than the remainder of the egg which
may sometimes be shelled off in a few concentric layers like the
fleshy part of an onion. Outside the boundaries of both sperm

EMBRYOLOGY OF CRYPTOBRANCHUS 105

area and blastodise there is left a crescentic area which retains the
usual color of the heavily yolk-laden portions of the egg; this
region often shows numerous fissures in the yolk, running parallel
to the margin of the sperm area. These fissures separate the
layers previously mentioned. In position and outline this area
corresponds very nearly to the ‘gray crescent’ of the frog’s egg
(Roux, ’83, 85, 87 and ’03; Schultze, ’00; see also Jenkinson, ’09,
p. 80 and fig. 43).

Within eight to twelve hours after fertilization the sperm
pits have become indistinct and, as a rule, they all disappear
before the first cleavage, though cases have been found as late as
the fifth cleavage stage. Meanwhile the sperm areas also become
indistinct, losing the dark line which serves as a boundary and
gradually blending with the surrounding surface of the egg. Fif-
teen or twenty hours after fertilization, it is usually impossible
to orient the egg with respect to the point of entrance of a sperma-
tozoon; before the egg is ready for first cleavage it has resumed
the general appearance of radial symmetry which it had before
fertilization.

The sperm areas have not been observed in living material, but
the examination was made without the aid of a binocular micro-
scope, an instrument which has proved of great value in thesur-
face study of the fertilization stage with preserved material.

In preserved material a space sometimes appears between the
blastodise and the vitelline membrane which elsewhere closely
invests the egg. An examination of living eggs at intervals from
fertilization to first cleavage shows that normally the vitellinemem-
brane fits closely about the entire egg. The condition noted in
preserved material is due to the subsidence of the blastodise;
the vitelline membrane does not spring away from the egg after
fertilization, as occurs in some lower forms.

If one remove an unfertilized egg from its gelatinous envelope
and immerse it in water, and place almost in contact with it a drop
of seminal fluid, one observes that the spermatozoa by means of
slow writhing movements disperse gradually in all directions.
There is no evidence of attraction by the egg, but spermatozoa
coming in chance contact with it adhere to its surface, so that in

106 BERTRAM G. SMITH

time there are more spermatozoa at the surface of the egg than at
a little distance from it. As previously noted, spermatozoa are
found in capsules that do not contain eggs; in this case there is
no possibility of attraction by the egg.

In Cryptobranchus, as in other amphibian eggs, there is no
preformed micropyle. In eggs fertilized in a natural manner, the
spermatozoon may enter the egg at any point. More sperm
pits have been found in the marginal region of the blastodise,
about midway between the equator and the animal pole, than
elsewhere, indicating that this zone may be especially favorable
to the entrance of the spermatozoon; but if any selective influence
is at work, it cannot be’a strong one, for spermatozoa have been
found penetrating the egg close to the second polar spindle, and
at various points in the lower hemisphere, even at the vegetal
pole. Sperm areas are best developed about those sperm pits that
occur near the margin of the blastodise. In only one case has a
sperm pit at the vegetal pole been found surrounded by a sperm
area. Sperm pits are often more numerous on one side of the egg
than on the opposite side, indicating a chance inequality in the
exposure of the egg to the seminal fluid.

All the statements in this section regarding penetration of the
egg by the spermatozoa have been confirmed by sectioning eggs
which have first been carefully described externally.

D. SUMMARY

A germinal area is first visible in the ovarian egg taken about
the middle of August. Thegerminal area is usually situated on the
more exposed side of the egg, toward the periphery of the ovary;
it has at first a diameter of about 60°, and increases gradually in
size until about the time of first cleavage; it has then a diameter of
about 145°.

In ovarian eggs examined about the first of September, the
germinal vesicle is usually visible at the surface, in the center of
the blastodisc; it disappears shortly before the egg leaves the
ovary.

Soon after the eggs have reached the uterus, a sharply-defined
pit appears at the animal pole; this pit persists up to the time of

EMBRYOLOGY OF CRYPTOBRANCHUS 107

first cléavage. About the time of the appearance of the pit
at the animal pole, the egg becomes capable of fertilization.

The point of entrance of a spermatozoon (the ‘sperm pit’) is
easily recognizable in both living and preserved eggs. In pre-
served material, the influence of the spermatozoon on the egg sub-
stance is indicated in surface views by the differentiation of a
large circular area (the ‘sperm area’) surrounding the sperm pit.
This area is recognizable by a slight difference in color and by the
presence of a bounding dark line; it increases in size until it covers
nearly a hemisphere of the egg, then disappears.

In artificially fertilized eggs, and presumably in eggs fertilized
in nature, polyspermy usually occurs.

There is no evidence of attraction of the spermatozoon by the
ege.

The spermatozoon may enter the egg at any point, but sperm
areas are best developed about those sperm pits that occur near
the margin of the blastodisc.

V. THE INTERNAL HISTORY OF THE EGG BEFORE CLEAVAGE

The present section deals with a few features concerned in
ovogenesis and maturation, and gives a more detailed account of
the fertilization phenomena.

A. OVOGENESIS

The material for this study consists as follows:

(a). For the early stages it was found best to use larval and
immature post-larval females, with a body-length ranging from
9 to 38 em. (two years old and upward). Females with a body
length of more than 38 cm. are almost always sexually mature.

(b). The residual eggs of spent females taken in September
furnished ovocytes slightly older than those of the largest imma-
ture females taken in August and September.

(c). Mature females taken during July and August furnished
material for the late stages of ovogenesis.

There remains a period during the last year of development,
extending from October toJune inclusive, which is not represented

108 BERTRAM G. SMITH

in the material. Since during this time, which includes the win-
ter months, development is least active, the lack of these stages is
of minor importance for the purposes of the present paper.

1. The formation of the follicle and the egg membranes

The young ovary of Cryptobranchus is essentially a sae with
thick cellular walls. In a 9 em. larva the ovarian wall (see figs.
13 to 17) shows structural differentiation as follows: (a) an inner
and an outer limiting membrane of flattened epithelium; these
membranes are connected by (b) a network of cells of a character
similar to those comprising the limiting membranes, though usually
not so greatly flattened; within the meshes of this network are
found (c) young ovocytes in various stages of development.

In the ovary of a 9 em. larva, more or less clearly defined
groups or cysts of very young ovocytes (see fig. 13) may be found,
each group surrounded by a thin epithelial membrane, the cyst
membrane. All the ovocytes of each group or cyst are presum-
ably the product of a single primary ovogonium. Epithelial cells
also occur within the cyst. Within many of these cysts, develop-
ment has gone further, and some or perhaps all the ovocytes have
undergone an increase in size which involves both nucleus and
cytoplasm (see fig. 14). Within each cyst, one ovocyte usually
outstrips its fellows, and becomes surrounded by a layer of epi-
thelial cells which form the follicle (fig. 15).

With a further increase in size of the ovocyte, the follicular
layer assumes the character of a definite membrane with some-
what flattened cells, and that portion of the cyst membrane in
contact with the ovarian membrane shows an increase in the num-
ber of its nuclei and is more clearly differentiated as a separate
layer (see fig. 16).

With a still greater increase in size, as shown by the most
advanced ovocytes of a 9 em. larva and in later stages, the egg
presses the overlying membranes into the central cavity of the
ovary, so that the ovocyte comes to be suspended as in a sac, and
is more nearly surrounded by the cyst and ovarian membranes
(see figs. 17 to 21). In all three membranes, an increase in the

EMBRYOLOGY OF CRYPTOBRANCHUS 109

Figs. 13 to 16 Cross-sections through the wall of the ovary of a 9 em. larva of
Cryptobranchus allegheniensis. X 300. 06. v., blood vessel; c. w., cyst wall;
ep. (right), inner epithelial membrane of the ovarian wall; ep. (left) , outer epithelial
membrane of the ovarian wall; ep. c., epithelial cell of the cyst; fol., follicle cell;
ovuc., ovocyte.

Fig. 13 A cyst containing young ovocytes and epithelial cells occupies the
central part of the figure.

Fig. 14 A cyst containing slightly older ovocytes.

Fig. 15 An ovocyte surrounded by the newly-formed follicle.

Fig. 16 An ovocyte and follicle slightly more advanced than the one shown in
preceding figure. ;

110 BERTRAM G. SMITH

number of nuclei keeps pace with the increase in extent. In a
35 em. female (see figs. 21 and 22), the nuclei of the follicular
membrane are the most numerous and least flattened; those of the
eyst membrane and inner ovarian membrane are both decidedly
flattened. Somewhat rarely, the cyst membrane is ruptured
by the expansion of the ovocyte. According to King (’08) in
Bufo the rupture of the cyst membrane takes place regularly at an
early stage.

Fig. 17 Cross-section through the ovarian wall of a 9 cm. larva of Cryptobran-
chus allegheniensis, showing one of the most advanced ovocytes. X 300. Letter-
ing as in the preceding figures.

The ovocyte in the advanced growth stage is thus surrounded
by a single-layered follicle, suspended in a flask-shaped two-
layered sac of which the inner layer is the cyst membrane, the
outer layer is the inner epithelial membrane of the ovarian wall.
In a broader sense, the entire three-layered structure may be
called a follicle, and the neck of the flask-shaped sac may be called
the stalk of the follicle. This triple-layered wall persists without
any radical change in structure up to the time of maturation.

EMBRYOLOGY OF CRYPTOBRANCHUS tata

In the later stages of the development of the ovary, its walls
anastomose by the formation of cross-walls or partitions, divid-
ing the ovary into compartments or perhaps pockets; by these
cross-walls the course of the inner ovarian membrane is greatly
complicated.

The ovocyte of a female of 26 cm. and younger is apparently
a naked cell, possessing no proper membrane. In females with a

Fig. 18 Cross-section through the ovarian wall of a 26 em. Cryptobranchus
allegheniensis, showing one of the most advanced ovocytes. > 180. n., nucleolus;
v., vitelline body; ep., inner epithelial membrane of the ovarian wall. Other
lettering as in figs. 13 to 16.

body length of from 30 to 35 em. there occurs a rapid development
of two non-cellular membranes closely investing the egg within
the follicle. The inner of these two membranes exhibits a radial
striation and is the zona radiata; at the time of maturation it

becomes a simple cell wall to the egg. The outer membrane,
clear and homogeneous, is the zona pellucida; it persists as the
‘vitelline membrane’ of the embryo.

112 BERTRAM G. SMITH

The zona radiata and the zona pellucida begin to form simul-
taneously, shortly before the appearance of yolk granules. In
the most advanced ovocytes of a 35 em. female, these membranes
are well established and a narrow zone of yolk has appeared near
the periphery of the ovocyte (see fig. 22).

The zona radiata arises from the peripheral cytoplasm of the
ovocyte. In its early stages its inner boundary is not sharply
defined; its staining reaction is like that of the egg cytoplasm;.
aside from its cross-striation its structure, like that of the egg
cytoplasm, is finely granular. The zona pellucida, on the other
hand, is formed de novo as a product of cellular activity. In the
ovary of a 35 em. female its staining reaction is different from that
of any other structure present: with the borax-carmine Lyons-
blue picric-acid mixture it becomes green, while the ground-sub-
stance of the follicular, cyst and ovarian membranes stains blue.
Since, later, a membrane exactly resembling the zona pellucida
in character sometimes, though not typically, forms between the
cyst membrane and the follicle (see fig. 32), it seems reasonable
to conclude that the zona pellucida is the product of the follicle
rather than of the egg.

In the most advanced ovocytes of a spent female there is usu-
ally an increase in the thickness of the zona pellucida, while the
zona radiata shows signs of degeneration—there is a slight loss
in the distinctness of the radial striations. In adult females
taken in July and August, there is a further loss in the distinctness
of the striations of the zona radiata. In ovocytes taken just
before maturation, with the germinal vesicle close to the surface,
the zona radiata has in some eases almost lost its radial striation,
is decreased in thickness, and is becoming a simple cell wall to the
ege.

The literature on the zona pellucida and zona radiata of the
amphibian egg has been reviewed by Waldeyer in Hertwig’s (06)
Handbuch and needs no summary here.

The ovary of a young Necturus 20 em. long, killed August 25,
gives stages corresponding to those of a 35 em. Cryptobranchus.
The follicular layers and mode of attachment of the ovocyte to
the ovarian wall are practically the same as in Cryptobranchus,

EMBRYOLOGY OF CRYPTOBRANCHUS 115%

with the exception that there is a marked difference in the appear-
ance of the nuclei of the follicle proper: in Necturus these nuclei
are more numerous, and in form are spherical or even elongated in
a radial direction, instead of being flattened in the direction of the
circumference of the egg as in Cryptobranchus. The follicle
of Necturus more closely resembles that of the selachian egg in an
early stage (see Hertwig’s Handbuch, ’06, figs. 105 and 195). The
zona pellucida and zona radiata are much alike in the two urodeles;
the striations of the latter membrane are rather more distinct
in Necturus.

2. The establishment of polarity, and the progress of axial
differentiation

As already noted in the surface study of the ovarian egg, the ovo-
cyte ready for maturation shows its telolecithal character in the
presence of a superficial germinal area, in the center of which les
the germinal vesicle, while the remainder of the egg is heavily
laden with yolk. It is the purpose of the present section to trace
the changes by which this axial differentiation is brought about.

In the ovary of a 9 em. larva, vitelline bodies (see King, ’08)
are recognizable in the cytoplasm of the ovocytes in all stages
present, but are not very numerous hor conspicuous even in the
most advanced -ovocytes of such an ovary (see figs. 13 to 17).
In the largest ovocytes, the germinal vesicle is usually somewhat
excentrically situated, but with no constancy in the direction of
excentricity. Faintly-staining nucleoli are distributed quite pro-
miscuously thoughout the germinal vesicle, in the later stages with
a slight tendency toward forming a ring at the periphery.

In the most advanced ovocytes of a 26 cm. female (see fig. 18)
there is less excentricity in the position of the germinal vesicle;
the nucleoli are most numerous at the periphery. There is an
increase in the number and size of the vitelline bodies, which are
more numérous on the side toward the central cavity of the ovary.
After fixation in Zenker’s fluid, both nucleoli and vitelline bodies
take the nuclear stain, though faintly. In the ovary of a 27 cm.

114 BERTRAM G. SMITH

female, fixed in Bouin’s solution, the nucleoli take the nuclear
stain very faintly; the vitelline bodies take the cytoplasmic stain.

In the most advanced ovocytes of a 30 ecm. female (figs. 19 and
20) the germinal vesicle is quite centrally situated—a position
which it retains until a very late stage of ovogenesis. The nucle-
oli, which still stain but faintly, are nearly all at the periphery,
where they form a uniform ring. The vitelline bodies shown in
the figures now stain brilliantly with borax carmine used after

Figs. 19 and20 Sections through ovocytes and ovarian wall of a 30 em. Crypto-
branchus allegheniensis, showing the follicle and the distribution of vitelline bodies
and nucleoli. > 90. n., nucleolus; v., vitelline bodies.

Zenker’s fluid; in general they are much more numerous on the
side toward the periphery of the ovary, in the region of the future
animal pole. Some of the vitelline bodies are very large; these
usually occupy an equatorial position, but are sometimes found
on the inner side of the ovoeyte. Comparison with the preceding
stage suggests that the vitelline bodies originate or the inner
side of the ovocyte and migrate to the outer side; that they reach
their greatest development midway in the course of migration, and
break up to form the smaller and more numerous vitelline bodies

EMBRYOLOGY OF CRYPTOBRANCHUS 15)

in the region of the future animal pole. But scattered throughout
the cytoplasm are occasionally to be found other bodies, resem-
bling the vitelline bodies but more irregular in form and staining
very faintly. While it is possible that these bodies are different in
kind from the brilliantly-staining vitelline bodies, their appear-
ance suggests that they are stages in the degeneration of the latter.
The faintly-staining bodies, though seldom numerous, are more
frequently found in regions poor in deeply-staining vitelline
bodies. These observations enable us to offer an explanation of
the distribution of vitelline bodies, alternative to the theory of
migration: a wave of development of vitelline bodies, followed
by a wave of degeneration, may sweep from the inner to the outer
hemisphere of the ovocyte. But whether migration is real or
only apparent, the fact remains that the region of most abundant
deeply staining vitelline bodies has shifted from the vicinity of
the future vegetal to the future animal pole of the ovocyte. This
change is perhaps an expression of polarity; if so, it is the first
indication of polarity that I have observed. However, it is not
at all certain that polarity is not present at an earlier period;
in particular the history of the chromatin has not been sufficiently
studied, moreover it is of course possible that a physiological
polarity of the cell may precede its manifestation in a visible
form.

In the ovary of a 34 em. female, fixed in Flemming’s solution,
the distribution of vitelline bodies is much the same as noted in
the 30 em. female; in form the vitelline bodies are sometimes oval
or irregular, but never mulberry-shaped as is sometimes the case
with Zenker’s.

In the ovary of a 35 em. female, yolk granules are beginning to
form in the most advanced ovocytes; other ovocytes nearly as
large contain no yolk. In neither of these two stages are vitel-
line bodies in the typical condition present, but they are sometimes
found undergoing a process of degeneration—they lose the inten-
sity of their staining reaction, become irregular in form, and dis-
appear. The disappearance of the vitelline bodies at the time of
the formation of yolk suggests a correlation between the two phe-
nomena; but so far as I have been able to observe, the final stages

_ JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1

116 BERTRAM G. SMITH

in the disappearance of the vitelline bodies are not closely asso-
ciated with the formation of yolk granules, nor have I found any
undoubted ‘yolk nuclei,’ such as have been described by King
(08) for Bufo. In view of the diversity in the methods of yolk-
formation described for different amphibians, this result is not
altogether surprising.

Fig. 21 Section through an ovocyte and ovarian wall of a 35 em. Cryptobran-
chus allegheniensis, showing the follicle and the distribution of nucleoli. In this
ovocyte the vitelline bodies have disappeared, but yolk-formation has not yet
begun. X 60. n., nucleolus.

In the more advanced ovocytes of a 35 cm. female (see fig. 21).
the nucleoli stain deeply with borax carmine used after Zenker’s
fluid. There is usually a marked concentration of the nucleoli
on the side of the germinal vesicle toward the periphery of the
ovary. Account must be taken of the fact that shrinkage of
the germinal vesicle also proceeds, as a rule, most extensively
on this side, leaving a large space, while the opposite side remains
in contact with the cytoplasm. This greater shrinkage on the

EMBRYOLOGY OF CRYPTOBRANCHUS Wf

outer side in part accounts for the greater concentration of nucleoli
on this side, but it is inadequate to account for all of it; moreover
the axis of excentricity in form due to shrinkage does not always
correspond accurately to the axis of excentricity in the arrange-
ment of the nucleoli.

This excentric distribution of material marks an axis which
corresponds, roughly at least, to the polar axis at the time of
maturation; the nucleoli accumulate on the side which is to become
the animal pole, and thus perhaps afford a second indication of
polarity. King (’08) found this condition in Bufo at the time when
the nucleus was moving from the center of the egg to the animal
pole, and suggested the possibility that the accumulation of most
of the nucleoli in one part of the nucleus might have something to
do with this movement. In Cryptobranchus this concentration
of the nucleoli begins long before the migration of the germinal
vesicle to the surface, and indeed before the formation of any
yolk; it is most marked in the advanced ovocytes of a 35 cm.
female, when the yolk is just beginning to form. As will appear
from the study of later stages, this arrangement of the nucleoli
does not persist during the actual migration of the germinal vesi-
cle; nevertheless the early occurrence of axial concentration of
nucleoli is significant.

In the ovary of a 35 cm. female, we find that occasionally,
through the folding of the ovarian wall, an ovocyte has been
thrust deep into the central cavity and has come in contact with
the nutrient ovarian wall of the opposite side. The side opposite
the stalk of the follicle now becomes the side best nourished, and
here the nucleoli accumulate. Thus nature’s experiment shows
that the accumulation of nucleoli, and perhaps polarity, is not
something predetermined in the egg, or even fixed by the relation
of the egg to the ovarian wall within which it develops, but is
a phenomenon depending upon larger environmental relations
which probably have to do with nutrition; for as a consequence
of the changed position of the egg the nucleoli accumulate on the
opposite side from that favored by the original environment.

In the ovocytes of immature females with body lengths of from
35 to 38em., the yolk first appears ina narrow zone nearthe periph-

118 BERTRAM G. SMITH

ery and parallel to the newly-formed zona radiata, but sepa-
rated from the latter by a narrow layer of clear cytoplasm (see
fig. 22). At this time the ovocyte has a diameter of from 1.5 mm.
to 2mm. The yolk zone is divisible into two layers, an outer

Fig. 22 Portion of a section through one of the most advanced ovocytes of a 35
em. Cryptobranchus allegheniensis, showing structure of the membranes sur-
rounding the egg and the distribution of yolk granules. X 340. Thestripshown
extends about half-way to the germinal vesicle. c., cyst membrane; cy., yolk-
free peripheral zone of cytoplasm; ep., inner epithelial membrane of the ovarian
wall; fol., follicular membrane proper; z. p., zona pellucida; z. r., zona radiata;
y. and y’., layers of fine and coarse yolk granules respectively.

layer of fine yolk particles and an inner layer of coarse yolk par-
ticles, separated by a narrow region poor in yolk.

In the largest residual eggs (2 to 3 mm. in diameter) of spent
females, the yolk-laden zone has extended inward further than

EMBRYOLOGY OF CRYPTOBRANCHUS 119

outward; a very narrow zone of clear cytoplasm persists at the
periphery, and a much broader zone containing only a few scat-
tering yolk granules surrounds the germinal vesicle. The middle
portion of the yolk zone is now filled with coarse yolk granules;
its margins consist of fine yolk particles. The germinal vesicle
is still centrally situated, and the arrangement of yolk zones and

Fig. 23 Meridional section through one of the most advanced ovocytes of an
adult Cryptobranchus allegheniensis killed July 6. The bounding line repre-
sents the zona radiata. X 20.

cytoplasm is concentric. The nucleoli are still distributed at the
periphery of the germinal vesicle, with only a slight tendency
toward concentration at the outer side.

In the largest ovocytes of adults killed July 6 (see fig. 23),
the germinal vesicle occupies a position midway between the
center of the egg and the periphery, on the more exposed side of
the egg, toward the stalk of the follicle. The animal pole is thus

120 BERTRAM G. SMITH

defined by a point on the surface, toward which the germinal
vesicle is moving. The cytoplasm is now everywhere thickly
interspersed with yolk granules; these granules are in general
coarse throughout the central portion of the egg, finer and
more densely packed at the periphery. Axial differentiation in
the arrangement of yolk particles is now for the first time evi-
dent in a slight thickening of the peripheral layer of fine yolk
particles in the region of the animal pole. This region is also
somewhat richer in cytoplasm than the remainder of the egg.
There is thus present the beginning of a germinal disc or blasto-
disc, which in later stages becomes visible in surface views as
the germinal area.

In the vegetal hemisphere a region of particularly fine and
dense yolk, crescent-shaped in meridional section, lies: mid-way
between the center of the egg and the periphery. This region I
shall call the ‘yolk cup.’ Its appearance suggests that it may be
a part of a once continuous zone completely enclosing the germinal
vesicle, and that, in the animal hemisphere, this zone has been
interrupted in consequence of the migration of the germinal
vesicle toward the surface. Probably the yolk cup is the physio-
logical equivalent of the concentric layers of dense fine yolk found
in the egg of the hen and various other vertebrates. Riddle (’11)
has shown that the alternate layers of yellow and white yolk in the
hen’s egg are due to a daily rhythm in nutrition; he has advanced
the same principle in explanation of the concentric layers of
yolk in the eggs of certain cyclostomes, selachians and reptiles.
In Cryptobranchus, from comparison with ovarian eggs taken in
the autumn after the close of the spawning season, it is evident
that in the stage under consideration the yolk cup marks the limits
of growth during the preceding winter; hence it seems very prob-
able that the yolk cup is the result of a seasonal variation in
nutrition, and represents a layer added during the winter months.

The nucleoli are still found mainly at the periphery of the ger-
minal vesicle, but with no constant tendency toward concentra-
tion in an axial position.

It has been noted that the animal pole, as defined both by the
center of the germinal dise and the point on the surface toward

EMBRYOLOGY OF CRYPTOBRANCHUS eA

which the germinal vesicle is moving, lies in general on the more
exposed side of the egg, within the stalk of the follicle. The ani-
mal pole thus lies iv. the opposite direction from that assumed in
the ovarian egg of the hen (Lillie, ’08, p. 29). According to King
(02) in the great majority of cases the egg of Bufo is attached in
the equatorial region by the stalk of the follicle.

From a comparison of this stage with the preceding one (the
ovocytes of a spent female), it is evident that yolk-formation pro-
ceeds concentrically about a centrally situated germinal vesicle
until the egg is nearly or quite filled with yolk, and that axial
differentiation in the arrangement of yolk particles does not
appear until a very late stage of ovogenesis, two or three months
before maturation. It is further apparent that the germinal vesi-
cle attains its final position, not through unequal growth of the
cytoplasm or excessive accumulation of yolk on the other side of
the egg, but by a process of migration.

In the ovocytes of adults taken July 20, the germinal vesicle
has migrated further toward the animal pole; it lies about one-
third of the distance from the surface to the center of the egg.
Both nucleoli and chromosomes are now aggregated at the center
of the germinal vesicle. The yolk-cup persists, and there is an
increase in the extent of the germinal disc. In some eggs a small
cone-shaped mass of dense cytoplasm, with the apex of the cone
pointing inward, lies immediately beneath the germinal vesicle.

In the ovary of an adult female killed August 17, the egg (fig.
24) has nearly reached its maximum size before fertilization;
a meridional section cut in paraffin has a diameter of about 6
mm. (It should be noted that a yolk-laden egg does not shrink in
paraffin to the same extent as ordinary tissues). The germinal
vesicle lies only a short distance from the surface, and is bounded
on the side toward the center of the egg by a large cone-shaped
mass of cytoplasm. The apex of this cone is continuous with a
slender meshwork of less dense but yolk-free cytoplasm extending
half-way to the center of the egg. Owing to a slight obliquity of
the slender cytoplasmic mass, it has not been found complete in
any one section; in fig. 24 it has been added, from adjacent sec-
tions, to the one chosen for the remainder of the drawing.

122 BERTRAM G. SMITH

Immediately beneath the zona radiata lies a peripheral layer of
yolk-free cytoplasm, which from analogy with the teleost egg I
shall call the ‘protoplasmic mantle.’ In the region of the vegetal
pole this is so thin as to be barely recognizable with a magnifi-
cation of 500 diameters; in the region of the animal pole it is thick-
ened to form a dise which I shall call the ‘cytodise.’ At the ani-

Fig. 24 Meridional section through an ovarian egg of an adult Cryptobranchus
allegheniensis killed Aug. 17. X 20. cy., eytodisc; y. d., yolk disc.

mal pole the cytodise reaches its maximum thickness of about
15u—a little thicker than the layer of follicle cells proper.

The remainder of the egg is filled with yolk. Underlying the
cytodise and occupying an area about 100° in diameter surround-
ing the animal pole, is a thick layer of fine but dense yolk which I
shall call the ‘yolk dise.’ The eytodise and yolk dise combined

EMBRYOLOGY OF CRYPTOBRANCHUS eS

represent the anlage of the germinal dise or blastodise, which later
comes to enclose the germinal vesicle and the cytoplasm accumu-
lated beneath it. Elsewhere a very thin peripheral layer of fine
yolk particles, continuous with the yolk disc, lies immediately
beneath the protoplasmic mantle. The interior of the egg shows
no particular change in the yolk.

The germinal vesicle is spherical when perfectly: preserved ;
when flattened this is due to shrinkage. The ground-substance
or nuclear sap appears homogeneous under a low power, but with

Fig. 25 Central portion of the germinal vesicle represented in the preceding
figure, enlarged to show details. The finely granular ground-substance of the ger-
minal vesicle is not shown. X 340. c. g., chromatin granules; chr., chromo-
somes; nu., nucleolus.

a magnification of 500 diameters it exhibits an extremely fine but
dense granular structure. The nucleoli are now nearly all aggre-
gated at the center; some few persist at the periphery, particularly
on the side toward the center of the egg. The chromosomes are
for the most part confined to the central part of the area occupied
by the nucleoli.

Among the nucleoli, though not closely associated with them,
there are now found very numerous and minute granules which

124 BERTRAM G. SMITH

stain like chromatin (see fig. 25). These are evidently formed in
close association with the chromosomes; in earlier stages chromo-
somes have been found covered with these granules before the
latter have appeared elsewhere.

In the ovarian eggs of an adult killed August 22, the most
marked changes in the general topography as viewed in meridional
sections are a slight advance in the migration of the germinal vesi-
icle toward the surface, and an increased thickness of the periph-
eral zone of fine yolk particles, particularly in the yolk dise.
In the vegetal hemisphere the protoplasmic mantle is no longer
recognizable as a separate layer; its constituents have mingled
with the peripheral layer of fine yolk particles. The cytodise
is reduced in thickness by the blending of its inner surface with
the yolk dise.

The narrow path of cytoplasm leading toward the center of the
egg from the apex of the cone of cytoplasm underlying the germi-
nal vesicle has disappeared; likewise the yolk cup is, as arule, no
longer present. In this stage there is a slight increase in the num-
ber of chromatin granules dispersed amongst the nucleoli; other-
wise the nuclear contents seem unchanged.

Ovaries taken during the last week in August and the first week
in September usually contain some eggs with the germinal vesicle
appearing at the surface. In the general organization of the egg
before the germinal vesicle actually reaches the surface, there are
few changes from the condition described for August 22. Fig.
26 shows the general topography of an egg with the germinal vesi-
cle very close to the surface. The cone of cytoplasm underlying
the germinal vesicle is beginning to mingle with the yolk; it is not
present in the section figured. Within the germinal vesicle the
nucleoli are massed more closely together at the center; there is an
increase in the number of chromatin granules, and apparently a
gradual disappearance of the chromosomes—in some eggs they
could not be found.

At the close of the period considered, axial differentiation is
evident in the following arrangement of material: (a) the excen-
tric position of the germinal vesicle and the cone-shaped mass
of cytoplasm underlying it; and (6) the formation about the animal

EMBRYOLOGY OF CRYPTOBRANCHUS 25

pole of a germinal disc or blastodisc consisting of two layers, a
very thin peripheral layer of yolk-free cytoplasm which has been
called the cytodise, and underlying this a thick lenticular layer of
mingled cytoplasm and dense fine yolk which has been called the
yolk dise. At any given level the egg is radially symmetrical
about the axis of polarity. In general the egg has progressed from
an alecithal through an isolecithal to a telolecithal stage.

Fig. 26 Meridional section through an ovarian egg of an adult Cryptobranchus
allegheniensis killed Sept. 6, 1910, showing organization just before the germinal
vesicle reaches the surface. XX 20.

The establishment of polarity, with axial differentiation, is an
event of great morphogenetic importance, since the formative
materials for the embryo are being segregated in the vicinity of the
animal pole. Through later changes in the distribution of this
material the animal pole comes to mark the anterior, the vegetal
pole the posterior end of the future animal; hence the establish-
ment of polarity defines the principal axis of the embryo.

126 BERTRAM G. SMITH

The changes that immediately follow—the appearance of the
germinal vesicle at the surface, the rupture of its membrane, and
the reorganization of the germinal dise with the incorporation of
materials brought from the interior of the egg by the nucleus—
lead up to maturation and will be considered in the account of
that process.

3. Resorption of ovocytes; the follicle cells in a phagocytic réle.

In young females nearing maturity (about 38 em. body length),
a few ovocytes reach an advanced stage of development, becoming
filled with yolk and attaining a size nearly as great as the ovocytes
of an adult. These precocious ovocytes fail to undergo matura-
tion changes, and during the breeding season begin to degenerate,
or rather to be resorbed, together with some of the less advanced
ovocytes only partially filled with yolk. Viewed in the living
ovary, these degenerating ovocytes are colored a very bright yel-
low or orange. Digestion and absorption of the yolk granules
is accomplished through the medium of the cells of the follicular
layer proper, which become greatly enlarged and function as
phagocytes, thereby reversing their usual réle as nurse cells to
the egg.

The first step in the process of degeneration of the ovocyte
is the disappearance of the zona radiata; the later stages are
illustrated by figs. 27 to 30. The follicle cells enlarge, by increase
both in the size of the nucleus and in the amount of cytoplasm.
The zona pellucida is ruptured; at the same time it becomes irreg-
ularly thickened, a circumstance which may be interpreted either
as a shortening of the fragments due to the release of tension,
or as a step in the process of dissolution. The rupture of the
zona pellucida allows the yolk to come in contact with the follicle
cells; the latter engulf the yolk particles, and become surrounded
by thin walls. About this time the zona pellucida disappears,
and the follicle cells are left as large yolk-filled cells resembling
columnar epithelium, forming a continuous layer around the egg.

Digestion of the yolk particles is completed first at the outer
margin of the follicle cells, while the inner margin continues to
engulf yolk. The included yolk granules stain less deeply with

EMBRYOLOGY OF CRYPTOBRANCHUS p27

: 3 Ss fi
aed POC BOL) y {
CON PAS OS EOE CES

e

()

28

30

Figs. 27-30 Changes at the periphery of an ovocyte in the process of resorp-
tion; the follicle cells are shown in a phagocytic réle. Fig. 27, read from left to
right, shows the beginning of the process (compare with fig. 22 showing the nor-
mal condition of the follicle). The remaining figures, taken from different ovo-
cytes, show successively later stages. X 244. y., yolk; z. p., zona pellucida.

128 BERTRAM G. SMITH

haemotoxylin than the others, and in sections stained with the
borax-carmine lLyons-blue picric-acid mixture the included
granules take the cytoplasmic stain more deeply.

The follicle cells later become greatly elongated, and the cyto-
plasm takes the form of a faint meshwork with large spaces.
Ingestion of yolk continues at the inner ends of the cells, while
the remainder of the cell functions as a long tube to convey the
products of digestion to the periphery. The follicular layer
remains one cell in thickness until the cells have reached a length
of about 250 u; with a further increase in thickness it becomes
broken up into a meshwork of cells, amongst which are numerous
capillaries. Ovocytes have been found in which this meshwork
of cells reaches nearly to the center, and the remaining yolk is very
small in amount.

In the adult female occasional eggs, though of full size, fail to
escape from the ovary. Judging from their external appearance
these ovocytes undergo resorption in the manner just described.
A somewhat similar process of resorption has been described in
the eggs of cyclostomes and fishes (Biihler, ’02).

4. The organization of the egg shortly before the appearance of the
germinal vesicle at the surface

At this point it may be well to summarize briefly the condition
of the ovocyte in the stage immediately preceding the appear-
ance of the germinal vesical at the surface (see fig. 26).

The egg lies within a triple-walled follicular sac whose cellular
membranes have undergone little change since they first became
well established. The stalk of the follicle, and in general the
animal pole of the egg, lie toward the periphery of the ovary.

The zona pellucida persists unchanged, except for a slight in-
crease in thickness; the zona radiata shows signs of atrophy, and
in some cases is assuming the character of a simple cell wall.

The nucleus or germinal vesicle has migrated from the center
of the egg to a position near the periphery, ordinarily on the side
toward the stalk of the follicle. During the migration of the
germinal vesicle a cone-shaped mass of dense cytoplasm has

EMBRYOLOGY OF CRYPTOBRANCHUS 129

collected beneath it, and is now beginning to mingle with the sur-
rounding yolk.

A germinal disc or blastodise is evident in surface views of living
material as a circular area, lighter in color than the rest of the egg,
about 60° in diameter and situated on the more exposed side of
the egg. Inmeridional sections it is shown to consist of two layers:
a thin peripheral layer of cytoplasm, the cytodise; underlying this
a thick lenticular mass of mingled fine yolk particles and cyto-
plasm, the yolk disc. The germinal vesicle lies at the center of
the yolk disc.

The yolk dise is continuous with a thin peripheral layer of fine
yolk granules, mixed with cytoplasm, which lies in contact with
the zona radiata everywhere except in the region of the cytodise.

The remainder of the egg is filled with coarse yolk granules
mingled with fine yolk granules and a small amount of cytoplasm.

The nucleoli are grouped at the center of the germinal vesicle,
and amongst them are numerous chromatin granules. In some
eggs chromosomes are found at the center of the group of nucleoli,
in others the chromosomes have disappeared.

A point on the surface overlying the center of the germinal
vesicle marks the animal pole. The general arrangement of
materials is radially symmetrical about the axis of polarity, with
differentiation proceeding in the direction of this axis.

B. MATURATION
1. The germinal vesicle at the surface

Meridional sections through ovocytes with the germinal vesicle
at the surface show little change in the details of structure from
the condition previously described. The germinal vesicle is
usually somewhat flattened against the periphery, and a portion
of its surface is in direct contact with the zona radiata. Masses of
a wavy fibrous material are occasionally found in the nuclear
sap. A few fragments of chromosomes are present in some eggs;
in others no chromosomes have been found. The nucleoli and
chromatin granules persist at the center of the germinal vesicle.

130 BERTRAM G. SMITH

2. The dissolution of the germinal vesicle, and the formation of the
jirst polar spindle

Material for the study of this stage was obtained from two
females in which the majority of the ripening eggs had left the
ovary, and were found distributed in the body cavity, oviduct
and uterus. The nearly mature eggs left in these ovaries were
found in every case investigated (nine eggs were sectioned) to
have the germinal vesicle ruptured and its constituents well
mixed with those of the blastodisc; in the majority of cases the
first polar spindle had already formed.

Fig. 31 Meridional section through an ovarian egg of Cryptobranchus alle-
gheniensis, shortly after the rupture of the germinal vesicle. Fragments of the
germinal vesicle are seen scattered throughout the blastodisc. x 18. p. s. J,
first polar spindle.

The rupture of the germinal vesicle and the distribution of its
materials throughout the blastodise must take place with consider-
able rapidity, since in eggs sectioned only the beginning and the
end of the process have been observed. Fragments of the nuclear
membrane, together with the wavy fibrous material previously
noted in the germinal vesicle and innumerable fine granules,
probably derived from the cell sap, become widely scattered
throughout the germinal disc (see fig. 31). During this process of
disintegration of the germinal vesicle the nucleoli and chromatin
granules are lost to view. It seems improbable that all the chro-
matin granules should again be segregated as nuclear material;
at any rate the rupture of the germinal vesicle affords an oppor-

EMBRYOLOGY OF CRYPTOBRANCHUS lie

tunity for the contribution of important nuclear material to the
cytoplasm.

The germinal dise or blastodise no longer shows a division into
two layers; the material of the cytodisc is intimately mingled with
that of the yolk dise. The cone of cytoplasm following the germi-
nal vesicle in its migration is likewise more or less thoroughly
incorporated into the blastodise.

The end result of the migration of the germinal vesicle to the
surface and its disintegration in that situation is now apparent.
All the material of the nucleus and a considerable amount of cyto-
plasm have been brought from the interior of the egg to the vicin-
ity of the animal pole, fragmented, and the débris more or less
scattered throughout the blastodise. Out of this complex there
soon emerges close to the surface at the animal pole the recon-
structed nucleus in the formof the first polarspindle. Onefunction
of migration is doubtless to get this nuclear material to the periph-
ery where a part of it may be disposed of in the maturation divi-
sions. A further adaptation is found in the fact that, later, the
egg-nucleus or female pronucleus is left in the center of the forma-
tive material of the blastodise. A third end attained by migra-
tion is that the formative material of the blastodise is added to by
cytoplasm following the germinal vesicle, and also by substances
derived from the germinal vesicle itself.

The zona radiata has become reduced in thickness, has lost its
striation and no longer shows a distinct limiting inner surface—-
its inner margin is irregular or blends with the peripheral cyto-
plasm of the egg. When the egg is shrunken away from the zona
pellucida the zona radiata usually remains organically connected
with the egg. The character of the zona radiata has changed so
radically that it will no longer be referred to by this name; it has
become a simple cell wall to the egg, and as such takes part in the
later process of cleavage.

The zona pellucida persists unchanged as the so-called vitelline
membrane of the egg at the time of fertilization and during the early
stages of embryonic development.

As in other amphibian eggs, only these two membranes, the
Zona pellucida and the cell wall formed from the zona radiata,

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1

32 BERTRAM G. SMITH

accompany the egg in its escape from the ovary; the process of
ovulation involves the rupture of the follicle which remains in the
ovary.

The occurrence of the first polar spindle was studied in two
females, (A) and (B), in which the eggs were distributed from
ovary to uterus inclusive. The first polar spindle was found in
eggs taken from the following situations: ovary, body cavity,
oviduct, and extreme upper part of the uterus; out of a total of
twenty-eight eggs studied, the first polar spindle was found in
thirteen cases. Five eggs taken from the lower uterus were studied ;
no first polar spindle was found.

In the case of another female (C), in which the eggs were all in
the uteri, no first polar spindle was found in three eggs sectioned.

Allowance must be made for the fact that in some cases in which
the first polar body is absent it may have been missed on account
of imperfections in the series. The results are sufficient to jus-
tify the conclusion that the first polar spindle is usually present
at the time the egg leaves the ovary and during its passage down
the oviduct, and that it disappears about the time the egg reaches
the uterus.

The first polar spindle (see figs.” 32 to 35) is formed with its long
axis either coinciding with the axis of polarity of the egg, or oblique
to this axis. The number of chromosomes is probably twelve
before any of them have divided. There is an outer ring of six
large chromosomes, surrounding a central group of six small chro-
mosomes usually found in a state of division; it is probable that
these six small chromosomes are not all of equal size. These
size differences of the chromosomes are interesting in the light of
well-known recent work indicating individual differences in the
chromosomes of many forms.

There is frequently present close to the cell wall overlying the
spindle a dise-shaped body with an irregular cross-striated struc-
ture, which, from its probable mode of origin, I shall call the ‘con-
tact dise’ (see figs. 33 and 34). This disc takes the cytoplasmic
stain, and seems to be of the same composition as the cell wall.
The adjacent cell wall is slightly thickened and sometimes shows
a cross-striation, reminding one of the zona radiata (compare the

EMBRYOLOGY OF CRYPTOBRANCHUS 33

effect on the cell wall of penetration by the spermatozoon, de-
scribed later). The presence of the contact disc is uniformly
accompanied by a deficiency of the spindle, which lacks an aster
at the end nearest the disc. In a few cases there seems to be a
small amount of sphere substance underlying the contact disc.
The inference seems to be that the contact disc is the product of
the aster of the first polar spindle modified by contact with the

32 33 34 35

Figs. 32-35 Meridional sections showing first polar spindle of Cryptobranchus
allegheniensis. Figs. 32 to 34 are from ovarian eggs; fig. 35 is from an egg taken
from the lower part of the oviduct. X 340.

Fig. 32 c. w., cell wall, formed from the zona radiata. z. p., zona pellucida.

Fig. 33 c.d., contact disc.

Fig. 34 The section cuts the spindle obliquely and includes all the chromatin
except one small chromosome belonging to the central part of the group, which is
left in an adjacent section. There are probably six large chromosomes forming a
ring, surrounding six small chromosomes in a state of division.

Fig. 35 A considerable part of the chromatin is left in an adjacent section.
There are probably six large and six small chromosomes, arranged much as in the
preceding figure.

cell wall. The function of the disc, if it have any function, may
be to anchor the spindle at the surface during the pulling-apart
of the two sets of chromosomes. Unfortunately for this hypothe-
sis the linin threads have not been traced from the chromosomes
of the first polar spindle to the contact disc; but since the latter
structure is never found except in conjunction with a polar spindle,
there is no escape from the conclusion that it is in some way related
boOvit.

134 BERTRAM G. SMITH

Sections afford no explanation of the faint dark spot or shallow
depression noted in surface views of the animal pole after the rup-
ture of the germinal vesicle and before the formation of the second
polar spindle. An actual depression overlying the first polar
spindle is rarely found in sections; if present in the living egg it
must be lost through shrinkage of the egg during the process of
preparation for sectioning by the paraffin method.

The yolk granules immediately adjacent to the cytoplasm
surrounding the spindle are distinctly larger than at the same level
elsewhere; they are doubtless brought from a deeper situation by
the migration of the nucleus.

The anaphase of the first polar spindle has not been observed,
and the first polar body has been found only in a state of degener-
ation, in conjunction with the second polar spindle.

3. The second polar spindle

The second polar spindle (see figs. 36-38) may be distinguished
from the first by the smaller amount of chromatin material, and
by the fact that a well-defined pit already noted in surface views
usually lies above it. This pit sometimes disappears in the late
anaphase of the spindle.

The débris of the first polar body is usually found at the bottom
or sides of the pit, outside of the cell wall; in some cases fragments
of its chromatin are found mingled with the contact disc of the
first polar spindle. The chromatin fragments stain but faintly
with borax carmine.

The contact disc of the first polar spindle has fused with the
thickening of the cuticle which overlies it. In the telophase of
the second polar spindle a new contact dise is formed which soon .
fuses with the old. In some cases linin threads have been clearly
traced from chromosomes to the contact dise of the early second
cleavage spindle, thereby sustaining the view of the origin of the
contact dise set forth in the account of the first cleavage spindle. -

No second polar spindle has been found in any eggs of the two
females, (A) and (B), for which the occurrence of the first polar
spindle was tabulated. This indicates that the second polar
spindle is not formed until after the eggs have been for some time

EMBRYOLOGY OF CRYPTOBRANCHUS 135

Oa 022. Of
a) eS o. fj 2 Oheok oO : t as ‘

I) ip é ‘07 [om (0) 0; 0; ie | J; y
nee ncpatuere® Suis ee oO I OBG85
OIEOLO DISA G EY ETO eOaOY SOS AOS 5308

38 39 40

Figs. 36t040 Meridional sections showing the second polar spindle, and the for-
mation of the second polar body and the egg-nucleus in Cryptobranchus allegheni-
ensis. X 340.

Fig. 36 Section through the second polar spindle of an unfertilized egg taken
from the uterus of aripefemale. z. p., zona pellucida; p. 6. J, débris of the degen-
erating first polar body. ‘

Fig. 37 Section through the second polar spindle of an egg killed 15 minutes
after fertilization. The section lies in the plane of the equator of the spindle.

Fig. 38 Late anaphase of the second polar spindle in an egg killed 23 hours after
fertilization. A considerable part of the chromatin is left in an adjacent
section.

Figs. 39and 40 Two consecutive sections through an egg killed 5 hours after
fertilization; the first figure shows the second polar body, the second figure shows in
addition the newly-formed egg-nucleus.

136 BERTRAM G. SMITH

in the uterus. In the third female (C) considered, in which all the
eggs were in the uterus, a second polar spindle was found in one
out of the three eggs sectioned. This result is sufficient to show
that sometimes, if not always, the second polar spindle is formed
while the egg is still in the uterus, previous to fertilization;
hence the penetration of the egg by the spermatozoon is not
requiredas a stimulus to the formation of the second polar spindle.

The question arises whether the second polar spindle is normally
or ever present after the penetration of the egg by the spermato-
zoon; in other words, do the processes of maturation and fertiliza-
tion overlap? We must first take into consideration the possi-
bility that eggs dissected from the uterus of a ripe female for pur-
poses of artificial fertilization may not be quite so far advanced as
eggs spawned and fertilized in a natural manner. Fortunately
it has been possible to check results obtained through artificial
fertilization by comparison with a case in which fertilization oc-
curred in a more natural manner. For the study of fertilization
three females, (C, D, and EF), were principally used; eggs from the
gravid uteri of the first two were artificially fertilized in the usual
way; the third female spawned with a ripe male while the two
were being carried in a pail of water.

Furthermore we must distinguish between what might be ealled
potential fertilization, the mere contact of the seminal fluid
with the gelatinous envelopes of the eggs, and actual fertilization,
the penetration of the egg proper by the spermatozoon. While
the act of fertilization is not consummated until the fusion of the
germ-nuclei, the influence of the spermatozoon is felt in many ways
as soon as it enters the egg cytoplasm, so that actual fertilization
may be said to begin as soon as the spermatozoon pierces the cell
wall of the egg. The time record is almost necessarily reckoned
from the moment of mixing of the two sexual elements, or poten-
tial fertilization; actual fertilization follows after an interval
necessary for the passage of the spermatozoon through the gela-
tinous envelope, which varies for the individual eggs and espe-
cially for eggs of different spawnings fertilized by different males,
and which can be determined only by a careful microscopical
examination of serial sections of each egg.

EMBRYOLOGY OF CRYPTOBRANCHUS 137

Out of twenty-one eggs from three females (C, D and £),
preserved at intervals extending from fifteen minutes to two and
one-half hours after fertilization, a second polar spindle was found
in eighteen cases, and one or more spermatozoa were found in
each of eleven eggs. The sectionsshow that thespermatozoon may
pierce the cell wall of the egg as early as fifteen minutes after con-
tact with the outer envelopes, though a longer time is usually
required.

Making allowance for faults of technique we may say that the
second polar spindle is usually and probably always present from
the time of fertilization up to two and one-half hours later, reckoned
from the moment of mixing the sexual elements; there is no essen-
tial difference in this respect between eggs artificially and naturally
fertilized.

Only early stages of the second polar spindle are found in eggs
up to and including one and one-half hours after fertilization;
exclusively anaphase stages are found in eggs taken one and
three-quarters to two hours after fertilization; the formation of
the second polar body and the egg-nucleus (see figs. 89 and 40)
is confined to a period between 4 and 8 hours after fertilization.
While a stimulus from the spermatozoon is not required to initiate
the formation of the second polar spindle, it is evident that the
later stages of this mitosis are passed through only after fertiliza-
tion; in other words, the processes of maturation and fertilization
overlap. Hertwig (’06) makes the general statement that in na-
ture the time of fertilization of the amphibian egg falls between
the formation of the first and second polar spindles.

4. The organization of the egg immediately before fertilization

At the time of spawning the egg is surrounded by the unchanged
zona pellucida or vitelline membrane; within this is a thin cell
wall, the transformed zona radiata, which is organically connected
with the egg.

There are few changes in the general appearance of the blasto-
dise since the condition described shortly after the rupture of the
germinal vesicle (see fig. 31). There is a more intimate incorpor-
ation of the materials of the germinal vesicle into the substance of

138 BERTRAM G. SMITH

the blastodise; shreds of non-formative material, such as frag-
ments of the nuclear wall and the fibrous material of the germinal
vesicle, are each surrounded by a closely adherent film of eyto-
plasm and are being absorbed. In eggs ready for fertilization
the second polar spindle is sometimes, though perhaps not always,
fully formed; it lies beneath a sharply-defined pit at the bottom
of which may be found the débris of the first polar body.

The peripheral zone of fine yolk particles in the vegetal hemi-
sphere remains as described in the late ovarian egg.

C. FERTILIZATION
1. The history of the egg-nucleus

The formation of the egg-nucleus is shown in figures 38 to 40;
the process is usually complete about five hours after fertilization.
About ten and one-half hours after fertilization (see figs 47 and
48) the egg-nucleus has increased in size and sunk into the blasto-
disc to a point one-third as far from the surface as the position later
occupied by the, copulation-nucleus (see fig. 52). A yolk-free
region, partly filled with cytoplasm, extends from the egg-nucleus
for a short distance toward the surface, indicating the path of
migration (fig.48). Atthis time the egg-nucleus stains but faintly.

Figs. 41 to 43 Vertical sections of eggs of Cryptobranchus allegheniensis,
showing penetration of the egg by the spermatozoon. X 240.

Fig. 41 From an egg killed 24 hours after fertilization. This figure is a recon-
struction from two adjacent sections: the upper half of the figure is drawn from one
section, the lower half from the other. The spermatozoon shown in the figure has
entered the egg about 50° from the animal pole where the second polar spindle,
shown in fig. 38, is in the late anaphase stage. Another spermatozoon in the same
condition as the one figured has entered the opposite side of the egg a little below
the equator.

Fig. 42 From an egg killed 3 hours after fertilization. The spermatozoon
figured has entered the egg a little above the equator. This egg contains in all
ten spermatozoa.

Fig. 43. From anegg killed 5 hours after fertilization. The arrow indicates the
direction of the path of the spermatozoon which has entered the egg about 30°
from the animal pole. The distance from the surface of the egg to the head of the
spermatozoon is about one and one half times as great as in the preceding figures.
The head of the spermatozoon is shown entire in this section; the tail persists in a
somewhat abbreviated condition, but is not shown in the section figured. This
egg contains another spermatozoon in the same condition.

139

EMBRYOLOGY OF CRYPTOBRANCHUS

i
D
18

9,700
aS

a SSSR Se Rese a <i

as SOS SOSSOSSS OOS SSSHSMWOSER. S o or
Oe ee es Ss ap pesearoaony
ates S -) We KOS Ww OX Soe

RAST S TOSS
Se BESO ¥q0)
SSRN

2 5%

og
02
1)

%

Oi

; 02, 205
KB

00
(a)

ON;

° = ?) os oS)
a RES
‘ Soe © ry ~ Sy =
Z OSSQOIG se Ewe, SOS OS fe
SOSMSRDVIVAS SS X S IF fs} : :
SESS OSS SEO SS BSSEN Ness Se ONS
Ue ONS ESSOCRT SS NN) SSO CSS WSSSV_OsS $8069 ren Oe E06 Ch)
@ fy st
4 i

BROOBIAQONTIOCOS 20 29>

. eae Picaete cease

=WSVOr lent iS S

=" eee eee Se
5 SSS ays ee

Xo)

Rarer se

)

ASH Ri ~~ Ss deo
en eee

= = Ae pees ©
= Rei O 8 S
= 6 eer i)
rs Y SIO, sie SEQ
y 50 WC) os

z Smee

9995
0
Oo SO

43

140 BERTRAM G. SMITH

2. History of the sperm-nucleus

(a). Penetration of the egg by the spermatozoon. As previously
noted, spermatozoa may be found entering an egg taken as early
as fifteen minutes after fertilization. In describing the process,
we may best begin with an egg taken about two and one-half
hours after fertilization (see fig. 41).

The zona pellucida or vitelline membrane is not affected fur-
ther than by the formation of a minute perforation which can
only rarely be found in sections. The zona pellucida is omitted
in the figures.

The cell wall of the egg becomes greatly thickened around the
perforation made in it by the spermatozoon. The thickened
region is conical in form, with the apex of the cone pointing in-
ward; its outer and central portions are cross-striated. The per-
foration persists as a conspicuous pore lying in the axis of the cone.
The entire structure greatly resembles a micropyle. i

Beneath this pseudo-micropyle the path of the spermatozoon
is clearly indicated by a yolk-free cytoplasmic region. The form
of this region, and the attitude assumed by the spermatozoon
itself, indicate that the course pursued by the spermatozoon is a
spiral one, with the axis of the spiral lying in a radial direction.

The spermatozoon at this time retains practically its normal
form. As in Axolotl (Fick, ’93) and Bufo (King, ’01), the tail
is not left behind at the surface; in Cryptobranchus it continues
toserve as an efficient organ of propulsion. The undulating mem-
brane persists, though it is not shown in the figure. The head at
this time stains very faintly with the nuclear stain. The acro-
some and middle-piece, always difficult to see with the magni-
fication employed for the study of thick serial sections, have not
been observed in this situation.

Surrounding the shaft of the spermatozoon for a short distance
behind the head there is a spindle-shaped yolk-free region contain-
ing cytoplasm. This cytoplasm is particularly dense about the
region of the middle-piece; from this locality as a center cyto-
plasmic strands, resembling linin threads, but finely granular,
radiate in all directions, but those extending backward are more

EMBRYOLOGY OF CRYPTOBRANCHUS 141

prominent. This pheromenon is much more marked in some
other cases than in the one figured.

In eggs taken about three hours after fertilization (see fig. 42),
the thickening of the cell wall has flattened to the form of a disc;
it is strongly striated, recalling the zona radiata from which it
takes its ultimate origin. The perforation made by the spermato-
zoon has disappeared. The cytoplasmic path of the spermato-
zoon has become filled with yolk, except for a broad shallow region
underlying the thickening of the cell wall. The head of the sper-
matozoon has become shorter and thicker, and takes brilliantly
the nuclear stain; the tail has become slightly shorter, perhaps
by the degeneration of the posterior portion. The radiations of
cytoplasm proceeding from the region of the middle-piece have
disappeared, but in the same locality there is a somewhat larger
spherical region of uniformly distributed yolk-free cytoplasm.

Five hours after fertilization (see fig. 48), the spermatozoon
has penetrated only a little deeper into the egg. The thickening
of the cell wall of the egg at the place of entrance of the spermato-
zoon has disappeared, but its site is marked by convolutions in
the cell wall. The protoplasmic path leading from the surface of
the egg to the spermatozoon has almost entirely disappeared, but
traces of it persist at intervals along the route. The head of the
spermatozoon is spindle-shaped and much shorter and thicker
than before; the tail persists, but is somewhat abbreviated. The
circular area of cytoplasm surrounding the head of the spermato-
zoon has expanded to form a large crescent, whose horns extend
nearly at right angles to the path of the spermatozoon. The
yolk granules underlying the crescent are decidedly coarser than
those above it. This suggests a correlation of the internal struc-
ture with the ‘sperm area’ seen from the surface: the horns of
the crescent produced would meet the margin of the sperm area
(compare figs. 9 to 11 with figs 48 to 45).

Seven and one-half hours after fertilization (see fig. 44) the pro-
toplasmie path is marked only by a region of sparsely distributed
yolk granules extending from the surface for about two-thirds of
the distance to the spermatozoon. The yolk granules are par-
ticularly dense in the region immediately above the crescent, and

142 BERTRAM G. SMITH

°, onne, oO 9092
remeron sumeneeRatEn
TAKer 20098 20% Ab O
BIO SEEQIOSFOOS ED 20
OSHS SO? On8 GESO Oe
B28" eam eee 070

j 0020

008 KO.

09000

OF
o

xe Y g 0; O 0
¢ 022, QZ OC) Os ass IY,
Rees RUN ae

O S
200s560006 ORR 2608

46 47 48

Fig. 44 Vertical section through an egg of Cryptobranchus allegheniensis
killed 73 hours after fertilization, showing a late stage in the penetration by the
spermatozoon. X80. This spermatozoon has entered the egg about 30° from the
animal pole, and its path inclines toward the axis of polarity of the egg. The head
of the spermatozoon is shown entire; the tail persists in an abbreviated and
perhaps fragmented condition, but does not appear in the section figured. An
aster is present in an adjacent section at a little higher level than the sperm head.
Another spermatozoon in the same general condition is found in the same egg.

Fig. 45 Vertical section through an egg killed 10! hours after fertilization,
showing the sperm-nucleus. 80. The spermatozoon has entered the egg about
25° from the animal pole. Fragments of the tail of the spermatozoon are to be
found in the vicinity of the sperm-nucleus, but are not shown in this section.

Fig. 46 A portion of fig. 45 enlarged to show the sperm-nucleus. X 240.

Figs. 47 and 48. Two consecutive meridional sections through an egg of Crypto-
branchus allegheniensis, killed 10} hours after fertilization, showing the egg-
nucleus. This nucleus is situated about one-third as far from the surface as the
copulation-nucleus shown in fig. 52. XX 240.

EMBRYOLOGY OF CRYPTOBRANCHUS 143

are here finer than elsewhere at the same level. The crescent
has become larger, and thicker at the ends than in the middle.
The head of the spermatozoon is shortened to the form of a thick
spindle and stains deeply; the tail persists in an abbreviated con-
dition. In the case of the spermatozoon shown in figure 44, an
aster is found at a slightly higher level than the sperm head and a
little nearer to theegg-nucleus. A study of the protoplasmic paths
of the aster and the sperm-head shows that they have separated
at a point midway in the path of the latter.

Ten and one-half hours after fertilization (see figs. 45 and 46)
the spermatozoon has become transformed into the sperm-
nucleus, which is amoeboid in form; the tail cf the spermatozoon is
represented only by fragments. At this time the sperm-nucleus
lies about half as far from the surface as the copulation-nucleus
shown in fig. 52. Immediately beneath the sperm-nucleus lies
a considerable mass of cytoplasm, perhaps formed at the expense
of the crescent which is dwindling except at the extreme ends.
The remains of the crescent, and the characteristic appearance of
the surrounding yolk, enable one readily to distinguish the sperm-
nucleus from the egg-nucleus. The sperm-nucleus is smaller
than the egg-nucleus, and like the latter does not stain deeply at
this time.

The cytoplasmic changes in the egg caused by the invasion of
the spermatozoon may be tentatively interpreted as follows:
Under the influence of the centrosome, whose seat appears in this
case to be in the middle-piece, egg-cytoplasm collects about the
neck of the spermatozoon. Here the centrosphere and event-
ually the entire aster isformed. As the spermatozoon invades the
deeper region of coarser yolk particles, the resistance offered to the
progress of the accompanying mass of cytoplasm causes it to
flatten out like a bullet fired against a wall, assuming a form cres-
cent-shaped in section. Presently the spermatozoon, during its
transformation into the sperm-nucleus, comes almost to a full
stop, allowing the mass of cytoplasm again to assume a spherical
form.

Numerous observers have described in both invertebrates and
vertebrates a rotation of the sperm head after it enters the egg,

144 BERTRAM G. SMITH

whereby the middle-piece is brought into position to precede
during the further process of migration. Fick (’93) has described
this process in Axolotl, and Dean (’06) has noted it in Chimaera.
King (701) found no indication of a rotation of the sperm head
in Bufo; possibly this condition is correlated with the fact that in
Bufo the centrosome appears to be located, not in the middle-
piece, but in the head of the spermatozoon. In Cryptobranchus
rotation of the sperm head apparently takes place at a rather late
stage in the process of transformation into the sperm-nucleus. In
the stages shown in figures 43 and 44, the greatly shortened sperm
head is usually placed with its long axis oblique or at right angles
to its former path, so that one-end points toward the egg nucleus.
But in these stages it has not been possible to trace any connec-
tion between the tail of the spermatozoon and its head, and since
the aster has already separated from the sperm head, in no case
can it be stated which end of the sperm head is the one pointed
toward the egg-nucleus.

The spermatozoon ordinarily enters the blastodise in a more or
less centripetal direction, and continues in this direction for a con-
siderable distance; sometimes its path inclines almost from the
beginning in an oblique direction toward the point of future union
with the egg-nucleus. In either case the axis of the spiral path
is ordinarily straight up to the time of the transformation of the
head of the spermatozoon into the sperm nucleus; the later course
of migration has not been followed. In an egg preserved an hour
and three-quarters after fertilization, a spermatozoon, which had
entered the blastodise unusually near the animal pole, described
a path which proceeded in a centripetal direction only a very short
distance, then curved sharply in a direction parallel to the sur-
face, toward the second polar spindle which was in the late ana-
phase condition. The form of the spermatozoon remained unal-
tered, and rotation of the sperm head had not commenced. This
case is instructive in showing that the factors tending to bring the
germ-nuclei together are active at a very early stage of fertiliza-
tion: the egg-nucleus was not fully formed, and the spermatozoon
had not begun its process of transformation into the sperm-
nucleus. Moreover it is evident that in this case at least the

EMBRYOLOGY OF CRYPTOBRANCHUS 145

‘copulation path’ (see Hertwig, ’06, p. 529) is not dependent
upon the rotation of the sperm head.

(b). Polyspermy, and the fate of the supernumerary spermatozoa.
Brief data regarding the occurrence of polyspermy have already
been given. It is possible that the method of artificial fertili-
zation increases the number of spermatozoa entering the egg; but
in nature the eggs are fertilized in a confined space, and I see no
reason to doubt that polyspermy is a common occurrence under
natural as well as artificial conditions. It is evident that we have
here to deal with physiological, not induced or accidental poly-
spermy (see Brachet, ’10), for the eggs develop in anormal manner.

While the distribution of spermatozoa entering the egg is largely
if not entirely a matter of chance, the location in which a sperma-
tozoon finds itself has much to do with its ultimate fate. Sper-
matozoa entering the lower hemisphere, especially in the region of
the vegetal pole, never penetrate far, and since they are found in
this hemisphere only during the first few hours after fertilization,
must quickly degenerate. In the urodele Hynobius, Kunitomo
(10) found that a spermatozoon entering at the vegetal pole
sometimes succeeds in reaching the egg-nucleus; but the careful
study of many eggs has convinced me that this never occurs in the
heavily yolk-laden and strongly telolecithal egg of Cryptobranchus.

Only in the blastodise have spermatozoa been found in the stage
characterized by the presence of the cytoplasmic crescent (figs.
43 and 44). Obviously, the conditions elsewhere are unfavorable
for the formation of any considerable mass of cytoplasm about
the spermatozoon. In the stage with a well-formed cytoplasmic
crescent, not more than two spermatozoa have been found in a
single egg. No accessory spermatozoa have been found in any
situation after the formation of a sperm-nucleus. The supernu-
merary spermatozoa thus have but a transient existence, and the
only advantage resulting from polyspermy is doubtless that,
in an egg so large, penetration by several spermatozoa is of value
in insuring fertilization.

The literature on the occurrence of polyspermy in the amphib-
ian egg has recently been reviewed by Kunitomo (’10). As
noted in a previous paper (Smith, ’11), I have found polyspermy

146 BERTRAM G. SMITH

occurring in eggs of Amblystoma tigrinum fertilized under nat-
ural conditions; the material secured for the further investigation
of this subject has not yet been worked up. Polyspermy seems to
be characteristic of heavily yolk-laden eggs lacking a preformed
micropyle.

SOLO .Ou ea 382 O% OOO OAGLON08
0 p08, AO Bar ON QO00%. LO) 0 08 20:

49 50

Fig. 49 Meridional section through an egg of Cryptobranchus allegheniensis,
killed 12 hours after fertilization, showing fusion of germ-nuclei. The sperm-
nucleus is probably the smaller one. X 240.

Fig. 50 Meridional section showing fusion of germ-nuclei in another egg killed
12 hours after fertilization. The sperm-nucleus is probably the lower and smaller
one. X 240. For the position of this copulation-nucleus in the blastodisc see fig.
52 which is drawn from the same section.

3. Union of the germ-nuclei, and formation of the first cleavage
spindle ;

In two eggs killed twelve hours after fertilization, the germ-
nuclei have been found in the process of uniting (figs. 49 and 50);
in these two cases the copulation-nuclei are at approximately the
same distance from the surface (see fig. 52), quite deeply situated
in the blastodise and a little to one side of the axis of polarity.
The smaller germ-nucleus is probably the sperm-nucleus. The
egg-nucleus stains brilliantly with borax-carmine; the sperm-
nucleus takes the stain less deeply. The sperm-nucleus especially
is surrounded with dense cytoplasm; in one case (fig. 50) this

EMBRYOLOGY OF CRYPTOBRANCHUS 147

exhibits a tendency toward radial striation and probably repre-
sents the aster.

The study of the paths of migration of the germ-nuclei and
the copulation-nucleus is not quite complete, but indicates that
the germ-nuclei come together at a higher level than that occu-
pied by the copulation-nuclei shown in the figures.

The first segmentation nucleus in a resting condition has been
found in an egg killed eighteen hours after fertilization; the first
cleavage spindle has been found in an egg killed seventeen hours
‘after fertilization.

4. Changes in the blastodisc

In eggs taken from fifteen minutes to ten and one-half hours
after fertilization, cytoplasm is accumulating in irregular patches
underlying the animal pole (fig. 51). During this period, practi-
cally all traces of the débris of the germinal vesicle disappear.
In places, the surface of the blastodisc is sometimes very irregular,
almost villous; this may be due to injuries resulting from the actual
or attempted entrance of spermatozoa.

In eggs taken from twelve to eighteen hours after fertilization
(copulation nucleus to first cleavage spindle) the cytoplasm is
gathering in a broken layer close to the surface of the blastodisc.
The beginning of this process is shown in figure 52. In Hyno-
bius, Kunitomo (710) has noted a somewhat similar condition.
During the latter part of the period considered the layer of cyto-
plasm becomes much thicker than is shown in the figure, but
retains its segmented character.

During the first two hours after fertilization there is a marked
increase in the thickness and extent of the blastodise as a whole
(see especially fig. 51). Evidently the greater part of this change
takes place before the egg has become oriented with the animal
pole uppermost, hence it is independent of any possible sorting
effect of gravity acting on the materials of the egg.

No marked changes have occurred in the lower hemisphere since
the egg left the ovary.

The later changes in the blastodise lead up to first ORM and
will be considered in that connection.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1

148 BERTRAM G. SMITH

D. SUMMARY

The follicular layer proper of the ovarian egg of Cryptobranchus
is formed from some of the deeper non-germinal cells of the ova-
rian wall which resemble the epithelial cells of the outer and inner
limiting membranes. The follicular membrane proper completely
surrounds the egg and is suspended in a two-layered flask-shaped
sac which projects from the inner surface of the wall of the ovary

§2

Fig. 51 Meridional section through an egg of Cryptobranchus allegheniensis,
killed 1$ hours after fertilization, showing the condition of the blastodise. The
irregular faintly stippled areas near the animal pole contain yolk-free cytoplasm.
x 18. p.s. IT, second polar spindle.

Fig. 52 Meridional section through an egg of Cryptobranchus allegheniensis,
killed 12 hours after fertilization, showing condition of the blastodise and position
of the copulation-nucleus. Yolk-free cytoplasm is segregated in a broken layer
near the surface of the blastodise. The copulation-nucleus is shown a little to the
left of the center of the figure. 18.

EMBRYOLOGY OF CRYPTOBRANCHUS 149

into the central cavity; in a broad sense, the entire three-layered
structure may be called the follicle.

The zona radiata is formed from the peripheral substance of
the egg proper; at the time of the rupture of the germinal vesicle
it becomes transformed into a simple cell wall, in organic connec-
tion with the egg.

The zona pellucida is formed as a secretory product of the fol-
licular layer proper; it persists unchanged as the ‘ vitelline mem-
brane’ of the embryo.

The earliest observed phenomena which may perhaps indicate
polarity occur in the ovarian eggs of young females of a body
length of 26 to 30 em., as a shifting of the region of most abundant
vitelline bodies from the future vegetal to the future animal hemi-
sphere. In the ovarian eggs of young females of a body length
of 35cm. there is a concentration of nucleoli on the side of the germ-
inal vesicle toward the future animal pole; this may perhaps
afford a second indication of polarity.

Yolk-formation begins in the most advanced ovocytes of young
females with a body length of 35 em.; the yolk is first laid down in
concentric zones. With respect to the position of the germinal
vesicle, the distribution of cytoplasm, and the size of the yolk
particles in the different zones, the egg exhibits radial symmetry
until after it is nearly or quite filled with yolk.

About the time when the egg becomes completely filled with
yolk, the germinal vesicle migrates from its central position toward
a point on the surface which is thus defined as the animal pole.
Coincident with the migration of the germinal vesicle, axial differ-
entiation of the cytoplasmic and yolk contents of the egg lead to
the formation of a germinal disc in the region of the animal pole.

In general the animal pole of the egg lies within the stalk of the
follicle and toward the periphery of the ovary.

In the late ovarian egg a structure called the yolk cup 1s inter-
preted as the physiological equivalent of the concentric layers of
dense fine yolk found in the eggs of birds and various other verte-
brates.

Shortly before maturation the germinal dise is temporarily
differentiated into two layers: a thin peripheral layer of yolk-

150 BERTRAM G. SMITH

free cytoplasm, and underlying this a thicker layer of especially
fine yolk particles rich in cytoplasm. Both layers are continuous
with much thinner layers of the same character surrounding the
remainder of the egg.

In the ovocyte ready for maturation, the germinal vesicle hes
close to the surface at the animal pole, and is surrounded by the
germinal disc. A mass of cytoplasm has accumulated beneath
the germinal vesicle during the later stages of its migration. The
arrangement of materials is now quite strongly telolecithal.

Shortly before the rupture of its wall, the germinal vesicle
appears at the very surface at the animal pole. The rupture of the
germinal vesicle takes place just before the egg leaves the ovary;
the cytoplasmic and yolk layers of the blastodise now mingle,
and the materials of the germinal vesicle, together with the cyto-
plasm brought with it from the interior of the egg, are incorpor-
ated into the blastodise.

Absorption of degenerating ovocytes is accomplished by means
of the follicle cells, which reverse their usual role as nurse cells
of the egg, and function as phagocytes. }

The first polar spindle is formed about the time the egg leaves
the ovary, and disappears about the time the egg enters the uterus.
There are marked size differences in the chromosomes.

The second polar spindle is formed shortly after the egg enters
the uterus; it lies beneath a deep pit readily visible from the
surface.

The penetration of the egg by the spermatozoon is not required
as a stimulus to the formation of the second polar spindle.

The late stages of the second maturation division, culminating
in the formation of the second polar body and the egg-nucleus, are
passed through only after the spermatozoon has entered the egg;
in other words, the processes of maturation and fertilization
overlap.

A structure resembling a micropyle is formed in the cell wall of
the egg around the perforation made by the entrance of the sper-
matozoon. The influence of the entering spermatozoon upon the
egg is shown by characteristic changes in the distribution of the
yolk and cytoplasm.

EMBRYOLOGY OF CRYPTOBRANCHUS 151

Physiological polyspermy is of normal occurrence. The super-
numerary spermatozoa lead but a transient existence.

Union of the germ-nuclei takes place at a point deeply situated
near the center of the blastodisc, and is followed by the segrega-
tion of masses of cytoplasm forming a broken layer near its sur-
face.

BIBLIOGRAPHY

BracHet, A. 1910 La polyspermie experimentale comme moyen d’analyse de
la fécondation. Archiv f. Entw., Bd. 30, no. 1.

Brauer, Aucust 1897 Beitrage zur Kenntniss der Entwicklungsgeschichte
und der Anatomie der Gymnophionen. Zool. Jahrb. Anat., Bd. 10.

Bupgett, J. 8S. 1899 Notes on the batrachians of the Paraguayan Chaco, with
observations upon their breeding habits and development, especially
with regard to Phyllomedusa hypochondrialis Cope. Quart. Jour.
Micros. Se., N.S., vol. 42; also Budgett memorial volume (Kerr, ’07 a).
1901a On the breeding habits of some west African fishes, with an
account of the external features in the development of Protopterus
annectens, and a description of the larva of Polypterus lapradei.
Trans. Zool. Soc. London, vol. 16, part II, August. Also Budgett
memorial volume (Kerr, ’07 a.)
1901b The habits and development of some west African fishes. Proc.
Camb. Philos. Soc., vol. 11; also Budgett memorial volume (Kerr, ’07 a).

Btuuter, A. 1902 Riickbildung der Eifollikel bei Wirbeltieren. I. Fische.
Morphol. Jahrb., Bd. 30.

Dean. BasHrorD 1896 The early development of Amia. Quart. Jour. Micros.
Sci., vol. 38.
1906 Chimaeroid fishes and their development. Carnegie Institution
of Washington.

DeBussy, L. P. 1904 Eerste Ontwikkelungsstadien van Megalobatrachus maxi-
mus Schlegel. Tijdschrift der nederlandsche Dierkundige Veree-
niging, Bd. 8.
1905 Die ersten Entwicklungsstadien des Megalobatrachus maximus.
Zool. Anz., Bd. 28.

EYCLESHYMER, ALBERT C. 1906 The habits of Necturus maculosus. Amer.
Nat., February.

Fick, R. 1893 Ueber die Reifung und Befruchtung des Axolotleies. Zeitschr.
f. wiss. Zool., Bd. 56.

Haraitt, CHas. W. 1888 Recent notes on Scaphiopus holbrookii. Amer. Nat.,
vol. 22.

Harrinaton, N. R. 1899 Respiratory and breeding habits of Polypterus.
Science, N.S., vol. 9, pp. 314-315.
1899 The life habits of Polypterus. Amer. Nat., vol. 33, p. 721.

LZ BERTRAM G. SMITH

Hay, O. P. 1888 Observations on Amphiuma and its young. Amer. Nat., vol.
22.
1890 The skeletal anatomy of Amphiuma during its earlier stages.
Jour. Morph., vol. 4.

Hertwiac, Oscar 1906 Handbuch der vergleichenden und experimentellen Ent-
wicklungslehre der Wirbeltiere. Bd. 1, Th. 1. Jena, Gustav Fischer.

IsHikawa, C. 1904 Beitriige zur Kenntniss des Riesen-Salamanders (Megalo-
batrachus maximus Schlegel). Proc. Dept. Nat. Hist., Tokyo Imper.
Mus., vol. 1, no. 2.

JenkKINSOoN, J. W. 1909 Experimental embryology. Oxford. Clarendon Press.

Kersert, C. 1904 Zur Fortpflanzung von Megalobatrachus maximus Schlegel.
Zool. Anz., Bd. 27, no. 10.

Kerr, J. Granam 1900. External features in the development of Lepidosiren.
Phil. Trans. Royal Soc., Ser. B, vol. ‘192.
1901 The development of Lepidosiren paradoxa. II. With a note
upon the corresponding stages in the development of Protopterus
annectens. Quart. Jour. Micros. Sci., N.S., vol. 45.
1907 a The work of John Samuel Budgett. Cambridge. Univ. Press.
1907 b The development of Polypterus senegalus Cuv. Budgett memo-
rial volume (Kerr, ’07 a).

Kine, Heren Dean 1901 The maturation and fertilization of the egg of Bufo
lentiginosus. Jour. Morph.,vol. 18, no. 2.
1902 The follicle sacs of the amphibian ovary. Biol. Bull., vol. 3.
1905 The formation of the first polar spindle in the egg of Bufo len-
tiginosus. Biol. Bull., vol. 9, no. 2.
1908 The ovogenesis of Bufo lentiginosus. Jour. Morph., vol. 9,
no. 2.

Kineasuey, J. S. 1899 Text-book of vertebrate zoology. New York. Henry
Holt and Company.

Krineassury, B. F. 1895 The spermathecae and methods of fertilization in some
American newts and salamanders. Trans. Amer. Micros. Soc., vol. 17.

Kuniromo, Kanar 1910 Ueber die Entwicklungsgeschichte des Hynobius
nebulosus. Anatom. Hefte, Beitr. u. Ref. z. Anat. u. Entw., Bd. 40.

LILui£, Frank R. 1908 The development of the chick. New York. Henry Holt
and Company.

McGreeor, J.H. 1896 Anembryo of Cryptobranchus. Proc. New York Acad.
Sci., Dec. 14. Abstract in Zool. Anz., Bd., 20, p. 29.
1899 The spermatogenesis of Amphiuma. Jour. Morph., vol. 15, Sup-
plement.

Osporn, H. F. 1888 A contribution to the internal structure of the amphibian
brain. Jour. Morph., vol. 2, p. 51.

Peart, RayMonpd 1909 The nature of the stimulus which causes a shell to be
formed on a bird’s egg. Science, March 12.

Pixr, Nicwouas 1886 Notes on the hermit spadefoot (Seaphiopus holbrookii
Harlan). Bull. Am. Mus. Nat. Hist., vol 1, no. 7.

EMBRYOLOGY OF CRYPTOBRANCHUS 153

Reese, AuBeERT M. 1903 The habits of the giant salamander. Pop. Sci. Mo.,
April.

1904 The sexual elements of the giant salamander, Cryptobranchus
allegheniensis. Biol. Bull., vol 6, no. 5.

RercHarD, JacoB 1903 The natural history of Amia calva Linnaeus. Mark
anniversary volume.

Rippiez, Oscar 1911. On the formation, significance and chemistry of the white
and yellow yolk of ova. Jour. Morph., vol. 22.

Roux, W. 1883 Ueber die Zeit der Bestimmung der Hauptrichtungen des Frosch-
embryo. Leipzig.

1885 Ueber die Bestimmung der Hauptrichtungen des Froschembryo
im Hi und iiber die erste Theilung des Froscheies. Bresl. irtz. Zeit-
sehr.

1887 Die Bestimmung der Medianebene des Froschembryos durch
die Kopulationsrichtung des Eikernes und des Spermakernes. Archiv
f. mikros. Anat., Bd., 29.

1903 Ueber die Ursachen der Bestimmung der Hauptrichtungen des
Embryo im Froschei. Anat. Anz., 23.

Sarasin, P. B. and C.F. 1887-1893 Zur Entwicklungsgeschichte und Anatomie
der ceylonischen Blindwiihlen, Ichthyophis glutinosus. Ergebnisse
naturwissenschaftlicher Forschungen auf Ceylon, Bd.2.

Sasakr, C. 1887 Some notes on the giant salamander of Japan. Jour. Coll.
Sci., Imper. Univ. Tokyo, Japan, vol. 1, part III.

ScuuutzE, O. 1900 Ueber das erste Auftreten der bilateralen Symmetrieim Ver-
lauf der Entwicklung. Archiv fiir mikr. Anat., Bd. 55.

SmirH, BertRAMG. 1906 Preliminary report on the Snbine ology of Cryptobran-
chus allegheniensis. Biol. Bull. vol. 11, no. 3.

1907 The breeding habits of mnie gatortes pence: Amer. Nat.,
vol. 41, no. 486, June.

1907 The life history and habits of Cryptobranchus Piechenicnas.
Biol. Bull., vol. 13, no. 1, June.

1908 The spawning hain of Chrosomus erythrogaster Rafinesque.
Biol. Bull., vol. 14, no. 6, May.

1910 The Jinn of the spermatophores of Amblystoma punctatum,
Biol. Bull., vol. 18, no. 4, March.

1911 The nests and larvae of Necturus. Biol. Bull. vol. 20, no. 4,
March.

1911 Notes on the natural history of Amblystoma jeffersonianum, A.
punctatum, and A. tigrinum. Bull. Wis. Nat. Hist. Soc., vol. 9, April.

TowNSEND, Cuas. H. 1882 Habits of Menopoma. Amer. Nat., vol. 16, no. 2.

WiepersHEIM, R. 1900 Brutpflege bei niederen Wirbeltieren. Biol. Centralb.,
Bd. 20. @

Wuitman, C. O. 1898 Animal behavior. Woods Hole Biological Lectures.

PLATE 1
EXPLANATION OF FIGURES

53 Cryptobranchus allegheniensis. Living eggs dissected from the ovary,
showing the germinal vesicle at the surface. The small spots within the germinal
vesicle are probably nucleoli. Ovarian membranes containing blood-vessels
wholly or partially cover the eggs. X 4.

54 Cryptobranchus allegheniensis. Unfertilized eggs with their gelatinous
envelopes, after two days’ immersion in water. Natural size.

154

CRYPTOBRANCHUS ALLEGHENIENSIS PLATE 1

BERTRAM G. SMITH

53

JOURNAL OF MORPHOLOGY, VOL. 23

PLATE 2

EXPLANATION OF FIGURES

55 Necturus maculosus. ‘Nest’ of eggs of Necturus. The stone to which the
eggs are attached has been removed from the water and set on edge on the wharf;
it is about 16 inches in diameter. The embryos are in an advanced stage of devel-
opment.

56 Necturus maculosus. Eggs and envelopes shown in their natural position
in the water, attached to a piece of board; natural size. From a drawing by
Professor Bashford Dean.

NECTURUS MACULOSUS PLATE 2
BERTRAM G. SMITH

BODY SIZE AND CELL SIZE’

EDWIN G. CONKLIN

From the Biological Laboratory, Princeton University

TWELVE FIGURES

In his very thoughtful and suggestive address at the Zoological
Congress of the World’s Columbian Exposition on “The In-
adequacy of the Cell Theory of Development”’ Professor Whitman
(93) pointed out the fallacy of the view, prevalent since the time
of Schleiden and Schwann, that ‘organization means cellular
structure and that ontogeny means cell formation.’ On the
other hand he maintained that ‘‘ organization precedes cell forma-
tion,” and that ‘‘the secret of organization, growth, development
lies not in cell formation, but in those ultimate elements of liv-
ing matter for which zdiosome seems to me an appropriate name.”

The truth and importance of this position were well supported
in his argument and have been justified by later work on the or-
ganization of the germ cells, as well as by the older work on the
Protozoa; but no one would have been more ready than Professor
Whitman to recognize the fact that this protest against a dominant
doctrine did not express the whole truth. It may be granted
that a certain amount of organization precedes cell formation
without granting that all organization does. Indeed we know
that the latter is not true; very much of the organization which
we see in higher organisms occurs only after cell formation, and,
in all probabilities, as a result of it. Indeed the organization
which is possible without cell formation is probably only such as
is found among Protozoa and germ cells.

What the full meaning of cell formation in the development
of higher organisms is we do not know, but it seems practically

1 Prepared for The Whitman Memorial Volume, but received too late to be
included.

159

160 EDWIN G. CONKLIN

certain that it is associated not merely with growth in size but
also with growth in complexity. The degree of differentiation
of an organism is not determined entirely by the number of cells
in its body, but cell number is certainly one of the factors in
differentiation. It is now known that the protoplasm of certain
egg cells is composed of different substances which may be
localized in different parts of the egg, and in this respect it
resembles the protoplasm of a protozoan cell. Lillie (02) has
shown that certain simple differentiations, such as the appearance
of cilia, may occur in eggs which do not undergo cleavage, but
these differentiations do not go beyond this primitive, pro-
tozoan stage. When these differentiated portions of the egg
protoplasm are separated from one another by cell membranes,
the differentiations increase in degree, if not in number. The
whole progress of embryonic differentiation is thus associated
with differential cell division.

Of course there are many cell divisions which are non-differ-
ential, in which the daughter cells are alike, and such cell divisions
are not essential to progressive differentiation, though they are
essential to growth. We are thus able to separate growth and
differentiation, for while the latter consists in an increase in the
kind of structures, the former consists in an increase in their size
and number merely. In embryonic development, however, these
two processes usually go hand in hand; and in many instances size
differences between cells are the earliest differentiation which
can be recognized.

Almost all species of animals and plants have a more or less
characteristic body size which, in spite of individual variations,
may be said to constitute the norm of the species. Specific differ-
ences in body size may be slight, or they may be enormous, as
in the case of the mouse and the elephant. In similar manner
different individuals of the same species may differ in a marked
degree in body size, the individuals which represent the extremes
of size being known as dwarfs and giants. The question whether
the difference between a dwarf and a giant, or between a small
sized and a large sized race or species, is due to a difference in
size of the ultimate units of structure, the cells, or to a difference

BODY SIZE AND CELL SIZE 161

in their number, or to both, is one of fundamental interest and
importance.

Some of the earliest observations in this field were made by
botanists and served to show that differences in body size are due
to the number of cells present rather than to the size of the indi-
vidual cells. Erich Amelung (’93) determined ‘‘that the larger or
smaller development of a plant body has no influence on the size
of its adult constituent cells.”” Sachs (’93) also called attention
to the lack of correlation between cell size and body size. Stras-
burger (’93) reached the same conclusion concerning the embry-
onic cells from the growing points of extremely large and small
individuals. He says: a

Not the cell size, only the cell number, is influenced by the different
size ofanindividual . . . . Itwassurprising to me to find that while
individuals of the same species always showed the same size of embry-
onic nuclei and cells, varieties of these same species might differ greatly
from each other.

In an interesting paper on one of the mutants of Oenothera
lamarckiana, O. gigas, Gates (09) comes to practically the same
conclusion. He says (p. 543):

In O. gigas we have an organism built of bricks which are larger and
whose relative dimensions are also altered in some cases. These two
factors will apparently account for all the differences between O. gigas
and O. lamarckiana, and the second factor may be one merely of readjust-
ment consequent upon the first. It is probable that the number of
cells is approximately the same in both cases.

This is evidently one of those cases, of which Strasburger speaks,
in which different species or varieties of the same species may differ
greatly from each other in cell size.

In the case of animals the earlier work on cell size was confined
largely to studies on nerve cells and fibers. Gaule (’89) concluded
from a study of the spinal cord of the frog that differences in the
size of individuals of the same species influenced only the size of
the ganglion cells but in no way influenced their number, which
might be considered a constant factor. Donaldson (95) after
summarizing a number of observations on the number and size
of nerve cells in man says (p. 192):

162 EDWIN G. CONKLIN

The determination of the number of neuroblasts occurs so early in
the history of an individual, and under such uniform conditions, that it
is very difficult to regard the environment as possessed of much power
to cause variation in this respect, and for this reason among members
of the same race a high degree of constancy in this character is to be
anticipated.

He regards differences in brain weight. as due to differences in
the size of the nerve cells, the number remaining constant. Hard-
esty (’02) found that in various mammals the motor nerve cells
from the spinal cord were largest in the elephant and smallest in
the mouse, while in animals which were intermediate in size the
nerve cells were intermediate. Levi (’05) has made an extensive
study of cell size in different species, particularly mammals.
He finds that the size variations of epithelial and gland ceils are
insignificant and are not correlated with botly size. Measure-
ments of ganglion cells gave entirely different results; here the
size of the cell varies with the size of the animal. Nerve fibers
and lens fibers show the same correlation with body size as do
ganglion cells. In muscle he finds that the diameter of the fiber
is usually larger in large animals than in small ones, though this
is subject to many variations. Levi points out the significant
fact that epithelial and gland cells divide throughout life, whereas
the other types of cells named cease to divide at an early age.

Morgan (’95) concluded that in echinid larvae derived from
isolated blastomeres the number of cells is approximately pro-
portional to the size of the blastomeres or egg fragments from
which they came. However in the case of partial larvae of Am-
phioxus Morgan (’96) held that one-half larvae contained about
two-thirds as many cells as whole larvae, while one-fourth larvae
had more than one quarter, but not quite one-half the total num-
ber in a whole larva. Driesch (’98, ’00) determined in a very
satisfactory manner that the number of cells in partial larvae
of echinids is one-half the normal number in one-half larvae, one-
fourth in one-quarter larvae, ete., while in larvae from two fused
eggs the number of cells is double the normal number. And since
these partial or double larvae all have the typical structure of a
normal larva he was led to the formulation of the ‘rule of the
fixed size of specific organ-cells.’

BODY SIZE AND CELL SIZE 163

Rabl (99) found that the size of cells of the crystalline lens
was practically constant, but the cell number a variable one,
depending upon the size of the organ. Boveri (’04) also found
that in dwarfs and giants of the human species the size of epithe-
lial cells from the tongue and of bone corpuscles from a phalanx
agreed perfectly with those from an individual of normal size.

Chambers (08), on the other hand, has questioned the view
that a certain cell size is characteristic of a species or race; he
finds that the size of an individual frog, and the size of its con-
stituent cells, depends upon the size of the egg from which it
came. In agreement with the earlier work of Morgan (’04), he
has found that the size of the frog’s egg may vary considerably ;
whether laid by the same individual, or by different ones, the
diameters of eggs of extreme size may vary as much as | :3.
The smaller eggs give rise to smaller tadpoles and frogs, which
are composed of smaller cells, than those derived from larger eggs.
He believes with Popoff (’08) that the initial cause of the varia-
tions in the sizes of eggs is to be found in unequal division of the
nucleus or plasma of the oocytes.

Popoff (’08) holds that sperm ‘cells as well as egg cells vary in
size. He supposes that when a large egg is fertilized by a large
spermatozoon a large individual with large cells results; whereas
from small eggs and small spermatozoa small individuals with
small cells arise. He admits that the operations of this law may
be obscured by two other conditions, viz., (1) various factors which
limit or inhibit growth, and (2) rich nourishment which influences
only the number of cells, and not their individual size. He
affirms, therefore, that body size is not inherited as commonly
supposed—it has no specific representative in the germ plasm;
only the size of cells, not the size of individuals, is inherited.

Berezowski (710) has made a study of cell size in relation to body
size in white mice, varying in age from ten days to five months,
and in body volume from 4 cc. to 25 ce. He finds that with the
growth of an animal there is an increase in the cell size and nuclear
size, particularly in the length of cells and of nuclei of the
intestinal epithelium adjoining the pylorus. In a second paper,

164 EDWIN G. CONKLIN

Berezowski (’11), shows that the cell size is slightly greater in
castrated mice than in normal ones.

In a series of briliuant studies, Jennings (’08, ’09, 710, ’11, ete.)
has shown that at least eight different races, or ‘pure lines,’ of
Paramecium may be distinguished by their size. The mean
length of the largest race is to that of the smallest about as 5: 1.
While the variations in length within each race is considerable,
the norm for each race is quite characteristic. He finds that the
differences in size between individuals of the same race are due
to growth and environment and are not inherited, whereas the
differences between different races are inherited.

In addition to the foregoing references there are doubtless many
other observations on the relations of cell size to body size scat-
tered through the literature. These references, however, are
believed to include the most important works on this subject,
as well as the principal conclusions which have been reached.

The following observations on the relative size and number of
cells from the bodies of different species of Crepidula, and from
different individuals of the same species were made many years
ago, and a brief note on this subject was published at that time
(Conklin ’96), and a somewhat more complete account in a sub-
sequent paper (’98).

1. CELL SIZE AND BODY SIZE IN DIFFERENT SPECIES OF
CREPIDULA

I have made no extensive study of cell size in relation to body
size in different classes of animals, most of my work having been
confined to different species of gasteropods, and principally to
the genus Crepidula. Without any special study on this subject,
however, it is quite evident from casual observation that different
classes differ widely in cell size, and that these differences are not
usually correlated with differences in body size. The great size
of cells in amphibians, nematodes and some insects is well known,
whereas in echinoderms, annelids and mammals the cells are
relatively much smaller. Even different species, and varieties
of the same species, may show considerable differences in cell
size, as Strasburger, Gates and Boveri have shown.

od

BODY SIZE AND CELL SIZE 165

a. Body size. I have studied in some detail the body size and
cell size of four species of the genus Crepidula, as well as of sev-
eral species of other prosobranch gasteropods. The mature
females of Crepidula are always larger than the males and are
firmly attached to some object from which they are unable to
move; the males on the other hand are smaller and have greater
power of locomotion. In the following measurements of body
size and cell size the two sexes were always carefully separated;
twenty mature individuals of the same sex were chosen and, after
having been removed from their shells, they were placed upon
blotting paper to remove any excess of water and were then

TABLE 1

Body size in the genus Crepidula

ACTUAL VOLUME OF RELATIVE VOLUMEIN | RELATIVE VOLUME IN
BODY SAME SPECIES OF DIFFERENT SPECIES
SPECIES ‘res bon et Re = Z s SAS tN Ses:
Male | Female Male Female Male Female
cc cc.

Ccouvexay 2.2.) 0.01 | 0.05 1 5.0 1.0 1.0
©vaduncasee yack k 2: 0.025 | 0 208 1 Sus 2.5 4.16
Cupless | 0.046 0.667 le aides 4.6 13.34
© formieatae.-..| 1.25 | 1.60 1 1.34 | 125.0 32.00

dropped into a known volume of water in a graduated tube; in
this way the average body volume of the twenty specimens was
determined.

b. Tissue cells. Various tissue cells from corresponding organs
and parts of organs of C. convexa, C. plana, and C. fornicata have
been measured with the Zeiss 1/1 micrometer eyepiece and 3
mm. homogeneous immersion objective. The average dimensions
of such cells are given in table 2, and these measurements show
plainly that there is no constant correlation between body size
and cell size in these different species; the size of many tissue cells
is practically the same in all species (e.g. most of the epithelial
cells) while in those cases where the cell size differs in different
species, the larger cells are not always found in the species of
larger body size. Certain cells vary so much in size, especially

166 EDWIN G. CONKLIN

in large organs where it is impossible to be certain that precisely
corresponding cells have been measured, that no great reliance
can be placed upon such measurements and comparisons; this is
the case with the epithelium of the mantle, stomach and kidney.
On the other hand in organs which consist of relatively few cells,
or where the individual cells are easily identified, much confi-
dence can be placed in these measurements; this is the case with
the cells of the gill filaments, osphradium, liver ducts, muscles,

and sex cells.
TABLE 2

Size of adult tissue cells in different species of Crepidula

i}

Cc. CONVEXA Cc. PLANA C. FORNICATA
VOLUME OF BODY 0.05 ce 0.65 ce. 1.5 ce.
Cell | Nucleus | Cell | Nucleus Cell. += Nucleus
Me a Me im ll im
Tissue cells
Ectodermal epithelium
Mantle near anus...| 6x 12 6x9 6x14) 6x9 z. - -
Gill filament, exter-|
nal ciliated cells..| 3x18 | 3x6 | 4x15 | 4x8 | 4x15 | 4x7
Cells internal to chi-. |
HMOUS OO... ic00..| 4.x 12 1 ae 6 4x15| 4x9 ANS ea ay eloxe (8)
Osphradium......... | 3 G6 | esate | Sx Thc) Sx 10" ae Bo eg
Retina, maximum | | ;
cells... 8y..| 6x 21 | eee | 6x24 | 6x6 | PR OeN Ge
Kndodermal epithelium
Liver duct,.....25-.. |) Gnas 2 5 x 24 7x21
Liver cells, without
Secrevion.....%.... | 6x 27 7 x 30 | = =
Liver cells, with se- |
eretron ers .049).% 2 15 x 50 | 15 x 59 12 x33
a 6 x 36 | 9 x28 | 6 x 27
Mesodermal cells
Blood corpuscles, |
MAKUMUM..5......| 6 6 6
Muscle fibers, maxi-
mum diameter..... 10 6 6

BODY SIZE AND CELL SIZE 167

Although the body size differs greatly in different species of
Crepidula, the volume of an adult male of C. fornicata being
one hundred and twenty-five times that of the male of C. convexa,
and the volume of an adult female of C. fornicata being thirty-
two times that of the female of C. convexa, the size of the tissue
cells is practically the same in all of these species. Of course it
follows that the number of cells is much greater in the larger sized
species than in the smaller sized ones.

c. Sexcells. Inone type of cell, the sex cells, there is a constant
and considerable difference in size between the different species.
In all stages of the development of the ova, from the last genera-
tion of oogonia to the fertilized egg, these cells differ markedly
in size in the different species. In C. convexa, C. fornicata, and
C. plana the diameters of the oogonia at the time when they are
preparing for their last division are 27 yw, 16 wand 12 wu respec-
tively. Similarly the first generation of oocytes, before the for-
mation of yolk begins, measure 57 yu, 42 w and 36 u respectively
in these three species. These differences in the diameters of the
oocytes are associated with corresponding differences in the sizes
of their nuclei, yolk nuclei and yolk spherules, as.is shown in
table 3.

There are also differences in the manner of yolk formation in
these three species; in C. plana and C. fornicata the yolk granules
appear to be formed pretty uniformly throughout the protoplasm
of the egg; in C. convexa the yolk is formed at the base of the
cell, where it is attached to the ovarian wall, while the portion
of the oocyte next the lumen of the follicle is a cap of protoplasm
which for a long time remains free from yolk.

Finally the dimensions of the fertilized but unsegmented eggs
of four species of Crepidula and one species of Fulgur, together
with the average number of eggs laid by each mature female,
and the total volume of these eggs as compared with the body
volume, is given in table 4. There is here no constant relation
between the size of the individual egg and the volume of the adult,
though in general the species of Crepidula of smaller body size
produce the larger eggs (v. Conklin, 97). On the other hand the
size of the egg is correlated with its mode of development, the

JOURNAL OF MORPHOLOGY, VOL. 23, No. 1

168 EDWIN G. CONKLIN

TABLE 3

Size of sex cells in different species of Crepidula

Cc. CONVEXA | C. PLANA C. FORNICATA

Cell Nucleus | Cell Nucleus Cell | Nucleus
| Sukie ek COT aes Ml ie eae
Oogonia, at last divi- | |
SION: fave teins iii 15 12 6 15 12
Oocytes I, before for-
mation of yolk...... .| 57 28 36 20 42 24
Oocytes I, maximum
yolk nucleus.... ..... il 6 12
Oocytes I, maximum |
yolk spheres......... 36 15 ne,
Ootids, after fertiliza- | . |
ZALION- yea .4 Sek ee 280 142 182
Spermatocytes I... ....! 9 om || 8 6
Spermatocytes I]... ... 8 6 | 7 5
Spermatid, chromatin
condensed........... 4 ae | 3 2
Maturespermatozoa | | | |
(Head. | 24 | 1A 12
EKupyrene | Middle
] piece | 30 30 30
(Tail. . .| 110
Olkeooaene if Length 66 54 54
\ Width. | 1.5 2 2

smaller eggs producing free-swimming larvae while the larger eggs
give rise to larvae which undergo metamorphosis within the egg
capsules and escape only when the adult form has been reached.

The larger eggs have a much larger quantity of yolk than the
smaller ones, but they also have a larger quantity of cytoplasm
and larger nuclei; even in the oogonial stages, before any yolk is
formed, these cells are much larger in C. convexa than in C. plana;
indeed these oogonia differ as much in size as do the mature eggs.
It would be a great mistake to suppose that the larger eggs differ
from the smaller ones merely in the quantity of yolk which they
contain, as is usually assumed. Such a marked and constant
difference in the size of egg cells, where other types of cells are so
uniform, is significant. There is some evidence that in the species

BODY SIZE AND CELL SIZE 169

Figs. 1to3 Unsegmented eggs of Crepidula plana, C. convexa, and C. adunea,
drawn to the same scale, to show relative sizes of eggs, germ nuclei, and cytoplas-
mic and yolk areas.

170 EDWIN G. CONKLIN

with the larger eggs a considerable number of oogonia, or of fol-
licle cells, fuse to make a single egg. In the very early stages of
the oogonia, in C. convexa, there is no marked difference between
the germ cells and the follicle cells; later they differentiate and the
follicle cells become more numerous than the oogonia. Both
follicle cells and oogonia are sometimes imbedded in young
oocytes, or even in the oogonia, and in such eases the size of the
oocytes is greater and their number fewer than in other species in
which such a fusion has not been observed.

TABLE 4
Number, size and volume of eggs as compared with volume of adult female
| RATIO
SPECIES NUMBER DIAMETER | VOLUME aaa | VOLUME | VOLUME
EGGS LAID EGG EGG aaa ADULT Q EGGS TO
| VOLUME 9
LM | cu. mm. cu. mm. | cu. mm.
C. plana....:..........)ea. 9000 142 | 0.0014891° 13.4 | 667 | 1:50
CHfOrnicatigeteaov cs... ca. 138200) © 182 | 0.003156 41.65 1600 | 1:38
CHiconvexa: 2st... }ca. 220) 280 | 0.011494) 2.53 50 | 1:20
CAAGUNG Herre toc... ca. 180 410 | 0.036087 6.50 208 > | Les32
Bulgurcanear..: 4:2...|c¢a: 750) 1600°° |'2.13 1597.5 |

The adults of the several species of Crepidula are so variable
in size, color and form that it is frequently difficult to distinguish
the species; however the egg size of each species is highly charac-
teristic and constant, and by this means I have been able to dis-
tinguish doubtful species, and in one case to show that a supposed
species (C. glauca Say) is only a locally modified form of C. con-
vexa, (Conklin, 798). I know of no other animals in which the
size and form of sexually mature individuals are so variable and
the specific egg size so constant as in the genus Crepidula.

A similar, though less marked, size difference is seen inthe male
sex cells of the different species of Crepidula. In table 3 the di-
mensions of the spermatocytes of the first and second order, the
spermatids, and the mature spermatozoa, both eupyrene and
oligopyrene, are given for C. convexa, C. plana and C. fornicata.
In the case of the eupyrene sperm the tail is very long, about
110 y» in C. plana, and it is very tenuous toward the end so that
I have not been able to measure it with certainty; however it is

BODY SIZE AND CELL SIZE Ak

easy to get the average length of the head and middle piece in all
three species. These measurements show that the male cells at
all stages are larger in C. convexa than in the other species named,
thus forming a parallel case with the egg cells of this species.
Popoff (08) maintains that there are small variations in the
sizes of the sex cells of different species, caused by inequalities of
division and by unequal growth during the growth period. He
supposes that when a large egg is fertilized by a large spermato-
zoon a large individual, composed of large cells, results; whereas
if the sex cells are smaller than usual the individual developing
from them will also be smaller. Applying this hypothesis to the
ease of Crepidula, we should expect to find that C. convexa,
which has larger eggs and spermatozoa than the other species
considered in table 2, would show a larger body size and cell size
than the other species; on the contrary the size of tissue cells
is not greater, and the body size is much less in C. convexa than
in C. plana and C. fornicata. In this case it is evident that the
egg size does not determine the body size nor the cell size of the
adult, but that differences in body size are due to varying rates of
growth and cell division in the different species. It is true that
I am here dealing with different species, whereas Popoff’s hypoth-
esis applied to different individuals of the same species, but it
would be a remarkable fact if so general a proposition as Popoff’s
should be completely reversed in two closely allied species. We
have not generally regarded specific differences as so fundamental.
Popoff admits that body size may be the result of conditions
favorable or unfavorable to growth. A study of the conditions
which lead to the production of dwarfs or giants in Crepidula
plana shows that here these factors are environmental, and not
germinal; I have discussed elsewhere (Conklin ’98) this case and
will summarize it later in this paper. Undoubtedly environ-
mental conditions have much to do with body size. But in the
case of different species, each with characteristic body size, the
factors which determine size cannot be merely environmental.
In all animals there is a limit, and in most cases a clearly defined
one, to body size and consequently to cell growth and cell division.
This limit may be imposed by unfavorable environment, or by

ve EDWIN G. CONKLIN

certain intrinsic conditions, which are for the most part unknown.
In some instances there is a direct relation between egg size and
body size, as in the male and female eggs of Dinophilus, phyllox-
erans, rotifers and spiders (Montgomery ’08). On the other hand
there is marked dimegaly of the sexes in each species of Crepidula,
as shown in table 1, without any corresponding dimegaly of the
sex cells, but it is possible that protandric hermaphroditism may
sometimes occur in these species (Conklin, 798, Orton ’09). It
is well known that egg fragments produce smaller embryos than
entire eggs, and Zur Strassen (’98) has shown that from two fused
eggs of Ascaris megalocephala a giant individual may result.
According to Morgan (’04) and Chambers (’08) frogs’ eggs which
are smaller or larger than usual give rise to individuals which
are smaller or larger than the mean. All of this shows that within
a species there may be a relation between body size and the size
of the ‘Ausgangszellen.’ But at the most this is only one factor
of several which determine body size, and in many cases, as in
the genus Crepidula, the other factors are the more important
ones.

In the case of different species or varieties, even though closely
related, it is evident that egg size in general cannot be directly
correlated with body size. Here the rate and duration of .cell
growth and cell division are the most important factors in deter-
mining body size.

d. Blastomeres. In the early cleavage of the eggs of these gas-
teropods the blastomeres are, cell for cell, the same, except for
size, in all the species, whatever the size of the egg may be. The
direction of cleavage and its relation to the chief axis of the egg,
the rhythm of cleavage and the relative sizes of daughter cells, the
constitution of the blastomeres, whether protoplasmic or deuto-
plasmic, and the ultimate destination of the individual blastomeres
is the same in all the species of Crepidula, (figs. 4 to 12). In all
of them the ectomeres are separated from the entomeres as three
quartets of micromeres, which contain most of the cytoplasm of
the egg but no yolk (figs. 4 to 7); in all of them the mesomere
(4d) arises from the left-posterior macromere, and contains both
yolk and cytoplasm; in all species, the entomeres are the four

BODY SIZE AND CELL SIZE L763

Figs.4,5 Hight-cell stage of eggs of C. plana and C. convexa, drawn to the same
scale, showing relative sizes of blastomeres, nuclei, spindles, etc.

Figs. 6,7 Twenty-four-cell stage of eggs of C. plana and C. convexa, drawn to
same scale and showing relative sizes of blastomeres, nuclei, and protoplasmic and
deutoplasmic regions of the egg.

174 EDWIN G. CONKLIN

large macromeres which contain little cytoplasm and almost all
the yolk. The early subdivisions of the ectomeres take place in
exactly the same way in the largest as in the smallest eggs,
though the individual cells are larger in the former than in the
latter. When the third and last quartet of ectomeres is separated
from the macromeres, the first quartet has divided, and, in the
smaller eggs of C. plana, the second quartet also, so that the
completely segregated ectoderm consists of a plate of sixteen, or
twenty, protoplasmic cells resting upon the great yolk cells, or
macromeres (figs. 6, 7). Since this ectodermal plate contains
most of the cytoplasm of the egg, its dimensions in the different
species give a fair idea of the relative amounts of cytoplasm in
these eggs. This plate is larger in the large eggs than in the small
ones, as table 5 shows, and of course the individual cells of which
it is composed are larger in the former than in the latter. It
will be seen by consulting table 5 that the diameter of the ecto-
dermal plate is considerably greater than the diameter of the cyto-
plasmic area of the unsegmented egg; this is due in large part to
the more complete segregation of the cytoplasm in the later stage
than in the earlier one, though in part it is due to the growth
of cytoplasm at the expense of yolk, as I have shown elsewhere
(23)

It is a matter of capital importance that all differential cleav-
ages of the egg are precisely similar in number and character in
all these species of Crepidula, whatever the size of the egg may
be. Not only are all the cleavages which give rise to ectomeres,
mesomeres and entomeres the same in all species of the genus,
but all subdivisions of these cells, which are differential in char-
acter are the same in all these species, so far as I have been able to
determine. It is only in the case of non-differential cleavages
that differences in the number of cells appear in the different
species. But while the differential cell divisions do not vary
in number under normal conditions, this number is not to be re-
garded as irrevocably fixed, for it may be experimentally altered,
as I shall show in another paper.

Although the number of cleavage cells is the same in all species
during the early cleavage stages, it comes to differ greatly in the

BODY SIZE AND CELL SIZE Le,

|

ee

(
(9
ae
3
te
OR
p D./ =
H

Figs.8to10 Corresponding stages in the eggs of C. fornicata, C. convexa, and C.
adunea, drawn to the same scale. The blastomeres correspond cell for cell except
that one additional ectoderm cell (the basal cell in the posterior arm of the cross,
shaded by transverse lines) is present in C. adunca (fig. 10) which is not present
in the other species. There are present: 39 ectoderm cells (40 in C. adunca),
6 mesoderm cells, 7 endoderm cells.

176 EDWIN G. CONKLIN

later stages, the number of ectoderm cells being greater in the
larger eggs than in the smaller ones. Up to the 52-cell stage the
number of cells is the same in the eggs of all the species examined;
at this stage one additional ectoderm cell appears in the posterior
arm of the ‘cross’ in C. adunca which does not appear until later
in the other species (the additional cell is the one shaded by trans-
verse lines in fig. 10). At the 82-cell stage four such additional
cells are present in C. adunca, two in the posterior arm of the
cross and two in the posterior ‘turret’ cells, all the other cells
being the same in all the species (figs. 11, 12). These additional
cells of the posterior arm of the cross and of the posterior
turret cells are all similar histologically as well as in ‘prospective
significance’ with the cells from which they were derived. ‘They
all become the large ciliated cells of the ‘posterior cell plate’
(Conklin, ’97, p. 109), and ultimately give rise to the large cells
which form the head vesicle of the larva. In the later stages of
cleavage many additional ectoderm cells appear in the larger
eggs which are not present in the smaller ones. Most of these
additional cells appear in the primordia of organs after these have
been established, and consequently represent only an increase in
cells of a given kind. These primordia form chiefly on the oral
side of the egg, whereas the greater part of the aboral hemisphere
is covered by the large cells of the posterior cell plate. As a con-
sequence the number of cells visible from the aboral pole at the
period of the closure of the blastopore does not differ greatly in
number in the different species; whereas the cells visible from the
oral pole differ greatly in number in the different species, there
being more than three times as many in C. adunca as in C. plana
at the time when the blastopore closes, and in later stages this
disporportion becomes much greater.

Table 5 gives the diameter of the egg and of the ectodermal
plate in different species of Crepidula, at the time when the ecto-
meres are first separated from the macromeres; also the approxi-
mate number of ectoderm cells, visible from the oral and the
aboral poles at the time of the closure of the blastopore, in the
different species.

BODY SIZE AND CELL SIZE 177

\\\

Ze
d

v \

Uy
fy
y

4

Figs. 11, 12 Corresponding stages in the eggs of C. fornicata and C. adunca,
- drawn to the same scale. 1n the former there are present 62 ectoderm, 10 meso-
derm and 7 endoderm cells; in the latter 66 ectoderm, 10 mesoderm, and 7
endoderm cells. The 4 additional ectoderm cells in C. adunea are 2 additional
posterior turret cells and 2 posterior basal cells, shaded by transverse lines.

178 EDWIN G. CONKLIN

TABLE 5

Cell size and cell number in the development of Crepidula and Fulgur

1-CELL STAGE | 24-cELL STAGE CLOSURE OF BLASTOPORE
SPECIES Diameter | pe ee Number of cells visible
Diameter of proto- | Diameter Wi ecrsrrisee = ne
of egg Bee | of egg plate Oral pole | Aboral pole
ax |
Ml Ke | ML be
@appllam are vac sald os (Cl eee meals 65) ca. 134) ca. SO 164, 124
C. fornicata...........| ca. 182) ca. 86] ca. 182|/ ca. 100] 262) 140
| re | | |
(SICOMVEK Boa. = = os-% 5, che ca. 280) ca. 90] ca. 280) ca. 144! 430 146
RO PMA CICA) Fo orc as oy oh ea. 410) ca. 120] ca. 410) ca. 200 530-164
. va rry| a DY: fod ©
Fulgur carica..........| ca. 1600 ca. 160) ca. 1600) ca. 320) ca. 5000 ?

This table shows that although there are many more ectoderm
cells in the larger eggs of Crepidula than in the smaller ones, this
inereased number is chiefly confined to the oral pole, where the
primordia of the various organs are located, and it also indicates,
as was emphasized above, that the increase in the number of cells
in C. adunca as compared with C. plana, is due to a greater num-
ber of non-differential divisions in the former species after the
primordia have been established.

It seems probable that the more frequent divisions of the ecto-
meres in C. adunca as compared with those of C. plana is associ-
ated with their larger initial size, but at present it is not possible
to determine why the smaller cells of the latter species continue
to grow and divide for a longer period than the larger cells of the
former.

e. Larvae. Finally the body size and cell size of fully formed lar-
vae of C. plana, C. convexa, and C. adunca are given in table 6.
In the case of C. plana the larvae measured were of maximum size
and were ready to escape from the egg capsules; in the other
species the larvae were of a corresponding stage of differentiation,
with velum and larval organs of maximum size, though these
larvae undergo metamorphosis within the egg capsules and do
not escape until they are adult in form. The body volume of
these larvae was roughly determined by measuring the length and
breadth of the body as seen from the dorsal side, the thickness of
the body being approximately the same as its width.

BODY SIZE AND CELL SIZE 179
TABLE 6

Body size and cell size of fully formed larvae of Crepidula

|

Cc. PLANA Cc. CONVEXA Cc. ADUNCA

Dimensions of body be bp ue

Datel wise ae Bech Rae an ve 240 400 550
“y Bre ac Ghistyrsee sete o.oo ae 230 335 480

Relative volumes. 22.2... ee. noe eee 1 3 10

Dimensions of cells |
OesopWAaGUSse etree sh ate ya bok eer Gx 12 9x 12 6x 15
Stomach (oylOris)) ca aseh 22 a2yc she ee = 6 x 12 xs 5 O15
Hoem@epttnelmumM) 0.2.23... c605. se Pe 4 x 27 4x 24 4x 30
Cie Mame: Ase A Pe See 6sal2 5 x 15
REGIA (DOSUsaWiall)) mets a-flas tsk Sete ew bh Gexell5 6 x 18
Srexe Galle (OY. ae Gere yeas eee 5 6

While the relative volumes of the unsegmented eggs of these
three species are as 1 : 7 : 24, the relative volumes of the larvae are
as 1 :3:10; in short, the growth of the embryo and larva of C.
plana has been more rapid than that of C. convexa and of C.
adunea, so that the great disproportion which existed between the
eggs of these species at the beginning of development, has begun
to disappear.

Likewise the cell dimensions of the larvae of these three species
_show that the great disproportion in size of the early cleavage cells
of C. plana, as compared with C. convexa or C. adunca, has begun
to disappear. The cells of the larvae of C. plana are but little
smaller than those of C. convexa and C. adunca, but they are
much fewer in number. Even among the larvae differences in
body size are due chiefly to differences in cell number, rather than
to differences in cell size, just as is true of adults.

Evidently growth in volume during embryonic stages has been
more rapid in C. plana than in the other species named. This
may be due to a greater absorption of water on the part of the
embryo of C. plana; and in accordance with this suggestion it
may be said that the cytoplasm of this embryo is less dense
and stains more faintly than that of the embryos of the other
species. Davenport (’97) showed, long ago, that the increase in
bulk of the frog embryo and larva, up to the time when it begins

180 EDWIN G. CONKLIN

to take food, is due to absorption of water; and in the embryos
and larvae of Crepidula it seems probable that varying rates of
growth may be due to this same factor. This increased growth
of the embryo of C. plana as compared with those of the other
species, continues through the larval and post-larval stages, so
that in the end the adults of the first named species become much
larger than those of the latter. On the other hand the number
of cell divisions during embryonic stages is much greater in C.
convexa and C. adunca than in C. plana, so that the embryos
‘and larvae of the former species are composed of a much greater
number of cells than in the case of the last named species.
But in post-larval stages cell divisions -become much more
numerous in C. plana than in C. convexa or C. adunca, so that in
the adult condition the first named species contains a very much
greater number of cells than either of the others.

Considering only the intrinsic factors of growth, one may say
that the size of an embryo is the result of the initial size of the egg,
but the size of an adult individual is due for the most part to the
duration and rate of cell growth and cell division. The size of the
germ cells is of little significance in determining the body size of
adult individuals in different species of Crepidula, though it may
possibly be of some importance in determining the body size of
different individuals of the same species and race, when all other
conditions are equal. In different races and species the important
factor in determining body size is the duration and rate of cell _
growth and cell division. Many extrinsic factors are known to
limit or promote such growth and division, but in the case of
species which differ greatly in size and in which body size is a very
characteristic feature it would seem necessary to suppose that
size is inherited,—that some of its causes must be intrinsic
ones. What these intrinsic factors are which limit cell growth
and division in some cases and which promote such processes in
others, is at present entirely unknown. However it seems prob-
able that upon such factors depends in the main the difference
between a mouse and an elephant, while the initial size of the
sex cells is of only minor importance.

BODY SIZE AND CELL SIZE 181

2. CELL SIZE AND BODY SIZE IN TYPICAL AND DWARFED
INDIVIDUALS OF CREPIDULA PLANA

I have already discussed in detail an interesting case of en-
vironmental dimorphism in C. plana (Conklin, ’96, ’97, ’98), but
since this work has apparently escaped the attention of other
workers on this subject I shall here quote from one of these papers
(98) at some length. The ordinary or typical form of

Crepidula plana is found most abundantly in dead shells of Neverita
inhabited by the large hermit crab, Pagurus polycarpus. In this posi-
tion individuals grow to a large size, mature females frequently reaching
a length of two inches and a breadth of one and one-quarter inches.
A dwarf race of C. plana occurs in the dead shells of Nassa or Littorina,
inhabited by the little hermit crab, Pagurus longicarpus; the largest
individuals of this race never exceed three- quarters inch in length and
three-eighths inch in breadth, i.e., they are about one-third the linear
dimensions of the larger form.

There is good evidence that these dwarfs are not a permanent variety
or race. In the first place there are no anatomical differences between
the two varieties save size only; secondly the eggs, embryos and larvae
of the two cannot be distinguished; . . . . finally, a few specimens
were found which showed by the shape and character of their shells
that at one time they had been typical dwarfs; afterwards, having been
detached, they obtained a new foothold on a larger surface, and their
shells increased i in size, the new portions of the shell being shaped sO as
to fit the surface upon which they had found anew home. In every such
shell one can recognize both the dwarf and the normal forms. The
dwarfs are what they are by reason of external conditions, and not be-
cause of inheritance; they are, in short, a physiological and not a mor-
phological variety.

The average body volume of a mature female of C. plana is 3 ce.,
while the average volume of a mature female of the dwarf variety is 34)
cc., 1.e., the average body volume of the typical form is about thirteen
times that of the dwarf. This disproportion in size would be much
greater if comparison were made between the largest individuals ob-
tainable in the two classes, since the dwarfs are much more uniform in
size than the type forms.

The dwarfs are perfectly formed in all respects, and all organs of the
body seem to be reduced in about the same proportions. It is interest-
ing to note that certain organs, or parts of organs, which are formed in
considerable numbers in the course of development, are reduced in num-
ber but not in size in the smaller individuals;? this is true of the number

“In agreement with these observations are the experiments of Miss Peebles
(97) on the regeneration of small pieces of hydra; in such cases one, two, three
or four tentacles are formed, depending upon the size of the regenerating piece.

182 EDWIN G. CONKLIN

of gill filaments, and the number of lobules of the liver and ovary. The
number of gill filaments in three individuals, which differed greatly in
size, was as follows:

Mature female... .......... Volume of body, 0.75 cc., Gill filaments, 204
Immature female.......... Volume of body, 0.05 cc., Gill filaments, 53
Dwarf female (mature)..... Volume of body, 0.05 cc., Gill filaments, 58

Table 7 gives the dimensions of certain tissue cells from sexually
mature individuals of C. plana of widely different body size. All
cells measured are from corresponding parts of similar organs.
In each instance the dimensions given represent an average of
about one hundred cells from at least four different individuals of
approximately the same size. I am indebted to Mrs. Anna N.
Bigelow, one of my former students, for the care with which she
performed the laborious task of making these many measurements.

With the exception of the ganglion cells and the muscle fibers,
the differences in cell size in these different individuals is slight

TABLE 7

Size of tissue cells of mature individuals of Crepidula plana of different body size

VOLUME OF BODY
Giant 9? | Dwarf 2 Typical 6
ON7o—- 81eGr | 0.05 ee. | 0.05 ce.
Me Me Me
Ectodermal epithelium | |
Mantle? NCAT ANUS,..!..0. Argan a eee 6x 13 Gala 7) Gexals
Gill chamber, epithelial lining.......... 6x 11 by oe IU 6x 12
HGGrepIbne GMs 214.5 selec 6x 14 Ge 1b" |Get
Gill filament, external ciliated cells....) 4x 15 C00 oui ee aes we od
Gill filament, cells internal to chitinous |
1206s bes os ee a ORR REEMA ns cic 5) 38 Ils) 4x12 4x12
Osphraaims 0.26.) eee ee 3 x 75 4x 75 4x78
Ganglion cells, largest in pedal ganglion | 17x23 | 9x20 | 10x 20
Endodermal epithelium |
Ravieriducititennte sts apcdke hl oe eee ete | 6x21 6 x 20 6x 18
Liver cells, withsecretion...............| 15x 59 16 x 58 17 x 66
Stomach opposite liver duct............ 13 x 54 12 x 54 11 x 43
IRCCUUMING pero: «ie nike eee 10 x 28 UNSs Py 10 x 25
Mesoderm cells
IWidmercellissaenn a. ot. on te eee eee 13 x 38 16%x 38) 4) 13.34
Muscle fibers from foot, diameter........ 6 | 5 5

BODY SIZE AND CELL SIZE 183

and insignificant, and these results, which were briefly reported
in both of my former papers on this subject (’96, 98), show that
differences in the body size of different individuals are due to the
number of cells present rather than to the size of individual cells.
To quote again from one of the papers referred to (’98, p. 438):

It is an almost impossible task to count the number of cells present
even in a very small organ. I have, however, been able to count the
number of cells present in cross sections of the rectum, and while the
size of the cells here, as everywhere, is the same in the large as in the
small individuals the number of cells is greater in the former than in the
latter.

Of all the cells of the body, the ova are most easily enumerated; they
are laid in capsules which can be easily counted, and each of which con-
tains a nearly constant number of eggs. Oft repeated observation shows
that without exception the fertilized, but unsegmented, eggs of the
dwarfs are of exactly the same size as those of the giants, but are very
much fewer in number; e.g. table 8 shows the averages obtained from
a larger number of observations.

It is notable that the number of capsules formed is nearly the same in
the two varieties, though there is a great difference in the number of
eggs inclosed in each capsule.

In Crepidula, therefore, the cell size is fairly constant, and variations
in the size of the body are due to variations in the number of cells present.

. Whatever the cause of the dwarfed form may be, it will be
noted that i in Crepidula it operates by stopping growth and cell division.

3. CELL SIZE AND BODY SIZE IN MALES AND FEMALES OF
C. PLANA

Marked as is the environmental dimorphism in C. plana, the sexual
dimorphism is even greater (table 1), the average female being almost
fifteen times as large as the average male. In all species of Crepidula
the males are smaller than the females, though the difference in size is
greatest in C. plana.

TABLE 8

Size and number of eggs laid by typical and dwarfed individuals of Crepidula plana

| | |
DIAMETER OF | NUMBER OF EGGS IN | TOTAL NUMBER

EGG | CAPSULES CAPSULES | OF EGGS
4 = i , | ntadiak cic ta
Cy plana (ty pele by ge a 51 176 9,000
¢ C. plana (dwarf). Wetec | 136* | 48 | 64 | 3,070

*More recent Bee ememionte. aan nae een gale ee. Ganee lenses, alow.
the eggs to be about 142 u in diameter before the first cleavage, as given dkewhere
in this paper.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 1

184 EDWIN G. CONKLIN

In the case of the males, as in that of the dwarfs, the smaller size of the
body is due to the smaller number of cells present rather than to the
smaller size of the cells. Careful measurements of the cells of the intes-
tine, stomach, liver, kidney, muscles of foot, epithelium of gill chamber,
epithelium of gill filaments, etc. show that the cell size remains the same
in the male as in the female (table 7). Whatever the ultimate cause of
the smaller size of the males may be, it operates in this case as in that of
the dwarfs, by causing a cessation of growth, and cell division.

It seems probable from the observations of Orton (10) as well
as of myself (Conklin ’98) that the small males of the genus Cre-
pidula may sometimes grow into the larger females, and that we
have here a case of protandric hermaphroditism. If so the small-
ness of body size and cell number in the males of this genus may
be considered to be youthful characteristics.

I have found no evidence that the difference in the size of adult
males and females is associated with differences in the size of the
eggs as is the case in rotifers, phylloxerans, and Dinophilus, and
if protandric hermaphroditism occurs in this genus, such dimor-
phism of the egg would not be expected.

4. CONCLUSIONS

In the genus Crepidula differences in body size may be very
great; the volume of the average male of C. fornicata is one hun-
dred and twenty-five times that of the average male of C. con-
vexa; and the volume of the average female of C. fornicata is thirty-
two times that of the average female of C. convexa. Within the
single species, C. plana, the volume of the average female is about
fifteen times that of the average male and about thirteen times
that of the dwarf female of the same species.

In spite of these great differences in body size, the size of tissue
cells is approximately the same in all species examined, and in all
individuals of both sexes and of very different sizes. In the main
differences in body size are due to differences in the number of
cells persent, and not to variations in the size of individual cells.
Ganglion cells and muscle cells form the principal exception to this
rule.

These results agree with most of the work which has been done
on cell size in relation to body size, and particularly with the re-

BODY SIZE AND CELL SIZE 185

sults of Levi (05). On the other hand Berezowski (’10) finds that
the size of intestinal epithelial cells is smaller in young mice than
in older ones, and that with the general growth of an animal there
occurs a growth in the height of these cells. However this obser-
vation does not contradict the conclusion reached in this paper;
indeed it is true of Crepidula, as of the mice studied by Berezow-
ski, that younger animals have smaller cells than older ones, as
will be seen by comparing the size of larval cells given in table 6
with that of adult tissue cells given in tables 2 and 7. It is well
known that the size of cells depends to a certain extent upon the
rate of cell division and the length of the resting period, and the
rate of division is slower and the resting periods longer in mature
animals than in young ones. In all my measurements I have,
so far as possible, compared animals of the same stages, so that the
developmental changes in the size of cells does not materially in-
fluence my results.

But while tissue cells maintain.a very uniform size in Crepidulae
of all species and sizes, provided they are of corresponding ages,
the sex cells differ enormously in size and number in the different
species, the species of small body size having in general larger and
fewer eggs than species of larger size. On the other hand within
the same species the sex cells are of approximately the same size
in all individuals, but they differ in number in animals of different
body size, just as the tissue cells do.

The larger eggs of C. convexa and C. adunca are larger in every
respect, having more cytoplasm as well as more yoik than the
egos of C. plana and C. fornicata. Even in the oogonia and early
oocytes, before yolk begins to form, the eggs of the former species
are larger than those of the latter. The spermatozoa and sperma-
tocytes of C. convexa are also larger than those of C. plana or
C. fornicata. Presumably the sex cells of C. convexa are larger
from the time of their first appearance, and it is possible that this is
due to their being derived from larger blastomeres, as well as to
the fact that the primitive sex cells divide less often in species
with large eggs than in those with small ones. It seems probable
also that the oogonia or oocytes ingulf a larger number of oogonia
and follicle cells in C. convexa than in C. plana.

186 EDWIN G. CONKLIN

The larger eggs give rise to larger blastomeres and to larger
embryos and larvae than do the smaller eggs. The sizé of tissue
cells is nearly the same in the larvae of the different species, and
this size is less than that of adult tissue cells; but the number of
cells in the larvae of different species differs greatly being approx-
imately proportional to the body volume of the various larvae.
Finally cell growth and division continue for a longer period in
species of Crepidula which have small eggs than in those which
have large ones, with the result that the former give rise to larger
adults than the latter. In the different species of this genus the
size of the germ cells does not determine the size of the adult
(Popoff, Chambers).

Within the same species differences in body size are due in the
main to differences in cell number, the cell size being approxi-
mately constant. Small individuals are as complete and perfect
as large ones, all the typical differentiations and organs being
present in the former the same asin the latter. But in parts which
are reduplicated, such as the gill filaments, lobules of liver, kid-
ney, ovary and testis, etc., these parts are more numerous in
large individuals than in small ones. In the parts which are
reduplicated, whether they be organs or cells, there is practically
no differentiation between the different members. Increase in
size is due to an increase in the number of these parts or cells,
without any increase in the total number of the various kinds of
differentiations.

On the other hand, differential cell divisions, such as are found
in the early cleavages of the egg do not vary in number in eggs
or embryos of different size. The study of the cell lineage of these
gasteropods shows that the cleavage is cell for cell the same in
eggs and blastomeres of all sizes.and species until ectomeres, ento-
meres and mesomeres are completely separated, and until differ-
ential divisions have given place to non-differential ones. So far
as cell division is associated with differentiation and morphogene-
sis in the cleavage period, the number and character of these
divisions do not vary in different species or individuals; so far
as it is associated with growth, and the ‘vegetative duplication of
parts,’ but not with differentiation, it may vary enormously. Pro-

BODY SIZE AND CELL SIZE 187

fessor Whitman’s (93) statement that, ‘“The organism dominates
cell formation, using for the same purpose one, several or many
cells,” is true within certain limits. But the differences in cell
number which are unimportant are only those which are asso-
ciated with growth and not with differentiation, with trophic as
contrasted with morphogenetic processes.

Supplementary Note. After this paper had gone to press I
received an important publication by S. Morgulis entitled
“Studies of Inanition in its Bearing upon the Problem of
Growth,” Arch. f. Entw.-Mech., Bd. 32, 1911. The author
finds that both thé size and number of cells are decreased in
starving animals. Experiments of my own on starved planarians
yield the same result.

LITERATURE CITED

Ametuna, E. 1893 Ueber mittlere Zellengréssen. Flora, Bd. 77.
BrerezowskI, A. 1910 Studien ueber die Zellgrésse, I. Arch. f. Zellforschung,
Bd.75:
1911 Studien ueber die Zellgrisse, II. Idem. Bd. 7.
Boveri, Tu. 1904 Ergebnisse ueber die Konstitution der chromatischen Sub-
stanz des Zellkerns. Jena.
CuamBeERS, RoperT 1908 Einfluss der Eigrésse und der Temperatur auf das
Wachstum und die Griésse des Frosches und dessen Zellen. Arch. f.
mik. Anat., Bd. 72.
Couton, H. 8. 1908 Some effects of environment on the growth of Lymnaea
columella. Proc. Acad. Nat. Sci., Philadelphia.
Conxuin, E. G. 1896 Cell size and body size. Abstract of paper read before
Amer. Morph. Society. Science, January 10.
1897 The embryology of Crepidula. Jour. Morph., 13.
1898 Environmental and sexual dimorphismin Crepidula. Proc. Acad.
Nat. Sci. Philadelphia.
1912 Cell size and nuclear size. Jour. Exp. Zool., vol. 12.
Davenport, C. B. 1897 The rdéle of water in growth. Proc. Boston Soc. Nat.
Hist., vol. 28.
Dr Variany 1892 Experimental evolution. Macmillan, London.
Donaupson, H. H. 1895 The growth of the brain. Scribners, New York.
Driescu, H. 1898 Von der Beendigung morphogener Elementarprozesse.
Arch. f. Entw. Mech., Bd. 6.
1900 Die isolierten Blastomeren des Echinidenkeimes. Idem., Bd. 10.
Gates, R. R. 1909 The stature and chromosomes of Oenothera gigas, DeVries.
Arch. f. Zellforschung, Bd. 3.

188 EDWIN G. CONKLIN

*Gauur, J. 1889 The number and distribution of the medullated fibers in the
spinal cord of the frog. Abh. d. Math.-physiol. Cl. d. konigl. Sachs.
Gesell. d. Wissenschaft.

Harpesty, 1. 1902 Observations on the medulla spinalis of the elephant, etc.
Jour. Comp. Neur., vol. 12.

Jenninas, H.S. 1908 Heredity, variation and evolution in the Protozoa. Proc.
Amer. Philos. Soc., vol. 47.
1909 Heredity and variation in the simplest organisms. Amer.
Naturalist, vol. 43.
1911 Assortative mating, variability and inheritance of size, in the
conjugation of Paramecium. Jour. Exp. Zool., vol. 11.

JENNINGS, H. S. and Harerrt, G. T. 1910 Characteristics of the diverse races
of paramecium. Jour. Morphology, vol. 21.

Levi, G. 1905 Vergleichende Untersuchungen ueber die Grdsse der Zellen.
Verh. anat. Ges., Bd. 19.
1905 Studi sulla grandezza della cellule. Arch. ital. anat. embriol., T.5.

Minot, ©. S. 1908 Age, growth and death. Putnams, New York.

Monreomery, T. H., Jr. 1906 The oviposition, cocooning and hatching of an
Aranead, Theridium tepidariorum. Biol. Bull. vol. 12.

Moraan, T. H. 1895 Studies of the ‘partial’ larvae of Sphaerechinus. Arch.
f. Entw. Mech., Bd. 2.
1896 The number of cells in larvae from isolated blastomeres of ‘Am-
phioxus. Idem, Bd. 3.
1904 Relation between normal and abnormal development of the em-
bryo of the Frog. II]. ldem. Bd. 18.

Orton, J. H. 1909 On the occurrence of protandric hermaphroditism in the
molluse Crepidula fornicata. Proc. Royal Soc., vol. 81.

PEEBLES, FLoRENCE 1897 Experimental studies on Hydra. Arch. f. Entw.
Mech., Bd. 5.

Pororr, M. 1908 Experimentelle Zellstudien. Arch. f. Zellforschung, Bd. 1.

Ras, C. 1899 Ueber den Bau und die Entwicklung der Linse, III. Zeit. wiss.
Zool., Bd. 47.

Sacus, J. 1893 Physiologische Notizen, VI. Ueber einige Beziehungen der
specifischen Grosse der Pflanzen zu ihrer Organization. Flora, Bd. 77.

Semper, C. 1876 Animal life as affected by the conditions of existence. Apple-
ton’s, New York.

SrrasBurGER, E. 1893 Ueber die Wirkungssphire der Kerne und die Zell-
grosse. Histolog. Beitriige, Bd. 5.

Wuirman, C.O. 1893 The inadequacy of the cell theory of development. Jour.
Morph., vol. 8. ~

Zur STRASSEN, O. 1898 Ueber die Riesenbildung bei AscarieeBienn Arch. f.
Entw.-mech. Bd. 7.

*Quoted from Donaldson (’95).

ON THE STRUCTURE OF CLINOSTOMUM MARGI-
NATUM, A TREMATODE PARASITE OF THE
FROG, BASS AND HERON

HENRY LESLIE OSBORN

From the Biological Laboratory of Hamline University, St. Paul, Minnesota

SEVENTEEN FIGURES

PREFATORY. NOTE

In a previous article an account was given of the distribution
in this country and Canada of Clinostomum marginatum to-
gether with some notes on its habits. A short time after the pub-
lication of that article (Osborn, ’11) Professor Linton informed
me that he had recently found specimens of Clinostomum mar-
ginatum in brook trout which were taken from Alder Lake, a
private preserve in the Catskill Mountains in New York. The
conditions under which the trout live are well described by
Linton (10). ‘“‘It is a lake of forty acres in the heart of the moun-
tains. The owner maintains a well equipped hatchery on the
stream below the outlet and allows no other fish than trout in
the lake.” It is thus clear that the infection takes place in the
lake, or, in other words, that the first stages of this worm and
its primary host are to be found there. The lake is visited by
fish-eating birds and thus we can readily account for the intro-
duction of the parasite. As pointed out in my previous article,
we possess no information as to the early stages in the life history
of Clinostomum. We do not know its first host nor anything
about its development. It is evident from the facts now known
as to the occurrence of the parasite at Alder Lake that the infec-
tion must come from some form living in that lake, very likely
some invertebrate serving as food to the trout. Occurrence in
a small lake narrows down the problem of discovering this missing

189

JOURNAL OF MORPHOLOGY, VOL, 23, NO. 2
JUNE, 1912

190 HENRY LESLIE OSBORN

chapter in the life history of our subject to very workable linits.
More favorable conditions for a study of the point could hardly
be imagined. Professor Linton’s communication also adds a rew
host and a new locality to our knowledge of the distributior of
this animal.

The present paper gives an account of the organization of
Chnostomum marginatum. In justification of this when two
accounts are already extant I may plead the fact that neither cf
them are fully adequate and in some points both are erroneous.
Clinostomum is a parasite of some of our most desirable game and
food fishes and it is especially obnoxious because it is lodged in
the edible portion of its host. In order to keep the paper within
reasonable size I have left out many histological items and it is
hoped that later, when certain points have received additional
study, a further account of the histology may be published.

EXTERNAL FORM AND DIMENSIONS

The outer form of this, as of most trematodes, is extremely
changeable. It has therefore seemed best to give a description
of the form and measurements of worms after fixation. There is
little difference in form and proportions of body between the late
immature stages from cysts in the fish and frog and mature worms
from the heron. The encysted worms appear to average very
little smaller. Figs. 1 to 4 enable one to obtain an idea of the
form of the animal. Fig. 1 is from a worm killed under compres-
sion, which, after carmine staining, has been mounted entire.
Figs. 2 and 4 are from horizontal and transverse series; they show
parts which are on different planes as if they were on the same
level and need to be checked by the transverse sections shown in
fig. 3. Fig. 3 shows transsections from seven levels, all drawn to
the same scale. They are from a series of about one thousand
sections and the accompanying number is that of the section in
the series, and shows, though only very roughly, the distance of
the sections from each other.

The body is subdivided into two regions separated at the level
of the ventral sucker. Anteriorly it is almost cylindrical, its
cross section being an ellipse, posteriorly it broadens considerably

STRUCTURE OF CLINOSTOMUM 191

and is frequently somewhat concave on the ventral surface. In
living animals the posterior region of the body at times becomes
momentarily flattened and may thin out to a sharp lateral fin
but this is a merely momentary form followed by the thicker
form seen in the sections. The constriction of the body at the
level of the ventral sucker is shown in Wright’s figure (’79, fig. 1)
and in that of Linton (’98, pl. 44, fig. 6). It is also found in the
other species of the genus as can be seen in several of the forms
figured by Braun (00).

The dimensions of thirty-nine individuals were obtained from
worms mounted whole in balsam and drawn in outline with the
camera lucida. The measurements were taken both from mature
heron material and from bass worms.

These figures correspond fairly with those previously published
except that the minimum one for length is the least thus far re-
corded. A part of the differences can doubtless be attributed

TABLE 1

Showing the length and width in millimetres of thirty-nine individuals of
C. marginatum

FROM BASS FROM HERON FROM HERON
NEE COREA CE. | FIXED IN CORROSIVE a Ua ca Gane) ACID
Taupe Wide, || Length | Width | Length | Width
31 1.2 HD) MMe ger) | eae
3.5 igs 3.0 lore #120) 5.5 | 1.5
3.8 is SB3oth) 1.0 6.0 MW
4.0 16 ah 1) 6.5 5s
4.0 1.6 4.0 0.7 6.5 2.0
4.0 1.6 4.0 0.9 7.0 Ise
4.0 1.6 eete0 1.0 | 720 i
4.1 1.6 4.1 1.0 T2 1 Gey;
4.5 & Lgl 4.5 1.0 | va) 128
5.0 ee 5.0 (ae Oe ie 17
5.5 1.5 5.0 1.2 Po ert 18
5.2 ie? 8.2 1.9
5.5 fez |
| ir rG:0 22 |
| | 6.5 il 5)
| 6.5 7

192 HENRY LESLIE OSBORN

to the great mobility of the animal, but the series is too regular
to be wholly due to mere differences of degree of extension of
animals of a constant length and indicates also the existence
here of a length variation such as is common in all animal groups.
They furnish further an indication that the encysted worms,
which are slightly younger, are smaller than the heron worms.
The average length of the eleven specimens from the bass is only
4.1 mm. while that of the sixteen worms from the heron, fixed
with the same reagent, is 4.59 mm. It is interesting to note the
larger figures for the chromic acid material. The average length
of these individuals is 6.77 mm.

The form of the anterior end of the body is remarkable. In
many distomes the walls of the body converge anteriorly and meet
at the mouth, here they run parallel until they intersect the mar-
gin of a peculiar area, the oral field, which closes the anterior end
of the body. In an animal in which the oral field is in the resting
position, as in fig. 4, it is oblique to the axis of the body, with the
dorsal side projecting somewhat beyond the ventral. It is this
obliquity to which Leidy’s generic name alludes. Fig. 1 shows
the margin of the oral field where it meets the side wall. Often
there is a shght depression in the margin of the field at this point.
In fig. 5 the field is retracted, a very frequent act of the living
animal; this section is from a specimen which was caught by the
fixing reagent in this act. A fuller account of the oral field struc-
ture will be given later.

The ventral sucker (figs. 1 w, 3 C, and 4) is a very conspicuous
organ, both in the whole animal and in sections and is much larger
than the oral sucker. In all living and preserved animals which
I have seen it is entirely enclosed within the contour of the body.
Its opening is always very distinctly visible and is usually tri-
angular, with one of the equal sides anterior and the apex pos-
terior. In some cases, however, as in fig. 1, the three angles are
rounded, or the opening (as in Linton, ’98, pl., 44, fig. 6) may be
circular or even almost square (Wright, ’79). The sucker has
a length of 0.7 mm. and the same width and measures 0.4 mm.
dorsoventrally. It is about half as thick as the body and reduces
its space very much, as shown in fig. 4 C. The reason for the

STRUCTURE OF CLINOSTOMUM 193

ereat size of the ventral sucker has not been indicated by the
behavior of the animal. I have not seen it used at any time. Its
histological structure is such that it would seem to be fully func-
tional and its great size indicates a function of considerable
importance but no activities have been observed in connection
with it.

POSITION OF THE GENITAL PORE

MacCallum (’99, p. 699) states that ‘‘at about the middle por-
tion of the body behind the ventral sucker the genital openings
[italics mine] are seen, close together, that of the female apparatus
being directly in front of the male.’ Also on page 703 he says
that the female genital opening is located ‘‘directly in front of
the male genital pore.” These statements certainly imply that
there are two genital pores, a condition not found elsewhere in
trematodes. The statements are however contrary to fact and
are not consistent with MacCallum’s figs. 3 and 7 where a single
genital pore is clearly shown, so that it is difficult to see how they
crept into his paper. The exact position of the genital pore was
determined for twenty-two individuals, the data for which are
shown in table 2. .

The total length and width/measurements are given and the
distance from the anterior end to the genital pore. In order to
make direct comparisons possible the position of the pore in per-
centage of total length is given in the column on the right. The
opening is thus shown to lie posterior to the center of the body in
every instance and to vary between 53.7 per cent and 68.3 per
cent as extreme limits. A part of this difference may perhaps be
attributable to individual differences of contraction or reagent
action but in addition to these we must attribute it in part to
variation in the actual position of the pore. If we take the aver-
age of these figures we should have 56 per cent as the point of
location. The fact that some of the worms of this table show a
greater length than any given in table 1 is because they are speci-
mens killed under compression and are consequently unnaturally
elongate. I have however admitted them to this table, as the

194 HENRY LESLIE OSBORN

TABLE 2

Measurements used in determining the position of the genital opening

CATALOGUE NUMBER | TOTAL LENGTH OF | GREATEST WIDTH | Pees lect a SOrAL GanCTEe
OF SPECIMEN SPECIMEN OF SPECIMEN | GENITAL OPENING TO bee
mm. . mm. mm.

b. 3:8 2 | 25 | 65.7
oe 4.4 | 1.3 | ORS 63.6
TREE Fare odes 6.0 | 1.5 | 3.4 56.6
750) ee 6.0 ou | 3.4 | 56.6
TRS oft s 6- fe ths) | 6.2 | 1.6 | 3.8 61.2
Un ee | 6.7 2.3 3.6 530%
ead oe ee 6.8 1.75 4.1 | 68.3
g.. | 6.9 1 | 3.8 |< 55eO
She ee « 6.9 1.75 | 3.8 | 55.0
aloe! | 6.9 2.0 Bas 55.0
he | 7.0 ies 3.8 55.0
Ny | ee, 18 4.2 | 58.3
762.. 7.5 1.8 4.2 56.0
750.1.2 Ti 2.0 4.25 56.6
te dvi 7.5 253 4.3 57.3
(iD. pee, ea 7.75 | 1.75 4.5 58.0
ley ks 1.9 4.5 57.7
ce UR ci 7.9 | 1.75 4.6 58.2
mee .| 8.3 2.1 4.9 59.0
oh tee eee | S18 Dia 5.3 60.2
750.1 | 9.2 28 5.1 55.4
TOES 5s 10.8 3.0 i i 56.4

compression may be supposed to have acted equally in all direc-
tions and so has not influenced these results.

A comparison of the figures of the different species of this
genus as given in Braun’s paper (00) shows that the position of
the genital pore differs very much in them, it being very near the
posterior end of the body in C. heluans (fig. 10 a) 80.3 per cent,
and very posterior also in C. dimorphum. It is nearly 79 per
cent in the maximum case of C. marginatum of the table just
given.

The genital pore opens into a chamber, the atrium (fig. 3)
in which the male and female genital systems end. The opening
of the terminal part of the uterus, the ‘metraterm,’ lies in this
atrium anterior to the position of the cirrus of the male system.

STRUCTURE OF CLINOSTOMUM 195

This fact corresponds with the statements of MacCallum except
in so far as he gives the impression that these openings are located
on the outer surface of the body. The excretory. pore opens (as
shown in fig. 4) dorsally very near the posterior end of the body.

THE STRUCTURE OF THE BODY WALL

In general the trematode body is encased in a wall made up of
a non-cellular cuticula, which may or may not be spinous, rest-
ing upon an outer zone of the parenchyma in which muscles run
in various directions. For convenience we may consider the
oblique muscles as marking the inner boundary of the wall though
there is no break in histological structure at that point. The
fibers of the oblique muscles lie in groups considerably spaced
from each other so that the central parenchyma passes up be-
tween the muscles to the cuticle. This well-known structure is
shown by Braun in Fasciola hepatica (’93, pl. 29, figs. 1, 2 and 3);
it is also found in Cotylaspis (Osborn, ’04, fig. 21) and in many
other forms.

In Clinostomum marginatum there is a decided departure from
the usual type which, since a similar structure has not been re-
ported for any other trematode so far as known, merits a detailed
description. Figs. 3, 4 and 5 show the relation of the wall to the
body asa whole. The wall seems to be distinctly marked off from
the central substance in these figures, due to the prominence of
the large oblique muscles.

The cuticle is as usual. It measures from 0.01 to 0.015 mm.
in thickness, is entirely structureless, is reinforced by spines which
ordinarily do not project beyond the surface. The spines are
acute and taper from a broader base seated on the deeper surface
of the cuticle. They are set close together. Twin spines of
smaller size sometimes occupy the position of one spine of ordi-
nary size, as if the amount of embryonic material apportioned to
one spine had been subdivided between two. Spines are found in
all parts of the general surface of the body, they are more numer-
ous on the ventral surface and on the posterior parts of the dorsal
surface. They are not found generally on the surface of the oral
field, with the exception of a small area immediately adjoining

196 HENRY LESLIE OSBORN

the mouth opening. The spines have a strong affinity for stains
and in the iron-haematoxylin preparations are deeply tinged by
it, while the cuticle remains unaffected. We know nothing of the
process by which spines are formed.

The principal peculiarity of the wall of Clinostomum is the
existence in its inner layer of an extensive system of cavities, an
extension of deeper cavities pervading the parenchyma every-
where, connected ultimately with the excretory collecting vessel.
A full deseription of these cavities will be given in connection
with the excretory system of which it forms a part. They are
conspicuous in longitudinal sections and can be seen in fig. 4 G
and figs. 6 and 7. The subcuticular cavities run in such a direc-
tion as to oncinele the body, with connections inward to the col-
lecting vessel as seen in transverse sections.

Organs in the cuticle, perhaps sensory. Certain cavities in the
cuticle (see fig. 8) are possibly parts of sensory organs. They
can be found in the areas in the oral field immediately around the
mouth and on the dorsal surface near the anterior end, but not
in the surface generally. In fig. 8 two of these are shown. They
are spherical cavities excavated in the substance of the cuticle by
which they are entirely enclosed except at the base of where they
are open to the parenchyma on which the cuticle rests. The
cavities thus have no communication whatever with the exterior.
A fine deeply stained fiber can be traced into these cavities from
the parenchyma. The indication from views like that of the
cavity on the right in the figure is that this thread expands into
a disk resting against the upper surface of the cavity. The best
interpretation of the function of these organs which we can make
on the basis of their structure is that they are the terminations
of a pressure sense apparatus, the fiber being regarded as a pro-
longation from a more deeply seated nerve cell.

We have very few references to such organs in the cuticle of
other trematodes. They are doubtless not uncommon but not
many forms have been examined for them. Nickerson (’95)
found a very similar organ in Stichocotyle, one of the Aspido-
bothridae. The organ shown in his fig. 15 differs only in size
from the one in fig. 8 of this article. Bettendorf also (97, fig.

STRUCTURE OF CLINOSTOMUM 197

30) found organs of much the same kind in the oral sucker of
certain distomes. Pratt (’09, fig. 10) copies a figure of a section
of the body wall of Ligula, one of the cestoda, from Zerneke. The
section made by the Golgi method shows nerve cells located some ~
distance below the cuticle from which threads run outward to
small spherical ‘sense organs’ located in the basal level of the cuti-
cle. These organs and those of Nickerson are similar in struc-
ture to those of Clnostomum. In Cotylaspis (Osborn, *04, fig.
33) the cuticle contains organs apparently of sensation but of a
different type from these. They are in the surface of the cuticle
and communicate with the exterior. They have a number of
stainable fibers which unite and pass as a single thread inward
through the cuticle and disappear in the parenchyma. Nicker-
son, in the article just referred to, in his fig. 14 has shown an organ
in the cuticle which communicates with the exterior.

Muscles of body wall. The usual muscles are present as in
trematodes at large. Figs. 6 and 7 show them in longitudinal
and transverse section respectively. In addition to them there is
a layer of longitudinal muscle, which les immediately below the
cuticle. This is an unusual layer of longitudinal muscle, the usual
one being located inside the circular muscle, while this is exter-
nal to it. We may designate it the outer longitudinal muscle
(mo) the other being then called inner (m7) in the figures. The
fibers of this outer longitudinal layer were seen by Looss (’85) and
are shown in his fig. 23. According to his figure they are very
much stronger than I find them in my sections. In my material
the fibers are exceedingly small, having a diameter of only 0.0009
mm. In fig. 7 they are shown under a magnification of 1100
diameters. Their size can perhaps be better appreciated by a
comparison with those of the inner longitudinal layer as seen in
figs.6 and 7. In the latter the fibers are cut transversely. These
fibers lie at equal distances apart, in a single layer, and in direct
contact with the cuticle.

Writers who have given attention to the finer structure of
trematodes (Braun, ’93; Otto, ’96; Stafford, ’96, to mention three
at random) agree that there are three layers which compose the
musculature of the body wall, viz: circular, longitudinal and

198 HENRY LESLIE OSBORN

oblique. I have recently made a re-examination of the sections
on which my paper of 1904 was based to test the possibility of
the coat being present in Cotylaspis; as a result I am entirely con-
vinced that there is no outer lougitudinal muscular layer. It thus
seems safe to conclude that Clinostomum is peculiar in the pos-
session of this layer, though a similar may perhaps be found later
in some other forms. My observations of the other coats also
confirm those already reported by Looss. The fibers of the cir-
cular coat lie in several layers (fig. 7); they are very small, though
larger than those of the outer layer, measuring 0.0012 mm. They
do not fall into groups or bundles like those of the inner longitu-
dinal layer. These fibers are seen in sections generally at various
levels between the sub-cuticular excretory cavities, which thus
seem to occupy an area produced by the expansion of that part
of the body wall, in correlation with the presence of these
cavities.

The inner longitudinal muscles lie much deeper than usual.
Instead of lying quite near the cuticle as they do in many cases,
and in close contact with the circular muscles, they are located
here, as shown in fig. 6, below the vessels of the excretory system
at a distance of 0.04 mm. from the outer muscles. The inner
muscles are thus seen to be pushed down to a considerable depth
below the bottom of the cuticle near which they usually lie. This
departure from the ordinary arrangement is clearly a structural
adaptation correlated with the presence of the sub-cuticular
vessels. We may further perhaps regard the presence of the outer
longitudinal muscles as a part of the same adaptation; they may
have been developed thus near the surface to offset the disad-
vantage due to the increased distance of the inner longitudinal
layer from the cuticle. ‘

The inner longitudinal muscle fibers are very distinetly grouped
into bundles alternating with intermediate areas from which they
are absent. Fig. 7, mi, shows one of these bundles in cross sec-
tion; it is made up of a cluster of fibers without other muscles in
close proximity. The fibers of the inner longitudinal muscle
differ in size, as can be seen in fig. 7; the largest ones are much
larger than those of the circular muscle, measuring 0.004 in diam-

STRUCTURE OF CLINOSTOMUM 199

eter. The oblique muscles running in the usual two directions
are more deeply located.

Certain interesting points were noted in the cytology of the
body wall muscles which will receive attention later in connection
with those of the parenchyma.

Wall structure in the oral field. 'The wall of the oral field pre-
sents a structure decidedly unlike that of the general surface of
the body. Figs. 3 and 5 show the wall under low magnification.
It is very much thinner, owing to the great reduction of all its
components. The cuticle becomes so thin as to be barely recog-
nizable. The spines, which are so general over the rest of the
surface of the animal, are entirely wanting on the oral field with
the exception of a small area immediately around the mouth
opening where spines of a much smaller size exist. Sub-cuticular
cavities so conspicuous elsewhere are scarcely recognizable. They
do not in any case take on the regular arrangement so usual else-
where in the body wall, but are merely irregular cavities under-
lying the surface and communicating internally with the vessels
of the excretory system. The musculature of the oral field does
not agree with that of the rest of the body. The various layers
are not continued from the wall into the field. Fibers can be
found lying parallel with the surface but they cannot be con-
nected with the fibers in the wall beyond. The longitudinal mus-
cles of the parenchyma (pml in fig. 3) run on anteriorly until
they meet the surface of the field to which they are then vertical.
They are shown in fig. 5 at ml running directly to the wall, their
position enabling them to act as retractors of the field as shown
in the figure.

Glands (?) in the body wall. There are certain nucleated cells
lying in the body wall, as shown in fig. 7 at gl, which seem to be
probably of a glandular nature. They are very long and slender,
consisting of a globular body, which lies on the level of the oblique
wall muscles, and a tapering portion which can be traced outward
to a termination on the inner surface of the cuticle. The outer
end of the cell may branch so as to present in sections two termi-
nations. No passage through the cuticle has been seen or any
indications of secretions passing from these cells to or through it.

200 HENRY LESLIE OSBORN

The globular body of the cell is entire on its inner side; that is,
there are no processes given off from it. The bodies of these
cells contain a large clear nucleus. There is no cell wall. The
cells stain readily with iron-haematoxylin. Their bodies which
lie on nearly the same level constitute a faint zone parallel with
the surface of the body.

The position and, to a certain extent the structure of these
cells, remind one of the cells found by Blochmann (’96) in ‘tre-
matodes and cestodes in a similar situation. I have not had
access to this paper of Blochmann, but several writers have re-
produced his figures, among them Pratt (09) who, in his recent
paper on the cuticula, copies a figure from Blochmann of the wall
of the cestode Ligula and designates ‘sub-cuticular cells’ certain
cells which show great resemblance to those of Clinostomum to
which I have just referred. There are some differences between
the cells in Pratt’s figure and those in fig. 7 of this article. In
Ligula the cell body is sharply angulated on its inner side and pro-
duced into fine threads, which are lost in the deeper parenchyma.
Externally also the cells soon taper to a very fine thread. In
spite of these differences however it seems quite reasonable to
regard these cells in their relations to the body structure as a’
whole as identical with the sub-cuticular cells of Clinostomum
just described. Benham (’01) gives a diagram of the structure
of the body wall of Ligula which has cells more like those seen in
Clinostomum. It is held by some writers that these are epithe-
lial cells which have sunken from a position originally on the sur-
face. The Clinostomum sections do not supply any evidence in
support of the view that these cells are epithelial in origin.

THE PARENCHYMA

The interspaces among the organs within the body are per-
meated by the usual network of branching fibers emanating from
large nucleated cells. In places where the parenchyma comes in
contact with the surface of the walls of various organs such as
the oesophagus (fig. 8) and the uterus, but not of all (not of the
intestine, for example) its fibers become much more numerous

STRUCTURE OF CLINOSTOMUM 201

and denser, so as to form a compact capsule for the support of the
parts beneath.

There are certain cells loosely clustered together in a mass
which lies in the parenchyma in the region directly anterior to the
ventral sucker. They are shown in figs. 1, 8,4 Band 5. These
cells are oval, and measure 0.03 by 0.015 mm. The nucleus, which
lies near the margin of the cell, is clear and round and contains one
or more nucleoli and a few minute grains of chromatin. These
cells lie among those of the parenchyma but differ from them in
appearance, having no processes and no connection with the fibers
of the parenchyma. Each cell is sharply bounded. They also
have no connection with the surface, no processes can be traced
from them going off toward the surface and their long axes le in
all directions. If they communicated with the surface the cell
bodies would point in that direction. There seems to be no doubt
that they are purely internal in their physiological action. They
are similarin cytological appearance in both bass and heron worms.
The cells contain a clear homogeneous material which has a
marked affinity for stains.

The physiological significance of this organ is entirely unknown.
Looss recognized these cells in the immature worms from the
fish cysts. He suggests (’85, p. 46 of separate) ‘“‘vielleicht sind
es die Anlagen von Driisen, die spaiter. . . . erste ihre
Funktion antreten werden.” But this suggestion cannot be
accepted since the cells are identical in structure in the mature
worms. MacCallum’s suggestion (’99, p. 700) that they are
parenchyma cells cannot be accepted, for they are not found out-
side of certain limits and parenchyma cells pervade all parts of the
body. The great number of these cells leads one to believe that
they are important. Their entire absence of connection with
other organs implies an independent function. It seems there-
fore most likely that they are concerned in some way in internal
secretion. |

Parenchymal muscles. The muscles of the parenchyma are
very well developed in Clinostomum. The usual two séts are
found, (fig. 3) namely, the longitudinal muscles and the dorsi-
ventral ones. There are no horizontal muscles. As the longi-

202 HENRY LESLIE OSBORN

tudinal muscles pass forward they ultimately meet the oral field
almost vertically to its surface and attach there so that they thus
become its retractors. Fig. 5 is a camera drawing from a speci-
men which died with the oral field retracted. In this section the
longitudinal parenchyma muscles can be traced forward directly
to the in-bent parts of the field. Observations of sagittal sec-
tions furnish evidence that, at least in many cases, a single muscle
reaches across from the dorsal to the ventral surface, for nucleated
myoblasts can be seen in connection with these muscles and thése
are grouped in the center of the body.

Some cytological features are well shown in the muscles of
Clinostomum. Both the inner longitudinal wall muscles and
the longitudinal parenchyma muscles frequently show transverse
subdivisions into stained and unstained zones such as has been
noted in other forms by various writers on trematode histology,
but in no case with which I am familiar are they shown so dis-
tinctly as here. Nickerson (’95) states that in Stichocotyle the
longitudinal muscles of the body wall appear to be tubular “with
nodes of deeply staining substance filling the lumen at intervals,”
and shows the appearance in fig. 16 of his paper. Stafford, too,
in Aspidogaster (’96, fig. 26) noted the presence of ‘transverse
lineations’ in the parenchyma muscles which he speaks of as
‘contraction centers.’ He does not note any in connection with
the body wall muscles. Bugge (’02) in his paper on the excretory
system in cestodes and trematodes incidentally mentions ‘Quer-
streifung der Muskelfasern’ which he observed in the circular and
longitudinal muscles of redias and cercarias, ‘‘wie wir sie bei
Arthropoden und andern Wirbellosen auffinden,”’ and also quotes
Cerfontaine, to whose article (in Bull. Acad. Sci. Belg., 27, no. 6)
I havenot had access, and Nickerson as having seen thesame thing.
In 1904 I saw and recorded (’94, figs. 11 and*12) a similar muscular
structure in Cotylaspis, a form related to Stichocoytle and Aspi-
dogaster.

Turning now to Clinostomum, figs. 9,10 and 11 are from immer-
sion objective camera drawings of longitudinal wall and paren-
chyma muscles. Figs. 9 and 11 are from the body wall and
parenchyma muscles respectively from the same series. Both

STRUCTURE OF CLINOSTOMUM 203

were drawn with the same objective but 11 is made with a higher
eyepiece. Such views are found generally in many different
series so that we are justified in regarding them as a normal fea-
ture of the cytological structure of Clinostomum muscle. In
fig. 11 it is clearly seen that the muscle is made up of several paral-
lel-sided filaments of considerable length, composed of a substance
which is not strongly influenced by haematoxylin and a second
deeply staining substance. The swollen globular appearance of
the latter leads us to believe that it is a peculiar ‘contractile sub-
stance.’ The fibers do not all present this appearance. One is
represented in fig. 10 in which also the myoblast and nucleus are
shown. The myoblast is large, measuring 0.014 mm. across, and
the nucleus has a diameter of 0.005 mm. Fibers can be traced
from such myoblasts. These appear differently from those in
figs. 9 and 11, showing a dark contour on the wall and a clearer
center. Incross sections the appearance is that of two substances,
a clearer central and a darker surface material. These seem to
be the ‘hollow muscles’ of writers. Fig. 10 shows the fibers,
probably in an uncontracted state, while 9 and 11 are contracted
fibers. A more adequate study of the cytology of this muscle is
beyond the scope of this article.

THE ALIMENTARY APPARATUS

Oral sucker. The mouth opening lies in the center of the oral
field and leads into the cavity of the oral sucker. The sucker is
nearly spherical and is very much smaller than the ventral sucker,
measuring 0.28 mm. long and 0.25 mm. across. It has the usual
cuticular lining and heavy muscular wall composed of fibers run-
ning in the various directions.

Oesophagus. The pharynx, which is generally present in tre-
matodes and usually follows close after the oral sucker, is entirely
wanting. ‘There is a short tube immediately behind the oral
sucker which, after running ventrally a short distance, makes a
dorsal bend to meet a transverse portion of the intestine. This
is the oesophagus. -The structure of the two bends is somewhat
different. The more anterior portion is very thin walled and is
lined with a thin cuticle continuous with that of the oral sucker.

204. HENRY LESLIE OSBORN

The posterior chamber is a globular dense body as seen in a whole-
mounted worm. The wall is thick and heavy, due not to the
presence of a heavy muscular coat as it would be if the organ were
a pharynx, but to the very peculiar structure of the cuticular coat.
The cuticle here, which is continuous with the thin layer of the
anterior chamber, suddenly changes its character and becomes a
mass of tall slender processes springing vertically from the wall
and projecting freely into the lumen of the organ. Their appear-
ance is shown in fig. 12. They bear some resemblance to the tall
processes of the epithelium of the intestine just above them, with
which they are directly continuous. They have every appear-
ance of having arisen from a cuticularized epithelium. In Coty-
laspis (Osborn, ’04) there are similar indications of an epithelial
origin of the cuticle which lines the oesophagus.

The posterior chamber of the oesophagus has almost no muscu-
lar tissue in its wall. A very few circular and longitudinal fibers
can be recognized, evidently strictly comparable with the muscles
of the intestine. There is however a coating on the inner surface
of the organ which is a condensation of the parenchyma at large.
This has a fine but definite boundary next the parenchyma.

Oesophageal glands. Numerous cells lie in close proximity to
the oesophagus (oegl in fig. 12) which are not ordinary parenchyma
cells. Their massing too goes to show that they constitute a
definite organ whose position requires us to regard it in its work
as in some way a part of the oesophagus. The cells are not angu-
lar like those of the parenchyma but have rounded outlines. Each
cell has a nucleus poor in chromatin and a distinct nucleolus.
The sharp line in the figure passing on the left side of this group
of cells marks the boundary of the denser parenchyma which
ensheathes the oesophagus. It will be seen that the cells are
located outside of this sheath and so are somewhat remote from the
lumen of the oesophagus. An organ of this sort is usual in trema-
todes; it is often called a ‘salivary gland.’ One writer however,
(Otto, ’96) questions the salivary function of the cells in the Am-
phistomes, since he-does not find any connection between them and
the lumen of the oesophagus. We shall however go on calling
these organs ‘oesophageal glands’ though we have no definite

STRUCTURE OF CLINOSTOMUM 205

information as to their physiological significance. It is possible
that they are merely mucin-forming organs and that they acquire
a temporary connection with the oesophagus.

The intestine. The intestine consists of a part crossing the body
transversely (fig. 4 A) which, after bending, continues as the two
long lateral caeca. The caeca lie in the center of each half of
the body and extend (fig. 2) to the level of the excretory bladder.
The walls of the caeca are not entire; blind pouches extend out-
ward from them. These pouches are not as large and distinct in
the material after fixation as they are in life. Fig. 13 a is a free-
hand drawing of the living organ in a specimen just liberated from
a bass cyst. The pouches arise on both sides of the intestine;
they are very numerous and close together and are not long and
slender. The form of these pouches distinguishes C. marginatum
from C. heterostomum. In the latter (Braun, ’00, fig. 1) there
are a few very long and slender diverticula which are confined to
the outer side of the intestine. In the presence of these intestinal
pouches Clinostomum resembles Fasciola hepatica and the pla-
narians. Fig. 15 is a camera drawing of the wall of the intestine.
The pouches are conspicuous in some places and absent in others.
This corresponds with the facts seen in life; in bass specimens the
pouches are contractile and at moments are drawn back into the
wall. The wall itself is contractile; in life its movements are very
conspicuous. The lumen is filled with a fine grained material
lemon yellow in color. This flows back and forward, streaming,
the pouches empty themselves of it or fill: with it and the contrac-
tions of the wall may obliterate the intestine entirely for a
moment.

The structure of the intestine wall from a fully matured heron
worm is shown in fig. 13. The epithelium presents two distinct
zones, a deeper basal one and, arising from it, a second zone of
separate columnar structures. The basal zone is a continuous
protoplasmic layer, in which distinct nuclei occur at somewhat
regular intervals, but without any walls dividing it into cells.
This basal syneytium takes the stain readily. The processes of
the outer zone show a relation to the nuclei below though they
are not always strictly over them. In a section, (fig. 13 on the

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

206 HENRY LESLIE OSBORN

left) this may be due to slight differences in level. It is planned
to treat certain cytological points connected with this epithelium
in a later paper so that for the present I will only state that these
processes are apparently amoeboid and capable of being projected
from the deeper body of the cell or retracted. They are clear and
barely stained. They are filled with minute black pigment
grains which have been traced to the decomposed blood corpuscles
of the heron on which the worm had fed. While the intestinal
epithelium in some instances shows the appearances just described
there are other cases in which its form is quite different as shown
in fig. 14. Here it is a low, level surface consisting of a layer of
protoplasm with imbedded nuclei. There are no division walls
and the layer has the appearance of a syncytium. Cells of the
type shown in fig. 15 are found in the intestine of worms from the
bass. I have interpreted them as being in a resting or non-diges-
tive state while those in fig. 13 are actively engaged in the work of
digestion. In many trematode sections and figures with which I
am familiar the cells of the epithelium are entirely distinct and
independent to their very base. They do not show any fusion
as if syncytial as is the case here. In connection with the struc-.
ture of Cotylaspis (04, fig. 19) I called attention to the entire
independence of the cells of the intestinal epithelium. " In Crypto-
gonimus, on the other hand, the cells of the epithelium of the intes-
tinal caeca are fused into a syncytium.

The circular muscular coat is very scanty, its fibers lie close to
the epithelium. The longitudinal coat is, also very feeble. Its
fibers lie at a distance in the parenchyma.

The cavity of the intestine of the worms obtained from bass
cysts is filled with thin, flat, four-sided crystalline bodies. As
soon as the worm has escaped from its cyst the strong peristaltic
contractions already mentioned force this substance backward
and forward. At frequent intervals portions of it are expelled
from the body through the mouth. In worms obtained from the
heron this material is not found in the cavities of the intestine.
In such worms on the contrary the intestine has been found to
contain a coagulated fluid substance with blood corpuscles scat-
tered through it, which upon careful examination were found to be

STRUCTURE OF CLINOSTOMUM 207

identical with blood corpuscles taken from the heron. This
substance is evidently food, but the content of the caeca of the
bass worm cannot be so considered. Its prompt rejection from
the body as soon as it is liberated from the cyst would be evidence
sufficient to justify this conclusion. Its erystalline form and the
fact that it is discharged as soon as the animal becomes free, point
to the hypothesis that the cavities of the intestine are made use
of for storage during encystment and that the substance therein
is a waste product.

THE EXCRETORY SYSTEM

The excretory system of C. marginatum has never been de-
scribed. The indications of it which are given in Looss (’85, fig.
22) are purely diagrammatic and somewhat misleading. The
system, moreover, presents some features which are very unusual,
so that the whole subject needs a careful revision.

The location of the excretory pore has already been noted. It
opens from a very short duct (fig. 15) which is in communication
with the v-shaped bladder. Internally the two branches of the
bladder receive the termination of the collecting tube in the center
of a flattened area. At the excretory pore there is an invagina-
tion of the cuticle which covers the outer surface of the body.
As this passes more deeply it gradually changes into a cubical
epithelium composed of nucleated cells identical in structure with
those which make the wall of the collecting vessel. At inter-
mediate points epithelial cells of the bladder show all stages of
degeneration in structure and pass insensibly into cuticle. There
is no muscular tissue in the walls of the bladder. Living speci-
mens were observed particularly with reference to contractions
in the bladder as I had found this organ in Cotylaspis interesting
in this respect (794, p. 216) but the walls were not contractile. In
correlation with this is the absence of a sphincter at the surface
pore (one is present in this place in Cotylaspis) and the presence
of a sphincter at the junction of the collecting vessel and the blad-
der. We may conclude from the position of the sphincter and the
non-contractility of the bladder that the latter in Clinostomum is

208 HENRY LESLIE OSBORN

merely a passage and not a place of storage and that the collect-
ing vessel is the functional bladder.

The collecting vessels are very large in the posterior region of
the body, but anteriorly their identity is lost. It is usual in trema-
todes for a collecting vessel to run from the terminal bladders for-
ward to a point near the anterior end of the body and then to
bend suddenly on itself and run backward again. The second
vessel, called the recurrent vessel, is supplied with a strong vibratile
apparatus, while the collecting vessel lacks these. In Clinosto-
mum. the collecting vessel is readily traced forward as far as the
ventral sucker. In fig. 2 it is shown on the right side omitted on
the left, in fig. 3 it is shown as far forward as the genital organs
and then omitted, in fig. 4 it is shown in sections B-G. It disap-
pears a short distance in front of section number 150. The level
of the vessel is seen from the cross-sections. It always lies exter-
nally to the caeca and generally slightly ventrally. Its diameter
is quite variable as in fig. 15, a camera drawing of a section pass-
ing in its plane for a long distance. The wall is epithelial and
muscular.

I am not able to give a definite account of the relation between
the collecting vessel and the recurrent vessel. I have devoted
much time to the study of this in different ways without being able
to follow the collecting vessel forward to where it meets the recur-
rent vessel. The body is too thick anteriorly to allow this point
to be seen in an entire compressed specimen. I have repeatedly
examined the youngest individuals I could find but without suc-
cess. The network of anastomosing vessels (described in a mo-
ment) are so complicated in the anterior of the body that it is
impossible to recognize the collecting vessel if, indeed, it has re-
mained distinct from them. It is, of course, possible that the
collecting vessel does not remain distinct but is lost in the network
of vessels.

Allusion was made above to the system of cavities which lie in
the body wall immediately under the cuticle. In living worms
just removed from bass cysts these cavities are filled with a
cream-colored fluid composed of minute highly refractive drop-
ets which has the effect of an injection, making it very easy

STRUCTURE OF CLINOSTOMUM 209

to distinguish the different vessels and their connections. In such
a preparation the collecting vessel as well as the smaller vessels
which are derived from it are readily seen. The appearance of
this system of vessels is shown in a free hand drawing (text fig. 1,
for which I am indebted to Mr. Faus Silvermale) made from
an unusually young live bass worm under slight compression.
The collecting vessel can be followed, its size diminishing as it
advances until it is lost in the network of its subdivisions. The
network shows a predominance of transverse vessels, though

iS mm
Fig. 1 Free hand drawing from a young, living, slightly compressed bass worm.
Zeiss 2 A.

with many communicating vessels connecting them, and some
which cross over to the other side of the body and become con-
tinuous with those of that side. In this preparation the recur-
rent vessel could not be seen, but as I shall point out later it
seems most likely that it is present and joins the collecting vessel
in this part of the animal.

In the posterior part of the body these vessels may be somewhat
more definitely subdivided into two sorts according to their des-
tination, a superficial system which runs to the surface and be-
comes the subcuticular vessels and a deep system which passes

210 HENRY LESLIE OSBORN

inwardly among the inner organs. ‘The origins of both kinds
can be seen in fig. 15; they arise at close intervals from the col-
lecting vessel on both sides. Some of the vessels of the super-
ficial system run directly to the surface (one of these is shown in
fig. 4 G) where they become circular vessels immediately under
the cuticle. Since these tubes encircle the body they are readily
seen in sections passing tangentially in the plane of the surface.
The tubes frequently anastomose and all communicate directly
with the collecting vessel. They are cut across in longitudinal
sections and produce the appearance seen in figs. 3 and 5.

Looss (’85, p. 49) expressed a suspicion that these subcuticular
vessels communicate directly with the exterior: ‘‘so halte ich
es doch fiir héchst wahrscheinlich, dass—der Excretionsporus
nicht die einzige Stelle ist von welcher dieses Maschenwerk von
Kanilen mit der Aussenwelt in Verbindung tritt und zwar sind
es die subcuticularen Maschen des Gefiissnetzes welche diese
Kommunikationen vermitteln.”” And later he says ‘‘Ich halte
es nun nicht fiir unméglich dass sie (i.e., subcuticular vessels)
auch nach aussen miinden,”’ ete. But he closes his account with
an admission to the effect that he has not succeeded in demon-
strating the presence of openings from them to the exterior. The
observations of the movements of the fluids in these passages
described above render it very certain that outlets from them
directly to the exterior do not exist; were such outlets present we
should undoubtedly have seen stuff from within issuing through
them.

Turning now from these superficial vessels to the deep ones we
find that they pass inward, permeating the parenchyma every-
where. The vessels of one side tend to remain entirely confined
to that side, they anastomose with one another but do not often
become continuous with those of the opposite half of the body.

Allusion has been made to the flow of the contents of the ex-
cretory vessels in life. Pulsations were seen in bass specimens
in the wall of the collecting vessel, forcing a stream out into the
dependent vessels. Later these streams reversed their direction
and the droplets course back again into the larger vessel. As

already noted there is no escape distally; the movement is an
ebb and flow.

STRUCTURE OF CLINOSTOMUM 211

An observation made recently upon a worm from a frog cyst
seemed to be inconsistent with a circulatory movement of the
droplets as just described. The cyst attracted my attention
because of its small size, it being globular, compact and only 1.3
mm. in diameter, quite unlike frog cysts and much like those
found in the bass. When the worm had been liberated and ar-
ranged in a compressor for observation it was seen that the vessels
of the anterior region were filled with highly refractive droplets.
These droplets were not in a state of flux but were stationary.
The pulsations mentioned above and the flow of droplets were
seen in bass specimens in the posterior region. In the frog worm
the vessels of the posterior region were empty. It is possible that
this worm was a very recent arrival in the frog and that the move-
ments of the droplets had not yet begun to take place.

A system of branches derived from the collecting vessel and
permeating the body in this way is very unusual in trematodes.
The usual structure is a network of minor vessels uniting to form
larger vessels which finally merge into a single collecting vessel.
In three widely separated forms however we find an arrangement
somewhat similar to that of Clinostomum. In a young stage of
D. echinatum Looss (’04, fig. 192) figures a collecting vessel much
resembling that of Clinostomum, especially in the anterior body
region where side branches are given off, the main vessel mean-
while continuing until it meets the ciliated recurrent vessel at the
extreme anterior end of the body. A comparison of Looss’ fig. 192
with that of the younger stage shown in 191 indicates that the
branching is a late feature in the life history, a fact of interest since
it is uncommon in trematodes at large. In adults of D.echinatum
(Looss, ’94, fig. 114) these vessels are very much branched but the
branches do not assume the form of a sub-cutaneous system like
the one so well developed in Clinostomum. In Cephalogonimus
also the excretory collecting vessel is branched. This point was
first noticed by Poirier (’85, fig. copied by Braun, ’93, pl. 20, fig. 9)
who says “‘Ces canaux lateraux comme le canal impar median,
emittant sur tout leur parcours des branches ramifies se dirigeant
vers le bords lateraux du corps. Ces ramifications s’entendent en
avant, jusque un peu au-dela du point de bifurcation de l’oesoph-

Da bes HENRY LESLIE OSBORN

age.’’ In his illustration, as well as in this description, there is no
recognition of subcuticular vessels like those of Clinostomum. In
a study of the parasites infesting the frogs of Minnesota I have
happened to find specimens of this genus. The study of living
and sectioned material of this form demonstrates that, while there
is an extensive system of vessels derived by branching from the
collecting vessels and one which bears considerable resemblance
to that found in Chnostomum, encircling subcuticular vessels are
not developed.

A third form with somewhat similar branching excretory ves-
sels was encountered at Chautauqua, New York. In the livers
of sun-fishes certain cysts were found which contained immature
flukes belonging to the holostomes. These forms are peculiar in
having a broad thin anterior body region bearing a resemblance
to the foot of a gasteropod mollusc and posteriorly a globular mass
carried vertically over it. The excretory pore is located at the
summit of the latter. In the thin anterior part there are a median
and two lateral longitudinal vessels, extending from which are
branching vessels extending everywhere in the foot, anastomosing
and forming a complete network. All of the vessels of this sys-
tem contained minute highly refractive droplets, similar in appear-
ance to those found in the excretory cavities of Clinostoma
which had been recently liberated from bass cysts. In the living
worms masses composed of these droplets were discharged from
time to time from a point located at the posterior end of the body,
the excretory pore, thus indicating that the passages are members
of the excretory system. In life the droplets were in constant
motion in the vessels, coursing rapidly in all directions as they had
been seen doing in Clinostomum.

This observation, taken in connection with they presence of such
droplets in. the encysted specimens of Clinostomum and their
absence in the heron specimens of Clinostomum and the free
living Cephalogonimus, constitutes an argument in favor of the
supposition that the droplets are composed of chemical wastes.
In an encysted organism these must be disposed of in a way that
will prevent their damaging the animal, accordingly they cannot
be discharged from the body in the ordinary manner but must be

STRUCTURE OF CLINOSTOMUM Dita

stored during the period of encystment. It is thus reasonable to
look upon the extensive equipment of spaces possessed by the
excretory collecting vessel as a storage apparatus. In favor of
this interpretation is the further fact that in both Clinostomum
and the holostome just mentioned the contents of these cavities
begin to be discharged as soon as the worm has escaped from its
eyst. I think that the substances contained in the intestinal
caeca may also prove to be waste matters and that these cavities
are also being employed for storage.

The recurrent excretory vessel. Reference has already been
made to a vessel which parallels the collecting vessel. It is
readily seen in the parts of the body behind the ventral sucker;
anteriorly it is lost in the maze of vessels which are derived from
the collecting vessel. Posteriorly (fig. 16) it bends sharply for-
ward, as the vessel into which the capillaries drain. This, which
I have called the recurrent vessel, is spirally coiled in all sections
and even shows this state in living animals. It is located exter-
nally and somewhat dorsally to the collecting vessel, but is much
smaller, having a diameter of only 0.02 mm. The wall is com-
posed of a very thin membrane. The tube is uniform in diameter
in all parts; unlike the collecting vessel the wall possesses no con-
tractility, there being no muscular tissue present. The wall of
this vessel is supplied at close intervals with peculiar ciliary organs.
In life these vibrate at a very rapid rate so that they become visible
only after their vitality has become lowered. Then it is seen
that the organ is attached posteriorly in the wall of the vessel,
the rest being free and pointing anteriorly so that its vibration
produces a current running forward in the tube. These organs
are located in the recurrent vessel at close intervals. Bugge
(02, fig. 62) finds that in certain cercarias occurring in certain
helices the chief canals are supplied with ‘Wimperschopfen’
which correspond with the organs just mentioned and in addition
that there is a lining of ordinary cilia clothing the rest of the inner
surface of the wall. There are no similar ordinary cilia in these
vessels of Clinostomum. The ciliary organs are many times longer
than the diameter of the vessel in whose lumen they lie. In
life I am unable to recognize individual cilia in them but in sec-

214 HENRY LESLIE OSBORN

tions after the application of iron-haematoxylin, cilia are clearly
seen as sharp black wiry looking lines.

In view of the fact that these ciliary organs produce strong
current which flows forward, we are compelled to suppose that
the recurrent vessel discharges directly into the collecting vessel,
although as already noted it has not been possible to recognize
the connection.

Flame-cells and. capillaries. The ultimate members of the ex-
cretory apparatus of Clinostomum are very imperfectly known
as yet. Much attention and time have been dedicated to the
effort to trace these parts in the living animal with very inade-
quate reward. Some glimpses of them have been obtained how-
ever both in life and in sectioned material. Flame-cells have been
seen; they are very tall and slender with a narrow base where the
elongate and narrow mass of cilia are attached. A detailed
account, with illustration of these flame-cells together with some
other finer points, must be reserved for a later article.

It has not been possible to determine the mode of arrangement
of the capillaries and connecting vessels. In some places the
capillaries have been recognized. They ran in a posterior direc-
tion. Vibrating ciliary organs could be seen within them. It .
was not possible in any case to trace these vessels to a point of
connection with the recurrent vessel and I feel very strongly
convinced that the recurrent vessel does not receive any branches.

EXCRETORY SYSTEM IN TREMATODES IN GENERAL

There is considerable difference in the plan of anatomical organ-
ization of the excretory systems of different trematodes. In all
there is a system of flame-cells and their capillaries and one, or
occasionally two (e.g., Aspidogaster), posteriorly located excre-
tory pores. But there is great difference as to the vessels lying
between the external pore and the capillaries. All degrees of
distance between the terminal bladder and the capillaries can
be found. In Opisthoglyphe endolobum (Looss, ’94, fig. 157) a
large forked chamber, confined to the posterior third of the body,
receives directly a vessel formed by the junction of the capil-

STRUCTURE OF CLINOSTOMUM Zits

laries. In Allocreadium isoporum (Looss, 794, fig. 15) a collect-
ing vessel can be recognized which reaches the first body third
and there receives the capillary vessels. In Gorgodera cygnoides
(Looss, ’94, fig. 125) a collecting vessel runs the whole length of
the body and at its anterior end meets a vessel which runs back-
ward a short distance before the connecting vessel from the cap-
illaries meets it. This might be considered as a short recurrent
vessel. In Distomum echinatum (Looss, 794, fig. 191) a fully
developed collecting vessel meets a still longer recurrent vessel.
In Harmostomum leptostomum (Looss, 794, fig. 118) the collect-
ing vessel is fully developed and the recurrent vessel runs nearly
to the posterior end before the two vessels enter it from the capil-
laries. Finally in Cotylaspis (Osborn, ’04, fig. 26) the recurrent
vessel as well as the collecting vessel, is fully developed, the cap-
illaries discharging into a canal which is a forward bend of the
recurrent vessel. We see from this brief survey of these different
forms that, with the gradual development of both collecting and
recurrent vessels, an increasing interval is interposed between
the capillaries and the exterior. The structure of these two
vessels is entirely different, one being entirely destitute of cil-
iary apparatus and furnished with muscular tissue, the other being
ciliated and devoid of muscle. Clinostomum has its place among
the forms with complete collecting and recurrent vessels, and in
addition possesses the remarkable system of branches derived from
the collecting vessel.

It would be possible to find a series of forms showing reciprocal
developments of excretory collecting vessel and bladder. Thus
in Stichocotyle there is no bladder and the collecting vessels are
very large; in the closely related Cotylaspis the collecting vessels
are narrow tubes and there are two well developed excretory
bladders which are rhythmically pulsatile.

While noting that the parts of the excretory system thus exhibit
a series of degrees in the development of complexity of organiza-
tion we must not forget that this series is found not in a group of
genetically related animals but among forms which are widely
separated in the system. On the other hand when we examine
forms which are closely related we find great differences. Thus

216 HENRY LESLIE OSBORN

in the three genera of the Aspidobothridae we find in Stichocotyle
(Nickerson, 795, fig. 23) no recurrent vessel, a very voluminous
collecting vessel and no bladder; in Aspidogaster (Stafford, ’96,
fig. 15) a partly developed recurrent vessel, moderate collecting
vessel, no bladder and two excretory pores; in Cotylaspis (Osborn,
94, figs. 5 and 26) a fully developed recurrent vessel, a small
collecting vessel and two well developed bladders. So that it is
impossible to attach any phylogenetic value to differences shown
in the organization of the excretory system.

THE REPRODUCTIVE SYSTEM

My observations are in substantial accord with those of Looss
and MacCallum with regard to the chief facts in the anatomy of
this system. Figs. 1, 2 and 3 show the parts in situ as they ap-
pear in various sectional planes. Fig. 17 is an outline of the
organs based largely on a study of the total preparation from
which fig. 1 was drawn. The two testes lie in the last body third;
ovary, shell-gland and ootype, oviduct, Laurer’s canal and yolk
receptacle are all compactly grouped in the space between the
testes. There is a peculiar uterine sack in the course of the uterus,
the yolk follicles are small and very diffuse. Since no detailed
account of the members of this system has ever been given I will
give here a brief description of it.

Genital pore. The genital pore lies in the mid-ventral line (fig.
4 EF). Its distance from the anterior end has been considered
already in this article. There is an atrial cavity below the surface
from which a single pore opens to the exterior. The relation of
these parts is shown in figs. 3 and 4 EL.

Cirrus sack. MacCallum’s figure shows the cirrus sack and its
contents very adequately. The sack occupies a position on the
right side of the animal in front of the anterior testis. At its
posterior border it receives the two vasa deferentia. The wall of
the sack is supplied with muscles whose powerful fibers lie so as
to form a circular and longitudinal coat. The sack contains a
tube of varied diameter, coiled so as to accommodate its length to
that of the sack. Posteriorly the tube expands to a larger, thin-
walled, non-muscular seminal vesicle filled with spermatozoa.

STRUCTURE OF CLINOSTOMUM AG:

Continuing it anteriorly is a smaller portion whose wall is sup-
plied with a very strong coat of circular muscles. It is followed
by a less muscular portion at the outer end of the sack. This
portion is surrounded by what are apparently to be regarded as
prostate cells. This part is not as strongly muscular as the middle
region of the tube. Following the usual nomenclature I have
designated the middle and outer parts respectively as prostate
portion and ductus ejaculatorius. It seems however that the
more glandular part and the more muscular parts are out of the
usual order. Thus in D. isoporum (Looss, 794, fig. 104), the outer
part is strongly muscular, and coiled and between it and the semi-
nal vesicle there is a small chamber with which the large prostate
cells communicate.

Testes. The testes are somewhat pyramidal in shape, their
bases slightly coneaved and facing each other. In some cases the
remaining surfaces are more or less deeply indented, in many
others they are entire. A sharp line bounds the testes. The
vasa deferentia have a wall of epithelium with flattened nuclei.
This epithelium can be recognized in the wall of the testis where it
connects with the vas deferens but no epithelium can be seen in
the wall of the organ elsewhere. Apparently trematodes differ
on this point. In Cotylaspis (Osborn, ’94, fig. 88) the wall of the
organ is. distinctly epithelial. Schwartz (’85) found nuclei in the
wall of the testis of certain early stages of trematodes while Ziegler
(83) claimed that the wall is non-cellular in Bucephalus and
Gasterostomum. |

The testes are filled with cells which, in some bass worms, are
almost completely filled by the very large nucleus, poor in chro-
matin, and with a very large, readily staining nucleolus which
indicate the inactive stage preceding spermatogenesis; in other
bass specimens with cells showing various phases of spermato-
genesis. In the heron worms the testes contain fully developed
spermatozoa scattered among the active cells.

Ovary. The size and form of the ovary as shown in fig. 26 of
Looss’ paper (’85) is unlike anything which I have found in my
material, in which it is oval, entire and measures 0.4 mm. by 0.2
mm. In MacCallum’s figure it is also small and entire. It is

218 HENRY LESLIE OSBORN

bounded by a thin non-cellular membrane, which encloses cells,
some of which, near the opening to the oviduct, are much larger
than the rest and are about to descend to the ootype.

Oviduct and uterus. The oviduct passes inwardly from the
ovary and crosses to the opposite side of the body. Its wall is
composed of cubical epithelium and circular muscle fibers. At a
point near the ovary there is a small sack, the spermatic receptacle,
opening from the oviduct, this narrows dorsally to a tube—
Laurer’s canal—runs to the dorsal surface of the body (fig. 4 G),
and opens to the exterior. The epithelium of the oviducal wall is
replaced by cuticle in Laurer’s canal, which beéomes continuous
with that of the general surface of the body. The oviduct, at a
point a little farther to the left, meets the duct coming from the
yolk receptacle. There is no marked change in the diameter of
the oviduct at this point but it is surrounded by glandular cells
and doubtless serves as the ootype. The duct from this point
continues as the uterus, at first without change in diameter or
direction, next with several loops it recrosses toward the ovary,
then abruptly bends again and runs a straight course, passes
externally to the anterior testis on its left side and runs forward
to enter a large sack which we may call the uterine sack. The
relation of the uterus to this sack is shown in fig. 3; it passes on its
dorsal wall for a distance and opens into it at about the center of
its dorsal surface.

Uterine sack. The uterine sack is a large cavity capable of
considerable distension; in the case of mature worms it is filled
with eggs, as in fig. 1; in bass worms (fig. 2) the cavity is merely
a narrow slit. The form of the cavity in transverse section is
shown in fig. 4 D. It extends posteriorly to a point near the an-
terior border of the anterior testis; anteriorly it does not reach the
ventral sucker. The outlet from the sack is located at its pos-
terior end. The histological structure of the wall of the sack is
quite unlike that of the uterus. In the latter there is a nucleated
epithelium and a coat of muscle fibers. In the sack the cavity
is lined with cuticle and there is a muscular coat consisting of cir-
cular and longitudinal fibers. In addition to these there is a
condensation of the parenchyma immediately surrounding the
uterus. The nuclei of these parenchyma cells lie in definite

STRUCTURE OF CLINOSTOMUM 219

lines parallel with the surface from which the fibers of the paren-
chyma radiate. A sharp line bounds this mass of specialized
parenchyma which thus constitutes a capsule enclosing the uter-
ine sack. The fact that the uterus enters the sack in the center
of its dorsal surface and not at the anterior end prevents us from
regarding the sack as merely a dilatation of the uterus. We must
however think of it as having arisen as a differentiation which has
taken place in a loop of the uterus. In most of the species of this
genus (Braun, ’00) there is a similar blind sack into which the
uterus enters and which extends blindly in front of the end of the
uterus. In one species however, (C. heterostomum, Braun, ’00,
fig. 1), the uterus passes forward to the posterior border of the
ventral sucker where it bends and runs straight back again to end
at the female genital opening. This is doubtless the more prim-
itive anatomical arrangement, and the one from which the sack
form has been developed. We note also in passing that this
species is more primitive, too, in possessing well developed diver-
ticula of the intestinal caeca.

The form of the uterine sack in the D. reticulatum of Looss,
as described and figured in his paper (’85), is decidedly different
from that which I have just described. The sack in that species
is elongated posteriorly to reach a point posterior to the posterior
testis (fig. 22). In fig. 26 we learn that the part of the sack which
leads to the exterior is a lateral offset from the main sack. This
posterior portion of the sack of Looss is wholly wanting in my
material. Cross sections (e.g., fig. 4 F) show that the sack does
not extend into the testis region of the body. This is an inter-
esting point. It does not seem possible to doubt the fact as
related by Looss for his form. In every other respect the form
D. reticulatum bears the closest resemblance in organization to
C. marginatum, and writers from Leuckart down have considered
them identical. Thus Leuckart (’89, p. 401) says D. reticula-
tum‘ Mit Leidy’s Clinostomum gracile zusammenfiallt.”’ Stiles
and Hassall (98) say ‘‘Looss described as Distomum reticulatum
a form which is evidently identical with Leidy’s Clinostomum as
Leuckart has already surmised,” etc. And MacCallum (99, p.
705). The description given by Looss of D. reticulatum

applies so exactly in every particular to the forms we

220 HENRY LESLIE OSBORN

have just considered [C. marginatum] that I have not the least
hesitation in concluding that they are the same.” The form of
the uterine sack however isvery different in D. reticulum from that
of C. marginatum or of any other member of the genus. Accord-
ing to fig. 26 of Looss’ article (’85) the sack is extended posteriorly
dorsally to the testes until it reaches a point posterior to the pos-
terior testis. This posterior development of the sack is a feature
not found in any of the species of this genus so far as I am aware.
In the several species included by Braun (’00) in his article on the
group, the sack ends in advance of the genital pore. If Looss was
not in error in regard to the form of the sack (and this seems very
improbable) then we must recognize that D. reticulatum differs
decidedly in this respect from the rest of the genus. However
in any case we should not attach very much importance to differ-
ences in the shape of an organ like the uterus or its parts.

Metraterm. There is a short slender tube running from the
sack to the atrium (figs. 3 and 4 #) which, following the nomen-
clature suggested by Ward (’94), may be called the metraterm.

Vitellaria. The vitellaria are shown in fig. 1. They are dif-
fusely scattered in all parts of the region behind the ventral sucker.
As shown by transverse sections they lie in a thin zone, concen-
tric with the surface and next to the outer wall of the body. They
are entirely absent from the anterior part of the body. The vitel-
laria are made up of ultimate follicles, all of them very small and
numerous measuring 0.07 mm. These are bounded by a distinct
membrane which encloses a few yolk cells which measure 0.0125
mm. in diameter. Usually the vitellaria cannot be seen in total
preparations made from bass worms, but sections of similarimma-
ture individuals show the follicles with their thin wall enclosed
cells whose structure is then identical in appearance with that of
the immature germinal cells of the testes and ovary. In sections
from mature worms the follicles contain similar in:mature cells
and also fully formed yolk cells with large nucleus and nucleolus
and a cytoplasm containing granules, some of them measuring
0.001 mm. in diameter. These are food granules; in producing
them the follicle cells differ from ovarian cells with which they
are very likely homologous.

STRUCTURE OF CLINOSTOMUM pea):

The yolk receptacle lies near the ovary and its duct reaches the
oviduct as already noted. The shell gland surrounds the ootype
at this point. Its cells radiate from the ootype and their long
tapering portions seem to communicate with its cavity, though it
is not possible to recognize absolutely the manner of connection.

The egg. The eggs measure 0.099 mm. by 0.66 mm. There
is a distinct operculum very near the end of the shell. The shell
in some cases is deeply stained by the haematoxylin and looks as
if composed of the same substance as the spines. In other cases
the shell is not influenced by the stain. The shells contain as
usual a fertilized cell derived from the ovary and several cells
derived from the vitellaria. The eggs, both those of the uterus
and the older ones of the uterine sack are practically undeveloped.
In some cases the true egg cells may undergo one or more of the
early stages of cleavage but in the vast majority of eggs no devel-
opment takes place during the time that they are lodged within
the body of the parent. In view of this we must recognize the
sack as merely a place for the storage of eggs. The reason for this
storage remains for the present unknown.

BIBLIOGRAPHY
BrenuaM, W. B. 1901 A treatise on zoology, edited by E. R. Lankester. Vol.
4. The Platyhelmia.

BETTENDORFER, H. 1897 Ueber Musculature und Sinneszellen der Trematoden.
Zool. Jhrb., Abt. Anat., Bd. 10, pp. 307-385.

BiocuMann, F. 1896 Die Epithelfrage bei Cestoden und Trematoden. Ham-
burg.

Braun, M. 1893 Bronn, Klassen u. Ordnungen; Platyhelminthes I, Trematoden.
Bd. 4, pp. 396-924.

1900 Die Arten der Gattung Clinostomum Leidy, Zool. Jhrb., Abt.
f. Syst., Bd. 19.

Buacr, G. 1902 Zur Kentniss des Excretionsgefiisssystems der Cestoden und
Trematoden. Zool. Jhrb., Anat., Bd. 16.

Leuckart, R. 1889 Die Parasiten der Menschen, u.s.w.

Linton, E. R. 1898 Notes on the trematode parasites of fishes. Proc. U.S.
Nat. Mus., vol. 10 pp. 507-548.

1910 The diagnosis of a case of parasitism in the brook trout. Proce.
Seventh Internat. Zool. Congress, Boston Meeting, 1907.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

pape HENRY LESLIE OSBORN

Looss, A. 1885 Beitrige zur Kenntniss der Trematoden, Distomum palliatum
n.s.und D.reticulatumn.s. Zeit. f. w. Zool., Bd. 41.

1894 Die Distomen unserer Fische und Frosche. Bibliotheca Zoolo-
gica. Abt. 16.

1899 Weitere Beitrage z. Kentniss der Trematoden Fauna Aegyptens.
Zool. Jhrb., Abt. f. Syst., Bd 12. pp. 521-784.

MacCauium, W. G. 1899 On the species Clinostomum heterostomum. Jour.
Morph., vol. 15, pp. 697-710.

NickEerson, W. 8. 1895 On Stichocotyle nephropis, a parasite of the American
lobster. Zool. Jhrb. Abth. Anat., Bd. 8.

Osporn, H. L. 1904 On the habits and structure of Cotylaspis insignis Leidy.
Zool. Jhrb. Abt. f. Anat., Bd. 21, pp. 201-242.

1910 On the structure of Cryptogonimus chyli, etc. Jour. Exp. Zool.,
vol. 9, pp. 517-536.

1911 On the distribution and mode of occurrence of Clinostomum mar-
ginatum, ete. Biol. Bulletin, vol. 20, pp. 350-366.

Orro, Ricoarp 1896 Beitrige. z. Anat. u. Histol. der Amphistomeen. Deutsche
Zeit. f. Thiermedicin. u. verg. Pathol, vol. 22. (Inaug. Diss. Leip-
zig. ) A

Pratt, H. S. 1909 The cuticula and subcuticula of Trematodes and Cestodes.
Americ. Naturalist, vol. 43, pp. 705-729.

Porrier, J. 1885 Contrib. al’histoire des Trematodes. Arch. Zool. Exp., ser.
2, vol. 5, p. 465.

1886 Trematodes nouveaux ou peu connus. Bull. de la Soc. Philom.,
ser 7, tom. 8.

SALENSKY, W. 1874 Ueberd. Bauu. d. Entwk. der Amphilina. Zeit. f. w. Zool.
Bd. 24, pp. 28-32.

Scowartz, W. 1886 Die Postemb. Entwk. der Trematoden. Zeit. f. w. Zool.,
Bd. 43.

STAFFORD, J. 1896 Anatomical structure of Aspidogaster conchicola. Zool.
Jhrb. Anat., Bd. 9.

Stites AND Hassauu, 1898 Notes on parasites, no. 48. An inventory of the gen-
era and sub-genera of the trematode family Fascioloidae. Archiv f.
Parasitology, vol. 1, pp. 81-99.

‘Warp, H. B. 1901 On the structure of the copulatory organs in Microphal-
lus. Univ. Nebraska. Zool. Studies, no. 48, pp. 175-187.

Wriagut. ’79. Contributions to American Helminthology, I. Proc. Cana-
dian Institute, vol. i.

Zieeuer, H. E. 1883 Bucephalus und Gasterostomum. Zeit. f. w. Zool., Bd.
39, pp. 537-571.

STRUCTURE OF

CLINOSTOMUM 22e

ABBREVIATIONS

crs, cirrus sack

cu, cuticle

dej, ejaculatory duct

epo, outer part of epithelium of intes-
tine

epi, inner portion of the same

exbl, excretory bladder

excv, collecting vessel of excretory sys-
tem

expo, excretory pore

exrv, recurrent vessel of excretory sys-
tem

gl, glands communicating with the sur-
face

gpo, genital opening

int, intestine

lc, canal of Laurer

mc, circular muscle of body wall

mi, inner longitudinal muscle of body-
wall

mo, outer longitudinal of the same

mob, oblique muscles of body wall

mpl, longitudinal muscles of the paren-
chyma

mpl, transverse muscle of the paren-
chyma

mt, metraterm

nv, nerve collar

nvs, sensory nerve endings

oegl, oesophageal glands

oes, oesophagus

os, oral sucker

otp, ootype

ov, ovary

pgl, parenchyma glands

pt, parenchyma sheath of wall of intes-
tine

pn, parenchyma cell nucleus

prs, prostate part of cirrus organ

ps, parenchyma sheath of oesophagus

spn, spines of body wall

fa, anterior testis.

tp, posterior testis ”

ut, uterus

utsk, uterine sack

vs, ventral sucker

vsm, seminal vesicle

vd, vas deferens

vt, vitellaria

vid, duct from vitellaria

vir, yolk receptacle

PLATE 1

EXPLANATION OF FIGURES

All the figures (except 10, 13 a, and 16) were drawn with the Abbe camera
lucida. Most of them have been reduced one-third in reproduction; the magni-
fications are after this reduction.

1 View from the ventral surface C. marginatum, from a specimen from the
throat of Ardea herodias, fixed under compression in aqueous corrosive sublimate,
borax-carmine. X 12.

2 A partly schematic view from the dorsal side, combining facts from several
sections from a frontal series. From a bass worm. The vitellaria are not yet
developed, the uterine sack is not dilated, the excretory collecting vessel is omit-
ted from the right side and the recurrent vessel from the left, parts on different
levels are shown on the same level. X 27.

bo
bo
—

PLATE 1

STRUCTURE OF CLINOSTOMUM

HENRY LESLIE OSBORN

o

Sanna

7 wae ‘4
woiegrs

>
o
x
©

225

PLATE 2

EXPLANATION OF FIGURES

3 View combined from sections of a sagittal series, showing together organs
which are on different levels, mouth, ventral sucker, genital organs and excretory
pore are median while the intestine and the collecting and recurrent excretory
vessels are lateral. X 27.

4 Sections from a transverse series. The numbers show the number of the sec-
tion in the series; A is in the level of the oesophagus; B, in front of the ventral
sucker; C, at the ventral sucker; D, at the uterine sack; H, at the genital pore;
F, at the anterior testis; G, at the canal of Laurer. From heron, corrosive and
acetic, 1ron-haematoxylin. X 27.

5 The center section of a sagittal series, from a worm which died with the oral
field inverted. Heron, after chromic acid fixation and iron-haematoxylin. 40.

STRUCTURE OF CLINOSTOMUM PLATE 2

HENRY LESLIE OSBORN

227

PLATE 3

EXPLANATION OF FIGURES

6 A longitudinal section of the body wall of the dorsal surface, showing the
position of the various excretory vessels with reference to the muscular layers.
Heron, corrosive, iron-haematoxylin. > 240.

7 Body wall from a transverse section near the center of the ventral surface,
showing the uni-cellular glands (?); chromic acid, iron-haematoxylin. > 1100.

8 Two sense organs of the cuticle from the dorsal anterior region of the body.
x 1100.

9 Part of one of the fibers of the inner longitudinal muscle of the body wall,
showing the alternation of stained and unstained substance. Heron, corrosive,
iron-haematoxylin. 560.

10 Myoblast and its nucleus and adjoining muscle fibers of one of the longitu-
dinal parenchyma muscles. Heron, chromic, iron-haematoxylin. X 1100.

11 Part of one of the parenchyma muscles from the same series as fig.9. > 1100.

12 From asection passing vertically to the posterior region of the oesophagus.
Heron, chromic. X 560. ‘

13. Section from a fully matured worm vertical to the wall of the intestine,
showing the pseudopodial inner borders of the epithelium; the darker shading of
the deeper ends of the cells indicates the distinction between the stained and little
stained parts of the cell, corrosive, iron-haematoxylin. X 1100.

13 a Free hand drawing from a living worm from bass, showing the lateral
pouches of the intestinal caeca.

14 The epithelium of the intestine from an immature worm showing resting
stage of the tissue. Corrosive, iron-haematoxylin. X 1100.

15 Reconstruction from several adjoining sections of a frontalseries, showing
the relation of the collecting to the subcuticular cavities and to the bladder, also
the recurrent vessel and the intestine. X 36.

16 Free hand drawing from the posterior ends of the chief excretory vessels
as seen inaliving worm from the bass under slight compression. XX Zeiss oc. 2,
ob. A.

17 Reproductive system as seen from the ventral surface, from total prepara-
tions. The vitellaria have been omitted.

228

STRUCTURE OF CLINOSTOMUM PLATE 3
HENRY LESLIE OSBORN

THE DEGENERATIONS IN THE SECONDARY SPER-
MATOGONIA OF DESMOGNATHUS FUSCA

B. F. KINGSBURY anp PAULINE E. HIRSH

From the Histological Laboratory, Cornell University

TWENTY-ONE FIGURES

In 1902 there was published a part of the results of a study
of the spermatogenesis in Desmognathus fusca in which the
occurrence of degenerations! in the secondary spermatogonia
was mentioned. A fire had destroyed the larger portion of the
preparations and photographs covering this portion of the sper-
matogenetie cycle, and no effort was then made to complete the
investigation by a detailed study of the spermatogonia.

Of the degenerations that occurred two were particularly attrac-
tive—the regressive changes in the spent lobule and the degener-
ations in the last generation of spermatogonia, after the cessa-
tion of the season’s transformation into spermatocytes. These
degenerations are described as occurring in spermatogonia of the
last generation, although they might quite as well, perhaps, as
will appear subsequently, be given as involving spermatocytes
at the very beginning of their growth period.

The occurrence of degenerations in the amphibian spermato-
genesis has been known since the pioneer work of Flemming (’87)
on the spermatogenesis of Salamandra. He described (p. 447)
the vacuolation and fragmentation of the nucleus and the dis-
solution of the cell, constituting a form of degeneration already
described by him (’85). He says:

Irgend eine Beziehung zur Spermatogenese kann diese Erscheinung
keinesfalls haben, da sie zu einer viel friiheren Zeit auftritt, als jene;
in den Praparaten von Hoden mit Spermatogenese, welche ich bisher
studiert habe, waren derartige Bilder nicht vorhanden, ich will jedoch

! Kingsbury, ’02, p. 108, p. 111.
231

Doe B. F. KINGSBURY AND PAULINE E. HIRSH

nicht behaupten, dass sie nicht auch gleichzeitig mit der Samenfiden-
bildung noch vorkommen kénnten. Nach dem ganzen Habitus aber
handelt es sich offenbar um Processe der Degeneration und des Unter-
gangs von Kernen und Zellen, die aus einstweilen unbekannten Ursachen
zur Zeit der Epithelwucherung in manchen Cysten eintreten, und die,
wenn auch in der Form nicht ganz gleichartig, am néchsten vergleich-
bar erscheinen mit der chromatolytischen Entartung der Kerne im ovari-
alen Follikelepithel, die ich kiirzlich an anderem Orte beschrieben habe
(Flemming, ’85).

A definite localization in the spermatogenetic cycle was thus
not given, though it was recognized that it occurred much earlier
than the period of ‘spermatogenesis’ at the time of cell prolifer-
ation. The cause was neither ascertained nor suggested.

Hermann (’81), in the same material of investigation, devoted
(p. 99) more attention to these degenerations, describing the
changes in the chromatin and achromatic substance in the nu-
cleus, their altered staining reactions (as was subsequently done
by Heidenhain and others). He noted the common occurrence
of the degeneration in a certain kind of cell (his ‘spermatocytes’),
the degeneration of entire lobules and the freedom from degener-
ation of the follicle cells. He regarded the degeneration as nor-
mal and, in connection with the question of its significance, he
commented on the extravagance in the outlay of germ cells, call-
ing in as illustration the atresia folliculi.

Driiner (’93) devoted an article of several pages to the attempt
to show that the degenerations in the testis of Salamandra de-
scribed by Flemming and Hermann, were due to a parasite (pro-
tozoan, coccidian) whose spores entered the nucleus. His view
will be commented on subsequently. Later workers on amphib-
ian spermatogenesis (Meves, McGregor, Montgomery, and others)
have not, so far as we are aware, noted the occurrence of such
degenerations as have been described in Salamandra.

Whatever may be the condition in other salamanders, aside
from the European form in which these degenerations have been
recorded as given above, in Desmognathus studied by us they are
of constant occurrence. The material collected and studied cov-
ers a period of ten or twelve years and is from localities in the
vicinity of Ithaca one to two miles apart. The occurrence of the
degenerations becomes striking when it is recognized that they

DEGENERATIONS, DESMOGNATHUS 233

occupy in this form a distinct place in the spermatogenetic cycle.
Their occurrence, in fact, is so closely associated with the annual
spermatogenetic cycle and with the ‘polarity’ of the testis in
Desmognathus that a brief description may be introduced, even
at the expense of essential repetition of descriptions which have
been previously published (Kingsbury, ’02).

The spermatogenetic cycle in Desmognathus may be said to
begin in the fall or late summer, after the extrusion from the
testis of the spermatozoa formed during that season. During
the fall and winter months there is a multiplication of the sperma-
togonia and a tardy growth of the spermatocytes I, which in
midwinter seems practically suspended. Some spermatocytes
undergo division but the maturation divisions appear to be often
abnormal and the resulting cells to degenerate. In the spring
the multiplication of the spermatogonia and the growth of the
spermatocytes begins actively, characterizing particularly the
months of March, April and May, while divisions of the sperma-
tocytes occur in May, June and July. The transformation of
the spermatids into spermatozoa preponderates in August and
September.

In the late spring or early summer, transformation of the ‘last
generation’ of spermatogonia stops, no more spermatocytes
beginning their period of growth, until fall, and it is after this
time, when the growth of new spermatocytes has ceased, that the
degenerations in question occur. The cells undergoing degenera-
tion have been spoken of as secondary spermatogonia of the last
generation although, since they come intermediate between the
spermatocytes I and the secondary spermatogonia, they might,
perhaps, be equally as well designated as young spermatocytes I.

The following out of the sequence of stages in the spermato-
genesis of Desmognathus is facilitated by the marked polarity
of the organ in thisform. This polarity seems to be rather charac-
teristic of many at least of the tailed amphibia (Meves, McGregor,
Kingsbury, ete.) and particularly perhaps of the smaller or more
elongated ones. The changes of spermatogenesis proceed as a
‘wave’ in a cephalo-caudal direction so that at the proper season

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

234 B. F. KINGSBURY AND PAULINE E. HIRSH

of the year, mature or maturing spermatozoa fill the lobules at the
cephalic end while primary spermatogonia occupy the caudal
filament of the gland, successive stages occupying in orderly
sequence the intervening groups of lobules. Compare Kingsbury
(02), text figure A.

In Desmognathus, in the spring and in some cases in early
summer, the transition from spermatogonia to mature spermato-
cytes I, as shown in longitudinal section of the organ, is gradual,
but as the season advances a ‘boundary plane’ between those
destined to furnish spermatozoa that season and those that will
hold over becomes more and more evident. The ‘spermatoge-
netic wave’ has stopped short ata particular point; the progressive
changes of spermatogenesis continue in one portion and lag or
are lacking in the other portion; and it is in those lobules just
behind (caudad of) this boundary that the degenerations are
constantly found.

How striking this boundary between the two regions becomes
may be seen by comparing figures 1, 2, 6 and 3 which reproduce
longitudinal sections of the testis: of Desmognathus from the
months of May, June, August and September respectively, the
transverse lines drawn upon the photograph indicating the plane
of boundary which in figure 1 is not yet established as such.
Figures 4, 7, and 5 give enlarged views of the region of the bound-
ary and in them the cysts and lobules filled with degenerating
cells are seen. Figure 7 is from a section neighboring that shown
in figure 4 which is an enlargement from figure 2, while figure 5 is
an enlargement from the section of figure 3. As is well recognized
in urodele spermatogenesis, the cells contained in and composing
a cyst are in the same or closely contiguous stages of develop-
ment, and while in some instances, cells occur singly or a few
together, in the great majority of cases, the contents of the entire
cyst or even lobule are undergoing degeneration together. In the
degenerating cysts and lobules however it is the germ cells and
not the follicle cells forming the cyst wall that degenerate, the
nuclei of the latter being apparently normal and distinguishable
amid the degenerating spermatogonia as Hermann pointed out
and figured. Figures 8 and 9 show degenerating lobules while

DEGENERATIONS, DESMOGNATHUS 239

figures 10, 11, 12, and 13 show cysts in different stages of degen-
eration. In most of these figures the nuclei of the follicle cells
may be recognized in the midst of the degenerating germ cells,
fragments and débris.

Although occurring constantly and abundantly in the summer
months in this region of the testis, the degenerations appear
sporadically at other seasons. They have been found in late
fall (November) and in the spring (April). They are not as
abundant at these seasons but occur in the same region, 1.e., in
the zone intermediate between the secondary spermatogonia and
the growing spermatocytes. It is obviously difficult to determine
whether such cellsare spermatogonia or are young spermatocytes,
or whether they are all destined to disintegrate or whether some
may ‘recover’ from the extreme ‘contracted’ condition of the
nucleus which is so characteristic of early stages in this form of
degeneration. Itwasthis that first suggested a comparison with
synizesis. This, however, will be discussed subsequently.

The determination of the exact sequence of changes that occur
in the cell is difficult as it involves the larger question of the struc-
ture and functional changes in the nucleus and cell body. No
attempt has been made to follow systematically the changes in
the cytoplasm which are elusive and attention has therefore been
turned more particularly to the nucleus where the effect is strik-
ing.

Degeneration in the nucleus seems to set in when the cell is
preparing for mitosis. In the prodromic stage the nuclei possess
a clear appearance with a definite delicate reticulum. It should
of course be appreciated that the point of immergence is inferred
from the structure of the apparently unchanged nuclei in the same
cyst or the same lobule. This inference may be seen to be justi-
fied if it be remembered that the cells in the same cyst are in nearly
identical stages, the same applying, but less closely, to the lobule.
Three such nuclei are shown in the upper left hand corner of figure
12. The definite changes of degeneration are initiated by the
contraction (collapse) of the nucleus such as is shown in figure 10.
There then appear in the degenerating nuclei numerous spherical
masses often lying in clear spaces as if in vacuoles (figures 11 and

236 B. F. KINGSBURY AND PAULINE E. HIRSH

12). With iron hematoxylin they retain the stain strongly as do
the nucleoli. The application of a more differential stain, such as
the Biondi-Heidenhain shows that they are not basi-chromatin
(figures 20, 21). In the next stage the nucleus ‘runs together’
into a more or less compact mass in which however the chromatin
and parachromatin portions are usually distinguishable. Very
often the latter in some globular form adheres to the chromatin
mass as though ‘squeezed out’ (achromatic body of Hermann).

There are many forms of the degeneration picture. Quite
common is the typical chromatolytic—or preferably and more
correctly karyolytic—nucleus described by Heidenhain (’90) in
which the chromatin collects peripherally as a shell, series of
globules, or very often a crescent, the achromatic mass being
central. Figure 9 shows one of these inadequately while figures
18 and 19 give them in more detail. The condensation and corre-
sponding shrinking of the nucleus during this stage is usually exces-
sive so that a considerable space intervenes between the nucleus
and the cytoplasm, which throughout appears scanty and consists
largely of a peripheral layer which may be connected with the
nuclear mass by strands.

Further changes consist in the dissolution by fragmentation
(and liquefaction?) together with a loss of staining power with
basic stains. The resulting mass of granular and globular débris
contains much fat which is undoubtedly responsible for many of
the vacuoles apparent in figures 7 and 8.

Significance. In considering the significance of these degener-
ations the first thought would be that they were pathological—
a result of an infection or an abnormality introduced into the life
conditions (lack of oxygen, insufficient circulation, etc). This,
indeed, is the interpretation of Driiner, who believed that the
degenerations were caused by protozoa infesting the nucleus and
causing their degeneration. His figures and descriptions how-
ever are not conclusive and in the absence of any experimental
evidence and in the light of the occurrence in Desmognathus of
these degenerations, at a specific time and in a specific region
such an interpretation becomes highly improbable. Likewise
there is no indication of interference with the blood supply which
might cause their degeneration secondarily.

DEGENERATIONS, DESMOGNATHUS Dorks

On the other hand, there is much that indicates that they are
physiological degenerations. The constancy with which they occur,
the definite time of year at which they are most abundant,
and especially the location in the testis and the position in the
spermatogenetie cycle. This last suggests to us that they bear
a relation to the regulation of the spermatogenetic process. Such
a regulation appears not to have received much consideration.
The investigator’s attention is usually so riveted upon the in-
tensely interesting intracellular processes involved in the periods
of spermatogonial multiplication, synapsis, reduction, ete., that
the possibility that these processes may be in some way coordi-
nated with the life-habits of the form is generally not discussed.
Nevertheless, it will be seen that spermatogenesis is correlated
with the mating habits in those forms at least which mate at
a definite season. If the multiplication of the spermatogonia and
transformation into spermatocytes I is initiated, accelera‘ed,
retarded or checked at a definite stage, it can mean nothing else
than that these processes are regulated growth processes similar
to those that lead to the establishment of definite body form.
Whether this regulation is intrinsic—within the germ cells them-
selves—or extrinsic is a question for the consideration of which
there is as yet little basis of fact although analogy would suggest
the latter. Possibly the emptying and degeneration of the first
maturing lobules and the growth of the interstitial cells accom-
panying this may be afactor. Degenerations, therefore, so closely
associated with a break in the continuity of the growth process
of spermatogenesis seem associated with its regulation, and this
view is strengthened by the fact that only the germ cells, not the
follicle nuclei, undergo the histolytic change.

Relation to synizesis. The marked contraction of the nuclear
contents into a more or less homogeneous mass which has been
described in so many plant and animal forms as occurring at
about the beginning of the growth period of the spermatocyte I
so closely resembles in its extreme form the rounding off of the
nuclei entering on the degeneration changes above described as
almost to force a comparison, and one of us (Kingsbury, ’02),
struck by the strong general resemblance, made a suggestion that

238 B. F. KINGSBURY AND PAULINE E. HIRSH

introduces for consideration in this connection one of the critical
stages of o6genesis and spermatogenesis which appeals to the _
writers as one of the most difficult of interpretation—the so-
called synapsis stage.’

The suggestion then tentatively made?* was that the resemblance
between the nuclei in synizesis and the degeneration under con-
sideration might be more than a superficial one, especially as they
both occurred at about the same time, associated apparently with
the end of the multiplication period; that synizesis Hself might
possibly represent a ‘running out of the spermatogonial stock.’
Synizesis would on this interpretation be a ‘beginning degenera-
tion’—with recovery, which passes over later in the season into
a degeneration leading to dissolution.

The following diagram or schema may serve to illustrate for
the form Desmognathus the comparison of synizesis and the
degenerations in question. Successive stages in the spermato-
genesis being indicated by the letters of the alphabet as given in
the legend below, while the idealized zones of the testis are num-
bered from 1 to 10.

2 The terminology employed in the discussion of this period of the gametogenesis
possibly calls for brief comment. Synapsis is used in the original sense of the
pseudo-reduction in the chromosome number interpreted as due to a joining
together in pairs. It isthereforeas used here equivalent with conjugation and syn-
desis. For the contraction of the nuclear contents the term Synizesis introduced
by McClung is employed.’ While synapsis and synizesis are usually reported as
occurring together at the beginning of the growth period of the spermatocyte, after
the last spermatogonial division, they are not in all cases so assigned. Mont-
gomery places synapsis in the telophase of the last spermatogonial division. Miss
KXing described it in the toad as occurring after the growth period of the spermato-
cyte, etc.

3 Page 108. “‘(c) when the spermatogonia cease to undergo transformation
into spermatocytes in the summer, the last cysts of spermatogonia apparently
undergo a chromatolysis and solution, and the boundary between the spermato-
cytes which are to form spermatozoa that season and the spermatogonia remaining
over until the next summer, is thus well marked.”’ Page 111. ‘‘It is suggested

‘therefore that the contraction figures [i.e., synizesis], instead of being construc-
tive and a fundamental phenomenon in the formation of the spermatocyte, may
be an expression of a ‘running out’ in the spermatogonium stock and represent a
tendency toward degeneration. We know as yet too little of the occurrence of the
contraction figures in different forms to draw any general conclusions; possibly
quite different phenomena may be here included.”’

DEGENERATIONS, DESMOGNATHUS 239

With the horizontal axis representing the ‘spermatogenetic wave
and the vertical axis the successive transformations with ad-
vancing season, oblique lines upward and to the right would give
the similar stages at different seasons. The ‘boundary plane,’
when it appears, breaks the continuity of such lines. Allowing
for the equalization of all stages in duration and extent which
such a schema necessitates, it nevertheless gives a good diagram-
matic representation of the process of spermatogenesis in the
testis of Desmognathus as is indicated by comparing figures 1, 2,
6 and 3, from the months of May, June, August and September.
To these might be added many others from the yearly cycle. It
will be seen the synizeses (#) in front of the ‘boundary plane’
is in line with the degenerations (D) behind the plane.

Spermatogenetic wave

1 2 5) + 5 6 7 8 9 10

n ido amee all k ] i + Ds ee b a
n myo Kk j i h # Day ce rae a
=|
B
é n Tian) s) 1 h g . Die cc b a
ates
abbas « ml 4 h g f Ss Dec b a
5
Ss.
Sito aml he) Se f E c bate a
oc
<
n m-l ¢g f Be oe c b a a
n ] fi f c c b a a a
a—primary spermatogonia b —secondary spermatogonia
c —secondary spermatogonia D —degenerations
E—synizeses f —primary spermatocytes
g—secondary spermatocytes h —spermatids
i —transforming spermatids j —maturing spermatozoa
k—nearly mature spermatozoa 1 —mature spermatozoa
m—spent lobules (degenerating) n —degenerated lobules

* —boundary plane

240 B. F. KINGSBURY AND PAULINE E. HIRSH

A second possible interpretation of synizesis that occurred to
the writers when considering the resemblance of the degeneration
figures to extreme synizesis, has been elaborated by Hertwig (’03) ;
that is, that synizesis and synapsis represent an abortive mitosis.
According to this view, on the one hand, synizesis represents an
‘attempt on the part of’ the spermatogonia to divide again—
which fails; while, on the other hand the reputed conjugation of
chromosomes occurring at about this time is but the imperfect
fission and subsequent fusion of daughter chromosomes of such
abortive division. There promises to be some time before there
is any complete agreement as to the facts, let alone interpretation.

As far as synizesis is concerned, the extent of the contraction
of the nuclear contents seems to vary, from a condensation in
which no detail of structure is discernible, to a tendency orly, on
the part of the nuclear structures to withdraw from the nuclear
membrane. Since Meves (’07), in his recent rather severe critique
of the synizesis and synapsis problem, is forced to admit such a
tendency at this stage of the growth of the spermatocyte, syni-
zesis must represent a real alteration of conditions, and is not an
artifact due to imperfect penetration or fixation. In Desmog-
nathus, we still locate synizesis in the beginning of the growth
period of the spermatocyte. The contraction of the chromatin
in many specimens is not marked so that in mary of the prepara-
tions it ean be interpreted as little more than a ‘tendency’ to con-
traction. Furthermore, as was stated by Kingsbury (’02),
synizesis is only well marked in the early summer, among the
last spermatocytes to enter upon the growth period that season.
In this connection figure 4 and particularly figure 7 may be
examined, as well as the more enlarged figures 8 and 9 which are,
however, not particularly characteristic and are not introduced

The suggestion was too briefly stated at that time to be easily interpreted. The
idea intended to be conveyed however was that in the ‘play of forces,’ whatever
their character, which determine the succession of spermatogonial divisions, the
termination of the period of multiplication must be thought of as due to a check-
ing of, or a loss of a power of nuclear synthesis—a ‘running out,’ as if from an ex-
haustion of ‘material’ which necessitates a long growth period—or leads to degen-
eration. The suggestion was hardly intended to have the force of a ‘claim’ as
Miss King (’08) states it. B. F. Kingsbury.

DEGENERATIONS, DESMOGNATHUS 241

in illustration of the synizesis figure in Desmognathus. The
limitation of synizesis to this period of the year is not due to the
fact that at this time the ‘‘testis contains relatively more cells
in this particular stage of development” as has been intimated
might be the case (King, ’07). To appreciate this it is necessary
to keep in mind the ‘polarity’ of the urodele testis which permits
a very exact location of given stages and determination of their
sequence. Thus the examination of longitudinal sections of
organs secured throughout the spring shows in each a succession
of eysts filled with cells in stages grading from the spermatogonia
to mature and dividing spermatocytes. As has been said, it is
only after the ‘boundary’ limiting that season’s production is
well marked that definite instances of synizesis appear—unless
indeed, the isolated cysts of cells with markedly contracted nuclei
that are found in testes from the spring months before the
boundary plane appears represent synizesis. In this event, the
cells recover and are not degeneration figures.

As far as synapsis (syndesis) or the ‘conjugation of the chro-
mosomes’ is concerned, it appears to be lacking in Desmognathus.
We have carefully reéxamined the question in extensive material
and fail to find any indication of a fusion of the chromosomes,
parallel or end to end, or, indeed, of a splitting of the chromatin
threads in the early growth period of the spermatocyte. The
splitting of the chromosomes of the spermatocyte I in preparation
for the first division appears quite early; but it becomes more and
more distinct and complete as the division is approached. The
changes of the growth period of the spermatocyte I occur essen-
tially as already described by Kingsbury (’02).

Whatever general agreement may be ultimately reached as to
the facts, i.e., the general occurrence of a union in the spermato-
cyte (or spermatogonial anaphase) of distinct chromosomes, end
to end or parallel, and the prevalence of a contracted condition
of the nuclear contents—the explanation of the phenomena re-
‘mains quite distinct, nor should it be confounded with whatever
teleological significance may attach thereto. Thus such an hy-
pothesis as the abortive mitosis interpretation of the synapsis
period by R. Hertwig seems particularly suggestive, since it pre-

242 B. F. KINGSBURY AND PAULINE E. HIRSH

sents the possibility of an explanation on the basis of a general
interpretation and treats the cell as a unit.

Synizesis, as an alteration in the morphology of the nucleus, can
be adequately approached only by a consideration of the ‘play
of forces‘ upon which the morphology of the nucleus depends and
in which a correlation with the cytoplasm must be intimately
involved. An adequate analysis of such forces has not, as far as
we are aware, been made. One is particularly impressed with the
existence of such forces when in karyolysis, as a result of their
suspension, the nuclear substance is free to follow the (simpler)
laws of its physical state and condense into spherical masses. It
is this which suggested that synizesis expressed a more or less
complete suspension of nuclear processes. Since the contraction
is toward the idiosome the impression is strong that, in the con-
traction, the relations of the nucleus to that portion of the cyto-
plasm in which the idiosome is, persist or exist, possibly in exag-
gerated form, while there is a more or less complete suspension of
nuclear-cytoplasmic relations over other portions of the nuclear
membrane.

The arrangement of the chromatin (chromosomes) oriented in
relation to the idiosomatice cytoplasm in the well known ‘bouquet
stage’ indicates that such a peculiar interrelation between the
nucleus and this portion of the cytoplasm exists (persists) through-
out the growth of the spermatocyte I.

Suggestions of such important correlations are naturally to be
found in the literature. Thus, Winiwarter (’08) recognized
synizesis as expressing a correlation between the chromosomes and
the idiosome, the latter exerting a real influence of attraction
upon the chromatin filaments, but affecting the cytoplasm as
well, since the mitochondria cluster around the idiosome. As
expressing the attraction he proposes the term centrotaxis, but
of its nature we are entirely ignorant as yet. Montgomery (’11)
suggests that synizesis, which he finds may occur during a large
portion of the growth period of the spermatocyte in Euschistus

4 By this somewhat unsatisfactory expression is meant the sum total of forces

that are undoubtedly operative in protoplasm—electrical attractions and repul-
sions, chemical affinities and reactions, osmotic tensions, etc.

DEGENERATIONS, DESMOGNATHUS 243

indicates possibly a rhythmic discharge of material from the nu-
cleus. The chromatin plate described by him likewise is indicative
of an idiosome-nuclear correlation.

Practically nothing is known regarding the frequency of occur-
rence in amphibia other than Salamandra and Desmognathus, of
degenerations similar to those described. Through the kind-
ness of Dr. Montgomery sections of the testis of Plethodon cine-
reus erythronotus were examined and comparable degenerations
were found to be present. Likewise they have been seen in the
testis of Salamandra atra. In these forms, however, no system-
atic study of the degenerations has been made in which there has
been seriously attempted the ascertainment of any definite rela-
tion to the process of spermatogenesis, the stage’at which they
occur, their relation to the annual cycle or their location within
the testis, nor has the relation of the spermatogenetic process to
the testis been studied.

Miss King (’07) has found no trace of such degenerations in
Bufo. She says: “I have never found a condensation of the
chromatin in the spermatogonia as Kingsbury has described for
Desmognathus, and I am unable to confirm his statement that
‘contraction figures do not occur constantly in spermatocytes.’”’
To this the following comments may be made; first, that work
upon one form cannot be relied on for confirmation or disproof of
work done upon another form. The spermatogenetic process
seems to be worked out in the anuran testis in a way quite dif-
ferent from that prevailing among the urodeles. In the toad it
is apparently intralobular; many different stages are found with-
in the confines of a single lobule.® The seriation of stages in such
a testis as the toad’s are much more difficult, and, it may be ven-
tured, karyolytic nuclei might easily be overlooked as they would
probably occur singly. This, however, from Miss King’s careful

» Cf. King, ’07, p. 346. ‘‘As arule all of the cells in a cyst are in approximately
the same stage of development, but a single follicle [lobule] may contain both
spermatogonia and maturing spermatids. A transverse section of the testis,
therefore, shows practically all stages in the development of the spermatozoa.’’
In Desmognathus in a transection all cells would be in approximately the same
stage, while in a single longitudinal section at the right season of the year, prac-
tically every stage might be seen.

244 B. F. KINGSBURY AND PAULINE E. HIRSH

study would hardly seem likely and it is far more probable that
if such degenerations occur, they do so at a later season than that
studied by Miss King (i.e., after September), possibly at the be-
ginning of hibernation, when, if ever, one might expect a checking
of the spermatogonial divisions to be accompanied by degenera-
tions. Granted that these are ‘physiological degenerations,’ it
should be appreciated that there is no reason for believing that
the factors upon which they depend would be operative in all
forms in the same way. The degenerations might or might not
oecur, which fact should be considered in making comparisons
between the processes taking place in the testis of the toad and in
that of the salamander where this may be particularly applicable.

The lavish outlay of germ cells and their wholesale degenera-
tion has been commented on by a number of writers. There may
be particularly mentioned: Winiwarter and Saimont (mammals),
Hoffman (’92), D. Hollander (’05) (birds), Bouin (01), Dustin
(07), Levi (’05) (amphibia).

The descriptions of Winiwarter who has given the most mono-
graphic description of mammalian oogenesis are particularly
interesting. In the rabbit (00) he described two epochs of de-
generations which were completely separated. Of these the first,
including typical and atypical karyolysis, is of particular interest
in this connection. The second occurred in the atresia folliculi.
Winiwarter found that the multiplication of the oogonia ceased
soon (about ten days) after birth. The degenerations of the first
epoch extended from the twenty-third day embryo nearly to
eighteen days after birth. The degenerating cells were found to
be, in the large majority of cases at least, oogonia which suc-
cumbed particularly at the time of their division. At what
point the karyolysis sets in he is not sure but he thinks that in all
probability it is at the equatorial plate stage. Comparison of his
results with the conditions in Desmognathus are quite suggestive.

In the cat (’08) the degenerations begin to appear singly. In
embryos of forty-five to fifty days large groups of degenerating
cells are present. Shortly after birth the multiplication of the
oogonia suffers an arrest and, coincident with this, there is a
recrudescence of degeneration. The groups of nuclei particularly

DEGENERATIONS, DESMOGNATHUS 245

affected are his poussieroides and transitory (i.e., presynaptic).
The change shows itself first in the nuclei; the fine network or fine
granulations forming larger masses grouped around the nucleolus.

CONCLUSION

In Desmognathus fusea at the time that the transformation of
spermatogonia into spermatocytes ceases, degeneration figures in
large numbers, involving whole cysts and lobules, may be found
among the cells that have ‘failed’ to transform thatseason. They
have a definite position in the testis as well as in the spermatoge-
netic cycle and seem to be closely associated with the regulation of
the spermatogenetic process.

Apparently similar degenerations have been reported in a num-
ber of different forms, in the oogenesis and in the spermatogenesis.

Such degenerations have undoubtedly a greater significance
in connection with the activities of the reproductive organs than
is generally recognized.

For their adequate treatment, however, spermatogenesis must
be dealt with in its relation to the whole organ and the whole
organism. Little help in interpretation, it is believed, may be.
expected from the ultra-chromosomal point of view, or even from
that of the cell theory.

January 20, 1912,

246 B. F. KINGSBURY AND PAULINE E. HIRSH

LITERATURE CITED

«
BizzozeRO 1888 Anwendung des Methylgriines zur Erkennung der chemischen
Reaktion und des Todes der Zellen. Virchow’s Archiv f. pathol.
Anat., Bd. 113.

Bowtin, M. P. 1901 Histogenese de la glande genitale femelle chez Rana tem-
poraria. Arch. f. Biol., T. 17.

1903. Spermatocytes en degeneration utilizes comme material alimen-
taire pendant la spermatogenese. Compt. rend. Soc. Biol., T. 55.

Drtner, L. 1894 Beitrige zur Kenntnis der Kern- und Zellen-degeneration und
ihre Ursache. Jenaische Zeitschr. f. Naturw., Bd. 28.

Dustin, A. P. 1907 Recherches sur l’origine des gonocytes chez les Amphibiens.
Arch. de Biol., vol. 23.

FLEMMING, W. 1885 Ueber die Bildung von Richtungsfiguren in Saiugetiereiern
beim Untergang Graafischer Follikel. Arch. f. Anat. u. Entw.

1887 Neue Beitrage zur Kenntniss der Zelle. Arch. f. mikr. Anat.
Dh. 2. Bdz29:

HeEIpENHAIN, M. 1890 Beitrige zur Kenntniss der Topographie und Histologie
der Kloake. Arch. f. mikr. Anat., Bd. 35.

1892 Ueber Kern und Protoplasma. Festschr. f. Kélliker.

HerRMANN, L. 1888 Ueber regressive Metamorphosen des Zellkerns. Anat.

Anz., Bd: 3.
1890 Beitrige zur Histologie des Hodens. Arch. f. mikr. Anat., Bd.
34.

Hertwie, R. 1903 Ueber Correlation von Zell- und Kerngrésse und ihre Be-
deutung fiir die geschlechtliche Differenzierung und die Teilung der
Zelle. Biol. Centralbl., Bd. 23.

HorrMann, C. K. 1892 Etude sur le developpement de l’appareil urogenital
des oiseaux. Verh. d. Koninkl. Akad. van Wetensch. Amsterdam,
sec. 2, vol. 1.

D’Houuanper, IF. 1905 Recherches sur l’oogenese et sur la structure et la signi-
fication du noyau vitellin de Balbiani chez les oiseaux. Arch. d. Anat.
Miner... Le. 7.

KinGc, Heten Dean 1907 The spermatogenesis of Bufo lentiginosus. Am.
Jour. Anat., vol. 7.
1908 The odgenesis of Bufo lentiginosus. Jour. Morph., vol. 19. .

Kineassury, B. F. 1902 The spermatogenesis of Desmognathus fusca. Am.
Jour. Anat., vol. 1.

Levi, G. 1905 Sulla differenziazione del gonocita e dell’ ovocita degli Anfibi

con speciale reguardo alle modificazione della vescicola germinativa.
Arch. Ital. di Anat. e di Embriol., T. 4.

DEGENERATIONS, DESMOGNATHUS 247

McCuuna, C. E. 1905 The chromosome complex of Orthopteran spermatocytes.
Biol. Bull., vol. 9.

McGreeor, J% H. 1899 The spermatogenesis of Amphiuma. Jour. Morph.,
vol. 15.

Meves, Fr. 1895 Ueber eigentiimliche mitotische Prozesse in jungen Ovocyten
von Salamandra maculosa. Anat. Anz., Bd. 10.

1907 Die Spermatocytenteilungen bei der Hénigbiene (Apis mellifica,
L.), nebst Bemerkungen iiber Chromatin-reduktion. Arch. f.mikr.
Anat., vol. 70.

1911 Ueber die Beteiligung der Plastochondrien an der Befruchtung
des Hies von Ascaris megalocephala. Arch. f. mikr. Anat., Bd. 76.

Montcomery, T. H.1903 The heterotypic maturation mitosis in Amphibia and
its general significance. Biol. Bull., vol. 4.

1906 Some observations and considerations upon the maturation phe-
nomena of the germ cells. Biol. Bull., vol. 6.

1911 The spermatogenesis of an hemipteron, Euschistus. Jour.
Morph., vol. 22.

Scumaus, H., und Ausrecut, E. 1894 Degenerationen von Mitosen. Ergeb.
d. allg. Pathol., Bd. 1.

1894 Physiologische Degenerationen. Ergeb. d. allg. Pathol., Bd. 1,
2te. Teil.

Winiwarter, H. v. 1900 Recherches sur l’ovogenese et lorganogenese de
V’ovaire des Mammiferes (Lapin et Homme). Arch. de Biol., T. 17.

WINrIWARTER, H. v. et Satmont, G. 1908 Nouvelles recherches sur l’ovogenese
et l’organogenese de l’ovaire des mammiferes (Chat). Arch. de Biol.,
T. 24, Ch. 4.

PLATE 1
EXPLANATION OF FIGURES

1 Longitudinal section of testis of Desmognathus fusca. May 24; fixed in
Hermann’s fluid; iron hematoxylin stain. From the primary spermatogonia in
the portion at the bottom of the photograph, one passes up through a gradual
succession of stages to spermatids at the top of the figure. The nearly mature
spermatocytes 1 just below them may be distinguished by their larger size. The
boundary between the cells destined to furnish spermatozoa that season and the
residual cells, is not yet evident in this specimen.

2 Longitudinal section, testis Desmognathus fusca, fixed June 7, in Hermann’s
fluid; iron hematoxylin stain. The interstitial cells about the degenerated lobules
occupy the upper portion, below them come developing spermatozoa; at the lower
end, spermatogonia. The boundary is becoming well marked. Enlarged photo-
graph of the boundary region is shown in fig. 4.

3 Longitudinal section, testis Desmognathus, fixed September 7 in Hermann’s
fluid; iron hematoxylin. The body of the testis occupied by spermatozoa, the
lower portion by spermatogonia. The boundary is thus conspicuous. Fig. 5
shows an enlarged view of the boundary. An intermediate stage is shown in fig. 6.

4 Testis Desmognathus. Photograph giving an enlarged view of the boundary
region of fig. 2. Spermatocytes 1 occupy the upper lobules, spermatogonia the
lower lobules, the separation being particularly clear on the right side. Three
lobules of spermatogonia at the boundary region show more or less extensive
karyolysis.

5 Testis of Desmognathus. Photograph giving an enlarged view of the bound-
ary region of fig.3. Spermatozoa occupy the lobules above the boundary, sper-
matogonia those below. Degenerations are seen in one entire lobule and in
portions of two others.

248

DEGENERATIONS, DESMOGNATHUS PLATE 1
B. F. KrincsBurY AND PAULINE E. HtrsuH

£
>
%

*

oee,*
ooh

3

.
*

. *. >
~ J oe 2

249

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

PLATE 2
EXPLANATION OF FIGURES

6 Testis of Desmognathus. Longitudinal section. Fixed in Hermann’s fluid
August 21; iron hematoxylin stain. Two lobes are shown, the lower end of the
one on the right being connected with the lower end of the one on the left. In the
larger lobe, the region of degenerated lobules occupies in the figure the upper
end. These are succeeded below by spermatozoa, maturing spermatozoa, trans-
forming spermatids, spermatids, spermatogonia, the boundary between the regions
occupied by the last two being well shown in both lobes.

7 Testisof Desmognathus. Longitudinal section. Fixed June 7 in Hermann’s
fluid; iron hematoxylin stain. View of the lobules in the boundary region. In
the upper three lobules are growing spermatocytes 1; the intermediate three
lobules show synizesis; below these comes the boundary and lobules of sperma-
togonia among which many cysts are undergoing degeneration.

8 Testis Desmognathus. Photograph of the boundary region, showing a
degenerating lobule, filled with débris, fat vacuoles, karyolytic nuclei and the nor-
mal nuclei of the follicle cells. Above the degenerating lobule the spermatocytes
1 indicate a slight condition of synizesis. The lobule below contains spermato-
gonia.

9 Testis of Desmognathus. Photograph similar to that shown in fig. 8, from
the surface of the testis (on the right side of figure). One degenerating lobule is
shown and portions of two others contain degenerating cysts.

yr

=v

DEGENERATIONS, DESMOGNATHUS PLATE 2
B. F. Ktycsspury AND PAULINE E. HirsH

PLATE 3
EXPLANATION OF FIGURES

10 Testis of Desmognathus. Photograph at high magnification (2 mm. apo-
chromatic, no. 2 Zeiss projection ocular) of a portion of a lobule of spermatogonia
in which one cyst is in an early stage of degeneration. The resemblance to marked
synizesis is striking.

11 Testis of Desmognathus. Photograph at high magnification (2 mm. apo-
chromatic objective no. 2 Zeiss projection ocular). To show the karyolytic nuclei
in a degenerating lobule and the normal nuclei of the follicle cells. The lobules
above contain spermatozoa; those below dividing spermatogonia.

12 Testis of Desmognathus. A cyst filled with karyolytic cells. The three
nuclei in the neighboring cyst (upper left hand corner) are probably just about to
enter upon degeneration.

13 Testis of Desmognathus. Photograph from a lobule of degenerating sper-
matogonia.

14,15, 16,17 Penand ink drawings of camera lucida sketches of typical karyo-
lytic cells. Theresemblance to the extreme synizesis nuclei described in other
forms is striking.

18, 19, 20, 21. Colored drawings from camera lucida sketches of typical karyo-
lytic cells. From preparations stained with the Ehrlich-Biondi-Heidenhain
triple stain.

DEGENERATIONS, DESMOGNATHUS PLATE 3
B. F. Kincssury AND PauLine E. Hirsa

THE EPITHELIUM OF TURBELLARIA

RR. . YOUNG

From the State University of North Dakota
SIX FIGURES

The existence of an epithelium in trematodes and cestodes has
been a much debated question for many years, while among those
who deny the presence of this tissue in these worms there is much
difference of opinion as to the origin of this condition, some main-
taining that the epithelium has been lost, some that it has been
metamorphosed into the cuticula, while others fail to express an
opinion on this point.

If we turn to the Turbellaria, the probable ancestors of tre-
matodes and cestodes, we find conditions which I believe give
a clue to the answer to this question. Most of these possess a
typical nucleated ephthelium, the cell boundaries of which how-
ever are difficult to observe without special methods; while in
the pharynx and its sack nuclei are often lacking in the epithelial
layer, having invaded the parenchyma to form an insunken epi-
thelium as has been experimentally observed by Jander (’97) in
the regenerating pharynx of Dendrocoelum.

In many forms the general epithelium presents conditions simi-
lar to those common in the pharynx, while in some cases it has
been claimed that neither nuclein or cell boundaries are demon-
strable even with special methods of technique. Thus Bohmig
(90) in several species of Alloeocoela was unable to demonstrate
cell boundaries in the epithelial layer by treatment with silver
nitrate, while the distribution of nuclei was very irregular.

The typical conditions in the turbellarian epithelium are well
represented in Planocera inquilina (fig. 1). The surface of this

255

9256 R. T. YOUNG

worm is covered by a layer of cilia averaging 5.7 1n thickness
dorsally and 5.4 ventrally.!. In fixation, they become matted
together to form a tangled mass, in which it is difficult to observe
individual cilia. Where they are inserted in the epithelium,
the characteristic basal swellings give the appearance of a thin
dense layer at the surface, which has been interpreted by various
writers as a cuticula. The epithelial cells have a fibrillar struc-
ture presenting numerous small spaces? which are very likely the
result of shrinkage.

The course of the fibrillae is more or less irregular, though in a
general way perpendicular to the surface, producing the striations
mentioned by various authors in the turbellarian epithelium.
They form a close network, varying in density from point to
point, in the meshes of which lie the spherical or ovoid nuclei with
a distinct chromatic network and a definite membrane which
stains similarly to the latter, and in most, but not all cases, ap-
pears to be complete. Frequently, but not always, the network
is condensed at one or more points to form false nucleoli. This
appearance may be the result of shrinkage. These nuclei aver-
age 4.4 by 5.5u in diameter.’ Besides nuclei, numerous rhabdoids
occur in the epithelium, an account of which does not concern
us here. The fibrillae appear to be continuous with basal exten-
sions of the cilia, but on this point I cannot speak positively.

Cellular outlines, faintly evident in surface views, are indis-
tinguishable in cross sections. I have not observed a differen-
tiation of epithelial and interstitial cells, as described by Lang (’84)
for polyclads. Nor can I distinguish the difference in size of
nuclei which he describes and figures.

1 Average of seven measurements. Variations in thickness of the ciliary layer
of from 3 to 8u occur. These are probably not altogether normal, but where indi-
cations to the contrary are lacking I have included them in my averages.

2 Hxcept in the denser surface layers.

3 Average of twenty measurements.

4 Occasionally I find an apparent fibrilla passing through the epithelium from
outer to inner surface of the latter. These may represent cell boundaries but
they are distinguishable from other fibrillae only by their extent from surface
to surface of the epithelium and by their more nearly vertical direction. In
some places where shrinkage has occurred apparent cell outlines may be seen.

EPITHELIUM OF TURBELLARIA ARI

The thickness of the epithelium as well as the shape of the
nuclei is largely dependent on the state of expansion or con-
traction of the worm. Dorsally it averages 10u, ventrally 8.7
in thickness.°

Directly beneath the epithelium is the basement membrane.
This is in most places a well marked layer averaging 2.7y in
thickness dorsally and 1.54 ventrally,’ but varying from 4, to
a mere line from point to point. Toward the edge of the body
it becomes very thin. Next to the epithelium the membrane
frequently shows a very distinct outline, on the inner side it is
less sharply differentiated from the parenchyma. It is evidently
differentiated chemically from both parenchyma and epithelium,
judging by the differerce in stain between it and these latter
tissues. In haematoxylin-eosin preparations, the latter are stained
light blue or gray, while the basement membrane is straw-col-
ored, being thereby very distinctly marked off from the other
tissues. In sections taken perpendicular to the surface, the
basement membrane appears nearly homogenous, but where the
sections are oblique or parallel to the surface a fibrillar struc-
ture is plainly visible. The course of the fibrillae, while more or
less irregular, is in general parallel to the surface and thus at
right angles to those of the epithelium. A continuity between
them and those of the parenchyma on the one hand, and the epi-
thelium on the other, I consider probable although [ am unable
to demonstrate it positively.

Jander (’97, p. 24) describes the origin of the basement mem-
brane as a “Verdickung der Netzstrange bis zu dem Maasse
die eine Basalmembran darstellt.”” The same
author (l.e¢., p. 27) deseribes the striations of the epithelium as
occasionally passing ‘‘durch die Basalmembran bis in die aitissere
Langsmusculatur hinein,” but qualifies this statement by adding
that ‘‘ist es auch nicht unmoglich irrtiimlicher Weise einen der-
artigen Zellplattenstreif in einen Bindegewebstrang zu verling-
ern.’ This view is supported by my own observations of the
fibrillar nature of this membrane and its probable continuity with

®’ Average of seven measurements varying from 4 to I4u.
5 Average of seven measurements.

258 R. T. YOUNG

the parenchyma, and by its replacement in Polychaerus caudatus
by a fibrous network directly continuous, through the sub-epi-
thelial muscle layers, with the parenchyma, as is the case in
Taenia serrata and its larva (Young, 08). Woodworth (91, p. 20)
on the contrary believes the basement membrane of Phagocata
gracilis to be a hypodermal product. In the fibrillar groundwork,
some deposit is probably formed, either by the epithelium or the
parenchyma, or by both, which intimately unites the fibrillae
into a homogenous mass, and is the cause of the differential
staining capacity of this membrane as described above. Nuclei
in the basement membrane, as described by Lang (84) are not
present here.7

Conditions in general similar to the above exist in several other
turbellarians studied by me (viz: Planaria maculata, Dendrocoe-
lum lacteum, Phagocata gracilis, Bothromesostoma personatum,
Mesostoma. tetragonum, Mesostoma sp. and Phaenocora(?).
In not all, however, are the nuclei as numerous as in Planocera
inquilina. While the abundance of nuclei depends in a large
measure on the condition of expansion or contraction of the worm,
still by comparison of several specimens of each species similarly
fixed, it is possible to construct a series with reference to nuclear
abundance leading from the last named form to those in which
nuclei are seldom or never found in the epithelium.

Such a form is Polychaerus caudatus (fig. 2) which presents
an advanced stage in the development of an insunken epithelium.
Beneath the cilia, which in preserved material appear to form
an almost continuous layer, is a loosely fibrillar, vacuolated
layer representing the epithelium, while a basement membrane
is not differentiated. The fibrillae form an irregular network,
showing no definite arrangement with reference to the surface,
and apparently in direct continuity with that of the under-lying
parenchyma on the one hand and with the bases of the cilia on
the other. I cannot speak positively regarding this however.
The vacuoles in the epithelium of this worm I believe are, largely

7 Regarding this membrane, Lang says however (1. c. p.64) “. . . . sieauf

vielen Praparaten ganz homogen aussicht, weil veile fiir die tibrigen Gewebe des
K6rpers treffliche Tinetionsmittel dieselbe d ffus fiirben.”’

EPITHELIUM OF TURBELLARIA 259

ABBREVIATIONS
b, basement membrane o, outer muscle layer
c, cilia p, wall of pharyngeal sack
e, epithelium pi, cavity of pharyngeal sack
ex, excretory duct s, cavity of seminal vesicle
7, inner muscle layer S1, remnant of epithelium of seminal
m, muscles of seminal vesicle vesicle

n, nucleus of insunken epithelium

All photomicrographs were made on Cramer isochromatic plates through a ray
filter of 2 per cent potassium bichromate, an are light being the illuminant
employed. The lens combination was a Bausch and Lomb 3g obj. and a Zeiss
No. 12 compens. oc., the camera being so adjusted as to give a magnification of
1000 diameters in each case. The positives were retouched at the microscope.

EXPLANATION OF FIGURES

1 Body wall of Planocera inquilina.

2 Body wall of Polychaerus caudatus.

3 Body wall of Bdelloura propinqua.

4 Wall of pharynx and pharyngeal sack of Planaria maculata.
5 Execretory duct of Dendrocoelum lacteum.

6 Seminal vesicle of Bothromesostoma personatum.

260

EPITHELIUM OF TURBELLARIA 261

at least, due to shrinkage, as suggested above for Planocera.
Occasional rhabdoids are scattered through the epithelium and
ciliary layer. The main point of interest, however, in the epi-
thelium of Polychaerus is the occasional occurrence of nuclei.
These are found mainly dorsally, but I have observed at least
one or two clear cases of their presence ventrally, one of which
is shown in the figure. There can, I think, be no question as to
the nuclear character of this and similar bodies in the epithelium.
Their structure is identical with that of the parenchyma nuclei.
They are round or oval in outline, average 4.7 by 3.5u in diam-
eter’ and contain a dense chromatic network surrounded by a
fairly definite membrane, which is however obviously incom-
plete in some cases. It is not impossible, however, that they
may owe their presence here to distortion of the tissue produced
by contraction of the worm during fixation. This would explain
their greater number in the dorsal epithelium where the tissues
(in my preparations) are much more distorted than they are
ventrally. The fact of their occurrence in the ventral epithelium,
however, where but little distortion has occurred and the further
fact of their occurrence in the epithelium of other Acoela asrecorded
by von Graff (91) renders it probable that their Secetoaa) pres-
ence here is normal.

In Bdelloura propinqua (fig. 3) there appears to be an entire
absence of nuclei in the epithelium examined. The surface of
this worm is covered by a layer of cilia similar in appearance to
that of Planocera already described. Below the cilia is the epi-
thelium which also, except for the absence of nuclei, has a struc-
ture similar to that of the last named species. Here is found
the same layer of basal swellings of the cilia at the points of their
insertion in the epithelium, the same fibrillar network with its
meshes approximately perpendicular to the surface, and con-
taining numerous vacuoles which are probably artefacts pro-
duced by shrinkage, as they are sometimes absent or but poorly
developed. Here too exists the same difficulty in determining
the precise relation between the epithelial fibrillae and the cilia

8’ Average of ten measurements.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

262 R. T. YOUNG

on the one hand and the basement membrane on the other. The
cilia are probably continuous with the epithelial fibrillae and pos-
sibly such a continuity also exists between the latter and those
of the basement membrane.

In vertical sections, it is impossible to distinguish cell bound-
aries in the epithelium, but in tangential ones occasional poly-
gonal areas may be seen which probably indicate its cellular
character. Treatment with silver nitrate, moreover, reveals the
cell boundaries very clearly.

Directly below the epithelium is a definite basement membrane
which, in most of my material, appears homogenous, but in one
specimen shows a very evident fibrillar character. The basement
membrane here is very lightly stained and appears as though
the ground substance had been in some way removed from the
meshes of the fibrillae, thereby rendering them apparent. These
form a loose network, the meshes of which run in general parallel
to the surface and stain similarly to those of the epithelium.
Externally, the membrane presents a scalloped appearance, which
may possibly be due to the insertion of, the epithelium, which
in fixation has shrunken a little away from the membrane,
thus bringing its scalloped appearance more plainly into view.
Whether there is any anastomosis between the fibrillae of the
basement membrane and those of either the epithelium or the
parenchyma is a point which I must leave undecided. That
this membrane is chemically different from the other tissues is a
point which is indicated by its differential stain. In sections
strongly counter-stained with eosin, the epithelium is red or
pinkish while the basement membrane varies from colorless or
light straw color to gray or pale brown.

The ciliary layer, epithelium and basement membrane show
wide variations in thickness, not only in different specimens, but
in different parts of the same specimen. These differences are
doubtless due in great measure to differences in amount of con-
traction of the tissues in fixation and to distortion produced
thereby; they may also be due in part to differences in the plane
of section; probably also to differences in development of the

EPITHELIUM OF TURBELLARIA 263

different layers. The following measurements® (in y) indicate
the variations referred to:

DORSAL VENTRAL
ilies: os sete oo ee oe eye oh & Bh Uo 2h eh,
Barthelimmy? {Az ce: 2... eee ee bem ai by (toy hates Ics Be ee es)
Basementemembranes.*..4.1)...54 ae 5, 4,11, 3, 3 Ay IIB a 3

At the edge of the body, where numerous glands open, the
basement membrane is reduced to a very narrow band.

While nuclei are typically present in the surface epithelium of
the Turbellaria, they are frequently either very rare in or entirely
absent from the pharynx and pharyngeal sack, and cell boundaries
ean only be demonstrated by special methods, the epithelium
thus assuming the appearance of a cuticula (fig. 4). The dis-
tribution of the cilia in this region is also very variable according
to accounts of various authors, these being in some cases present,
in others absent throughout the pharnyx; in some present on
its outer but not on its inner wall, while in others the reverse
holds true. In general, the epithelium of the gut forms a definite
one-celled layer, but in some cases (Béhmig, l.c. in Plagiostoma
bimaculatum, maculatum and sulphureum; Fuhrmann, ’98 in
Plagiostoma violaceum), etc., it appears to be intimately con-
nected with the surrounding parenchyma, while in the Acoela,
as is well known, the gut is replaced by parenchyma, the pharynx
opening directly into the latter. The oesophagus, the transition
from pharynx to gut, may have an epithelium of the ordinary
type, or an insunken one (Luther, ’04, ete.).

The statements regarding the epithelium of the excretory ducts
are again conflicting. In general an epitheliim appears to be
present, at least in the main ducts; but in some cases (Luther, l.c.),
the wall is not sharply differentiated from the surrounding paren-
chyma, while a non-nucleated terminal duct in Monoophorum
striatum is described by Bohmig (l.c.), so that the existence of a
typical epithelium in the excretory system of all Turbellaria

° These measurements are arranged in order from five worms, the first of each
set referring to one specimen, the second to another, etc.

264 R. T. YOUNG

cannot be considered as definitely established. This uncertainty
is probably due in part at least to the difficulty of demonstrating
the excretory system in preserved material. Its delicacy renders
its tracing, at least to the finer branches, very difficult in pre-
served material. So far as I have been able to find the excretory
ducts in my own sections, I have seen no evidence of a definite
epithelium. Their walls apparently consist of a collection of
protoplasmic strands in direct continuity with those of the sur-
rounding parenchyma, in which are imbedded occasional nuclei
(ige*5):

Conditions in.the reproductive ducts and glands show con-
siderable variation according to the statements of various authors.
In general, an epithelium is present, which is, however, very
variable in form. It may be insunken (penis of Byrsophlebs nana
von Graff, 03 and Geoplana pulla, von Graff, 91), and may lack
nuclei (penis of Yungia aurantiaca, Lang, l.c.), or both nuclei
and cell boundaries (Typhloplaninae, Luther, l.c.). The latter
author says (p. 98):

. . das Epithel des Atriums geht an der Spitze des Penis
in eine e Kernhaltige Plasmamasse liber, in der sich keine Zellgrenzen
nachweisen lassen . . . oft is sie (in the penis) nur noch
an den hier und da der Innenflache anliegenden platten Kernen zu
erkennen, in anderen Fallen gelingt es tiberhaupt nicht mehr ihr Vor-
handensein festzustellen.

In Plagiostomum reticulatum, von Graff (08, p. 2287) de-
scribes the epithelium of the ‘‘ausseren Penisrohrwandung”’ as
‘“‘vollends cuticulaihnlich” which ‘“‘fairbt sich nicht mehr und
erscheint vollkommen homogen: im Ductus ejaculatorius priisen-
tirt es sich als eine haarscharfe, stark roth tingirte Linie.”’

According to Luther (l.c.) in the Macrostomidae, an epithelium
is present in the antrum femininum alone, the gland ducts being
parenchyma spaces. ‘The same author finds in the Mesostomidae
and most of the Typhloplaninae an epithelium in the virgin bursa
copulatrix, which later degenerates, leaving the basement mem-
brane superficial; the latter then becoming strongly developed.
In the bursa stalk, however, the epithelium persists. The inner
wall of the bursa seminalis of Gyratrix hermaphroditus is lined by

EPITHELIUM OF TURBELLARIA 265

a plasma layer continuous with “die Vacuolisirte Ausfiillungsmasse

in welcher zerstreute Kerne neben Spermamassen
Bnechettst sind” (von Graff, ’05, p. 140). <A similar condition is
presented by the bursa copulatrix of Monoophorum durum (Fuhr-
mann, l.c.). Until more information is available regarding the ori-
gin of ovary, testis and yolk glands in Turbellaria, speculation
concerning the presence or absence of an epithelium in these or-
gans had better be postponed. At present, statements of different
authors are divergent, and in many cases not specific, on this point,
some, as Lang (l.c.), claiming its existence, while this is denied or
not mentioned by others.

In some places (pharyngeal sack, seminal vesicle), the wall
of the organ may be sufficiently distended to cause the disappear-
ance of the epithelium in places so that the underlying muscles
abut directly on the lumen (fig. 6). Similar conditions have been
described by Lang (1.c.) in the penis of Yungia aurantiaca, Luther
(l.c.), in the atrium copulatorium of Castrada segne, the bursa
copulatrix of Mesostomidae, etc.

We thus find in the Turbellaria a complete series of stages in
the modification of a normal epithelium to one with insunken
nuclei and cell boundaries.not readily demonstrable except by
special methods of technique,!® and similar transition stages
occasionally occur in the same species as pointed out by von Graff
(91, p. 6) in the following words:

. bei ihnen (Convoluta sordida and paradoxa) ein und
dieselbe Hirtungs- und Farbungsmethode (Himatoxylin z. B.) bald
zahlreiche Epithelkerne hervorhebt, bald gar keine oder doch nur

sehr wenige Kerne des Epithels tingirt, wenn auch in allen wbrigen
Theilen bei beiden Individuen die Tinktion eine gleich tadellose wire.

Von Graff (99) has called attention to this transition, point-
ing out the resemblance between the insunken epithelium of
Turbellaria and the ‘epithelium’ of trematodes and cestodes.
He says (l.c., p. 42):

Beriicksichtigt man dass in der urspriinglichsten Familie (Geoplani-

dae) . . . . ein normales Epithel angetroffen wird, so kann
man inden eingesenkten Epithelien ttberhaupt und speciell in dem der

+0 Not even then, in all cases, fide BOhmig (’90).

266 R. T. YOUNG

Kriechleiste nur einen secundiéren Charakter erkennen—einen Charakter,
der erst in den beiden am weitesten differenzirten Familien der Rhyn-
codemidae und Bipaliidae auftritt und bei letzterer seinen hochsten
Ausbildungsgrad erreicht hat. Hier ergreift der Process der Kinsenkung
bei manchen Formen, wie Plac. kewensis, das gesamte Kérperepithel
und man kénnte sagen dass diese Species und die ihr im Baue des Epi-
thels zunichststehenden Bipaliiden im Begriffe sind, die Epithelform der
Trematoden und Cestoden zu acquiriren.

My own observations emphasize this view of von Graff, which
has not yet received sufficient notice. I must, however, differ
from him in assigning an epithelium to cestodes and trematodes."
My own view is that we have progressing in the Turbellaria an
epithelial transformation leading to the condition in the former
groups in which the epithelium has been replaced by a cuticula.
I find this process occurring in ontogeny in the vagina and penis
of Taenia serrata, as I hope to explain more fully in a forthcoming
paper.”

A similar process has been described by Lénnberg (’91) in the
vagina of Abothrium rugosum and Tetrarhynchus tetrabothrius;
he has also pointed out the probable homology between the
turbellarian epithelium and the cestode cuticula.¥

The presence of nuclei in the cuticula of the primitive cestode
Amphilina (Salensky, ’74) and in several trematodes (Mono-
stomum mutabile—Braun, ’93; Distomum sp.—Macelaren, ’05;
Cotylogaster—Nickerson, ’02, etc.) suggests that these are forms
in which the outer layer has not yet been fully evolved into the
cuticula typical of these worms.

The many observations among both trematodes and cestodes
of the sloughing of the larval epithelium, this being later replaced
by a cuticle formed from underlying tissues (Leuckart, ’86,
Looss, ’92, 93, ’94, Pratt, ’98), ete., does not, I believe, detract
from the soundness of this view, because the homology of the
larval epithelium is by no means certain. Until more is known
concerning the germ layers of plathelminths, speculation on this
latter point is futile.

n See my discussion of this question elsewhere (Young (’08).

2 Here however there is apparently no insinking of nuclei into the paren-

chyma.
13 See also Ziegler (’05).

EPITHELIUM OF TURBELLARIA 267

The many variations of the turbellarian epithelium, not only
on the surface of the body but also in the digestive and reproduc-
tive apparatus, indicate the plasticity of this layer and the prob-
ability of the theory outlined above.

LITERATURE CITED

Béumia, L. 1890 Untersuchungen tiber rhabdocoele Turbellarien, II, Plagio-
stomina und Cylindrostomina, Graff. Zeit. wiss. Zool., Bd. 51, pp.
167-314.

Braun, M. 1893 Trematoda, Bronn’s Kl. und Ord. des Tierreichs, Bd. 4, 1 a.

FuurMANN, O. 1898 Neue Turbellarien der Bucht von Concarneau. Arch.
d’Anat. micros., Bd. 1, pp. 458-80.

GraFrr, L. von: 1891 Die Organisation der Turbellaria Acoela. Leipzig.

1899 Monographie der Turbellarien, IT, Tricladida terricola (Land-
planarien). Leipzig.

1904 Marine Turbellarien Orotavas und der Kiisten Europas, 1,
Hinleitung und Acoela. Zeit. wiss. Zool., Bd. 78, pp. 190-244.

1905 Idem, II, Rhabdocoela. Zeit. wiss. Zool., Bd. 88, pp. 68-150.

1908 Turbellaria, Bronn’s Klassen und Ordnungen des Tierreichs,
iByale Zi I Ge

JaNvER, R. 1897 Die Epithelverhiltnisse des Tricladen Pharynx. Zool. Jahrb.
(Anat. und Ont.) Bd. 10, pp. 157-204.

Lana, A. 1884 Die Polycladen (Seeplanarien) des Golfes von Neapel. Fauna
und Flora des Golfes von Neapei . . . . herausg. von der Zool.
Sta. in Neapel. Leipzig.

Leuckart, R. 1886 The parasitesof man . . . . Translated by William
E. Hoyle, Edinburgh and Philadelphia.

LonNBERG, E. 1891 Anatomische Studien iiber skandinavische Cestoden, I.
Kgl. Svenska Vetensk.-Akad. Handlingar, Bd. 24, no. 6.

1892 Idem, I]. Kgl. Svenska Vetensk.-Akad. Handlingar, Bd. 24,
no. 16.

Looss, A. 1892 Uber Amphistomum subclavatum und seine Entwicklung.
Festschrift Leuckart’s, pp. 147-67.

1893 Zur Frage nach der Natur des Kérperparenchyms bei den Tre-
matoden. Ber. k.siichs. Gesell. Wissenschaften (Math.-phys. Classe),
pp. 9-34.

1894 Die Distomen unserer Fische und Frésche. Biblioth. Zool., 16,
pp. 1-64.

268 R. T. YOUNG

Lutruer, A. 1904 Die Eumesostominen. Zeit. wiss. Zool., Bd. 77, pp. 1-273.

Mactaren, N. 1903 Uber die Haut der Trematoden. Zool. Anz., Bd. 26, pp.
516-24.

Nickerson, W.S. 1902 Cotylogaster occidentalis. Zool. Jahrb. (Systemat.),
Bd. 15, pp. 597-624.

Pratt, H. S. 1898 A Contribution to the life-history and anatomy of the
appendiculate distomes. Zool. Jahrb. (Anat. und Ont.), Bd. 11, pp.
351-88.

Sauensky, W. 1874 Uber den Bau und die Entwicklungsgeschichte der Am-
philina. Zeit. wiss. Zool., Bd. 24, pp. 291-342.

Woopwortn, W.M. 1891 Contributions to the morphology of the Turbellaria,
I, On the structure of Phagocata gracilis Leidy. Bull. Mus. Comp.
Zool. Harvard, vol. 20, no. 1.

Youne, R. T. 1908 The Histogenesis of Cysticercus pisiformis. Zool. Jahrb.
(Anat. und Ont.), Bd. 26, pp. 183-254.

ZiEGLER, H. HE. 1905 Das Ectoderm der Plathelminthen. Verh. deutsch. zool.
Gesellshaft, 15 Vers.

THE BILATERALITY OF THE PIGEON’S EGG

A STUDY IN EGG ORGANIZATION’ FROM THE FIRST GROWTH PERIOD
OF THE OOCYTE TO THE BEGINNING OF CLEAVAGE

Part [I

GEORGE W. BARTELMEZ
From the Department of Zoology, The University of Chicago

FORTY-SEVEN FIGURES

CONTENTS
ep TO LUGO MN era ie A a, Sea sedorale acas' Pace tReet a suSToR si aeiese) crate mye sera 270
THY eyieratell yore lian elovete Rie ee ASE eRe Rebel o.a-cis cuteretta Cio cheides alae cicad oro amend al
iB e prob lemameqerrcr hares sas ai ko ear RIES hates ara Vere ttener a ek tte oye anon PAS)
Visa th crommartermnas boris sonvse csseNinlacys a4 re x» abeeagem ea iaxe sto in a eite alee ie tena ai PHU
A. The primordial follicle. Oocytes less than 90u ................. 277
B. The second growth period. Oocytes to0.4mm................ .. 281
1. The lipoid spherules and the yolknucleus..................... 282
2. The origin of the stigma and the follicular blood supply........ 283
C. The period of differentiation. Oocytes to5.0mm................ 286
1. The eccentricity of the germinal vesicle...................... 286
Pe NELOO DIaSIITG ZOMER.) esc. hos Sayeeraee sos ya ite tentials cctnls: St: eve 287
3. The peripheral migration of the germinal vesicle......: a SP sects 287
4, The zona radiata and the rotation of the oocyte.............. 288
D: The final growth period. Oocytes to 20mm... . 55 45.0..5..054. 290
(eehevanntrals strum ulteeaaac «c, .cceeecoe neers ae eee eo ae mene es < clare 290
2. Correlations in the reproductive apparatus................... 291
3. The orientation of the follicle and ovulation................. 292
V. The bilaterality of the blastodise. (A summary of Part II)............ 296
A. The origin and development of the blastodise..................... 296
B. The blastodise during maturation and first cleavage.............. 298
plea Accra trrlipre te, rece ke 8 2 'iite: cine ni N Aime 2 Ae ec oak ote th Apeyeibeeh 299
Rielong ,ehalazal and embry One axes lo Me gin te & olka) s nie)e oy aes 300
Me MDISCUSSIONGATIC: SUNGMRAT Ys oc . sn. cagegen epee bees ex aos) ara eeso-ets wlwreyrs | 304
Aon Baleteralityciiy Verte DraAtesONe a... ct tear ison chelate 8 © averse yaya ets 304
Eee CERN (GE ER MLC KE. 5 Ws copes to oct. a een econ s SeR ove Ato lcln'e\e sheets = 307
Caer ENOL T SUOTAS AN 5 Reeds ots ls: Secaralteo igh ssi E RM ee emit Dis Dey Hare we ¥ wee CaN 309
Senn Onset Tyee ct Uk Shotts seis Gre we raid hd ats sialaaic taohSs Ussie eM ates ais ode wihlegs 310
269

270 sEORGE W. BARTELMEZ

I. INTRODUCTION

Whitman in his paper on the development of Clepsine (’78),
gave the first full and connected account of egg organization;
since then much evidence has accumulated to show that many
eggs are highly organized before cleavage begins and there are
cases in which the origin of this organization has been traced
back into the ovarian history of the egg. Thus it is well known
that the axis of bilaterality, one of the most fundamental mani-
festations of organization, appears, in the insect, while the egg is
still in the ovary and there is some evidence that the ovarian egg
in at least two primitive vertebrates has a bilateral structure.
These facts make the relations between the egg and the embryonic
axis of certain reptiles and birds very suggestive. The workof
Kolliker (’76), Duval (84) and others established the existence,
in the hen’s egg, of a fairly definite relation between the embry-
onic and chalazal axes and similar conditions were found in the
pigeon’s egg by Patterson (’09). This relation cannot have
arisen after the chalazae have been laid down in the oviduct;
I have met with only two considerations of the significance of
this fact, one in a paper written in 1893 which was found among
Professor Whitman’s unpublished manuscripts, the other is in
Lillie’s book on the chick (08). It has been found that in the
pigeon the position of the chalazae is determined before cleavage
begins and it follows accordingly that the pigeon egg is bilaterally
organized shortly after fertilization. EXvidence will here be pre-
sented to show that the antero-posterior axis of the pigeon not
only appears clearly at the time of fertilization but may also be
traced back into the ovarian history at least as far as the youngest
oocytes found in the adult ovary.

I am indebted to Professor Whitman for the opportunity of
working on this material and for many of the birds that were used.
It is difficult for a student of his to express the appreciation he
feels of the privilege of working under Professor Whitman. His
clear insight into fundamentals, his keen criticism and high ideals,
his whole personality made every conference with him an inspir-
ation. It is a pleasure also, to express my obligations to Prof. F.

BILATERALITY OF THE PIGEON’S EGG Dal

R. Lille for his constant interest, advice and criticism. I wish
to thank Prof. C. L. Bristol, Dr. J. T. Patterson and the members
of this Department for their suggestions and coéperation. Dr.
R. R. Bensley has shown me many courtesies in the course of the
work; his artist, Mr. A. B. Streedain copied figs. 42 and 43 from
the original drawings.

Il. MATERIAL AND METHODS

Harper (04), Blount (’09), and Patterson (’09) have discussed
in detail the many advantages the pigeon’s egg has for studies
on the early stages of development in the bird’s egg. The
technique employed is new for these eggs and will be discussed
briefly.

At the outset the necessity of studying the living material must
be emphasized. No conclusions can be based on the form-rela-
tions of the younger oocytes or of the blastodises that are derived
solely from preserved material. Certain distortions are unavoid-
able, even with the most careful treatment, and they must be
controlled by the study and measurement of the living cells. The
ereatest distortions are introduced in cutting thin paraffin sec-
tions, so these have been used only to add structural details and
for the illustrations. In the latter case they were used because
it is easy to photograph them; they are to be considered merely
as corroborative evidence as to the conditions observed in fresh
or creasote preparations and free hand sections. Entire ovaries
cleared in creasote, dissected and studied under the binocular as
solid objects afford an easier and more accurate means of inter-
preting form-relations than could possibly be obtained from the
study of sections, and this method, when properly controlled, is
to be highly recommended. In this three-dimensional study of
the oocytes the following technique was used: Fix the entire
ovary for not more than six hours in a mixture consisting of:

Saturated HgCl, in normal salt solution..............05......-: 94 cc.

LECT el asOLMGh VC LOM ONS oki: Ocy-\'d ipa ge GER AN es, alee ahs 6 ce.

Neutral formol (comm. formaldehyde solution neutralized with
WGI C! OE Ses Aer 9 + a Be ea RR yr pie oe OR! 6tol0ce.

272 GEORGE W. BARTELMEZ

The formol should be added at the time of using and the mixture
warmed to the body temperature. As has been said, this fixa-
tion only preserves the external form-relations of the oocytes.
Cytologically the microscopic picture is not true to life and, as
the photographs show, the nucleus is always more or less plasmo-
lyzed. After fixation and washing, the ovary is run up through
the alcohols to 95 per cent containing a trace of eosin, then crea-
sote is added very gradually during the course of a day or two and
the ovary finally studied in pure creasote. The most important
factor in avoiding distortions is the gradual clearing. The method
involves little shrinkage and the change in shape is within the
probable error in measurement and negligible as may be seen from
table 1.

The average shrinkage of twenty oocytes, ranging between 1
and 5 mm., was 4 per cent; in no instance was it. greater than 5
per cent. The shrinkage involved in paraffin imbedding and cut-
ting may be judged from the following: an oocyte of 2.5 mm. be-
fore fixation, measured 2.4 mm. in alcohol and 2.8 mm. in par-
affin. It was cut perpendicular to the long axis into 221 sections
10u thick. This means a total shrinkage of 11.6 per cent. The
effect of the compression in cutting may be judged from figure
27, a photograph of a median section of this egg. The two axes
shown were nearly equal before cutting.

To study the blood supply the ovary was injected with a freshly
prepared, 5 to 7 per cent solution of Higgins’ india ink in physio-
logical salt solution (see Evans, ’09). After light staining in
eosin it was cleared and studied in creasote.

For cytological control the recent mitochondria methods were
used. The most favorable fixative is Benda’s ‘modified Flem-
ming’:

One pernicent aqueousichromictacide..sseaaiee ee eee 15 ce.
WOsDeriGe Mu AQUCOUS OS na1CiaCid eerie ier ee ieereteicray evel iercia terre 4 ce.
GA CIAlRGOnIC IM CIC ..i.,..ic.schresiesy carvers ecole Ore MTEL Sy ><a ee ae 3 drops

This fixation gives a microscopic picture most nearly resembling
the appearance of the living oocytes under the immersion lens;
the protoplasm appears as a homogeneous ground substance in
which are imbedded granules of various sizes and: staining reac-

EGG

o

BILATERALITY OF THE PIGEONS

08 €X 06°EX 0G

GALOSVAUD NI

92x —9°SX
—9' TX —9 1x
Vex -Te@x—

maN

iP |

+9°% X 09 ZX —8G
09°T X OO TX +L°T
hie Se ee tears

06€X 06 EX GaP

TOHOOTYV
INGO Udd Ch NI

G9°ZX 99° SX —8'°S
09° 1X09 TX +2°T
02°§ X 02'S X 08's |

06°§X G6 EX Gao F

DNIHSVM UALAV

T WIdVL

OL'SX—L'GX
08°T |
0Z°€ X 06° X GE's |

06°& X 66° X GBF |

Hall NI |
@ALAOOO AO SNOISNAWIG

£08°Z | ( (anoy JT) (quad ted g) Jou

(gued aod 7) o1J00V
o}BUITTGNG

(yued ted QT) JOULIOW

(quod 10d QT} Ie0V
opeurrTqus

GATLV XI

274 GEORGE W. BARTELMEZ

tions. In the study of the blastodiscs however, the important
objects are to preserve the form relations and to obtain sharp
differentiations in the cytoplasm. The sublimate-acetic mixture
above mentioned is best for this purpose. The optimum time of
fixation is between forty-five minutes and one and a half hours.
The whole yolk is dropped into the warm mixture in such a way
that the blastodise floats down; it remains in this position owing
to the weight of the glass pin inserted to mark the orientation.
When the blastodise is to be photographed enough 95 per cent
alcohol is added so that the egg just rests on the bottom of the
dish. After fixation the egg is washed in repeated changes of 35
per cent and 50 per cent alcohol and hardened in 70 per cent over-
night. The next day a pentagonal block is cut out as Blount
(09) and Patterson (’09) direct. Material for sectioning is de-
hydrated in. 95 per cent alcohol, gradually cleared in bergamot oil,
and imbedded in 55 to 58° paraffin. Most sections were cut 62
micra thick and mounted with albumen fixative.

After sublimate fixation the best stain is Bensley’s (’11) neutral
gentian. The copper-chrom-hematoxylin method also gave good
results. Intra vitam staining with Janus green and neutral red
aided materially in studying the fresh primordial follicles under
the 3 mm. immersion lens. The blastodises of mature and fer-
tilized ova were studied and usually drawn in the life. They were
kept in the Patterson stage incubator (’09) in a mixture of physi-
ological salt solution and egg albumen, and observed with a
Zeiss binocular. Continuous observations could be made by this
means of stages between fertilization and the completion of the
first cleavage if proper precautions were taken to prevent chill-
ing the egg. For the first three hours and probably for much
longer the development is perfectly normal.

In studying the yolk spherules of the younger oocytes the ovary
was fixed for twenty-four hours in warm neutral formol (10 to
15 per cent) or in a 15 per cent solution of formol in 2.5 per cent
aqueous potassium bichromate. After washing, free hand or
frozen sections were made, stained in a saturated solution of
Sudan m1 in 70-per cent alcohol, counterstained with a weak alco-
holic solution of methyl green, and mounted in glycerine.

BILATERALITY OF THE PIGEON’S EGG 275

II. THE PROBLEM

Attention has already been called to the relation between the
chalazal and embryonic axes in the eggs of the hen and pigeon
and the matter is considered in detail in section VI. The general
character of the relation in the pigeon is shown in diagram I.

Diagram I A polar view of an incubated pigeon egg (‘yolk’) showing the rela-
tions between the embryonic and long (chalazal) axes. The end of the egg which
was directed toward the blunt end of the shell, where the air chamber is always
found, is toward the left. The end which passed down the oviduct first, with the
‘cloacal’ chalaza (c.c.) attached, is to the right and the head of the embryo is
away from the observer; c.c., cloacal chalaza.

The essential feature is that when the end of the egg opposite
the air chamber of the shell is held to the right, the head of the
embryo is away from the observer. The chalaza at this end is
frequently heavier, double, and more firmly attached than the
other; it is in fact the chalaza which is formed first, since the end
of the egg to which it is attached goes down the oviduct first.
This chalaza may accordingly be called the ‘cloacal chalaza’
(c.c.) for its end of the egg, in the oviduct and in the ovary as well,
is nearest the cloaca; the other will be referred to as the ‘infundib-
ular ehalaza.’ It is evident that one end of the chalazal axis is
different from the other with reference to the embryo, that is,
the chalazal axis is definitely related to the embryonic axis of
bilateral symmetry. Furthermore the chalazal axis is the longest
axis of the egg. This fact has not been recognized hitherto

276 GEORGE W. BARTELMEZ

although itis almost as apparent in the hen’segg as in the pigeon’s.!
The chalazal, better the ‘long’ axis, marks the axis of bilaterality
of the ovum as a whole; i.e., but one plane, that passing through
the long and polar axes divides it symmetrically. The evidence
for this is not only that one end of the long axis is definitely re-
lated to the embryo, but also that one end is morphologically dif-
ferent from the other as is shown below, p. 289. There are then,
two axes of symmetry in the incubated egg, one of the embryo,
the other of the mass of food yolk, which, though they do not
coincide, are definitely related to one another. Since the chalazae
are laid down very soon after the egg enters the oviduct, this rela-
tion must also exist at the time of fertilization.

Evidence will be presented to establish the following theses:

1. The bilateral symmetry of the ovum as a whole, manifested
by the long axis of oviducal and laid eggs, is present in ovarian
eggs at all stages of development from the primordial follicle on.
The long axis of ovarian eggs is the same as that of oviducal
eggs (p. 294).

2. The antero-posterior axis of the embryo is predetermined in
the ovary because the axis of symmetry of the blastodise of the
ovarian egg bears the same relation to the long axis of the entire
ovum as does the embryonic axis in the fertilized egg and in the
subsequent stages. |

It remains to be seen how much farther back in the life history
morphological evidences of bilaterality can be found. The evi-
dence at hand points toward the view that bilaterality as well as
polarity are inherent characters of the protoplasm and persist
from generation to generation.

1 Patterson’s diagram (’09, p. 68) shows that he observed the long axis in the
pigeon and Sonnenbrodt (’08) and Riddle give two dimensions in their measure-
ments of oocytes in the hen but neither author attributed any significance to the
matter at thetime. The relations do not seem to be so constant in the hen’s egg
as in the pigeon’s. Thus, in a series of over one hundred hen’s eggs it was found
that in almost one-third of the cases the head of the embryo was directed toward
the observer when the pointed end of the egg was held to the right. This matter
deserves further study, for it involves the question as to whether the end of the
hen’s egg that is to pass down the oviduct first is predetermined in the ovary as it
is in the pigeon.

BILATERALITY OF THE PIGEON’S EGG Da Er

It should be borne in mind in this discussion that the facts are
presented in the order opposite to that in which the analysis was
made and so apparently trivial differences in the youngest oocytes
fit into a general scheme of development and become quite appre-
ciable during the cell’s linear growth of 700 diameters.

IV. THE OVARIAN HISTORY

Most workers on the bird’s ovary have divided the ovarian
development into more or less clearly defined periods, usually on
the basis of the nuclear phenomena alone. The best of these
analyses are those of D’ Hollander (’04) for the embryonic stages
and of Sonnenbrodt (08) for the subsequent growth periods. The
classification that has been made here is based upon the phases of
physiological activity of the oocyte.as a whole, i.e., the four periods
described are distinguished by the different forms in which the
cell organization is manifested.

The primordial follicles, the youngest oocytes found in the
adult ovary, form the starting point in this discussion. They are
in practically the same stage of development as the oldest oocytes
of a chick two weeks after hatching or a pigeon several weeks
older, and they may be considered as the final stages of the first
period of growth after the differentiation of the oocyte as such.
This period, in the case of the primordial follicles, involves a slow
increase in size from 10 to 804 and the accumulation of deutoplasm
in the form of lipoid spherules. It corresponds to Sonnenbrodt’s
period V, in which the chromosomes stain deeply and have a
characteristic thickened form.

A. The primordial follicle: (oocytes to 0.09 mm.)

The form relations of the primordial follicles can be studied
satisfactorily only in the fresh condition and the following de-
scription is based primarily on the study of material dissected out
of the living ovary in salt solution, usually stained intra vitam and
examined with a 3 mm. Zeiss apochromat immersion lens. The
cover slip was so supported that the oocytes were not subjected

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

278 GEORGE W. BARTELMEZ

to any pressure. Under these conditions the following relations
may be observed: The polar axis is marked by the peripheral
position of the germinal vesicle; another axis can invariably be
distinguished, namely, one perpendicular to the plane of the polar
axis. This axis is also approximately parallel to the surface of
the ovary and it is distinetly greater than any other axis (fig. 1).
It will be referred to as the ‘long axis’ and its essential relation is
that the polar axis lies in a plane perpendicular to it. It is found,
not infrequently, that the oocyte is so oriented that the polar
axis is perpendicular to the surface of the ovary as well as to the
long axis and that the animal pole is also the attached pole, the
vegetative, the free pole (diagram II A). This relation is neither
constant nor essential, but there are many reasons for believing
that it represents a typical condition.

Under the high powers (2 mm. apo.imm.) the cytoplasm appears
as a suspension of various kinds of granules in a homogeneous
ground substance. With the aid of vital stains, three kinds of
granules can be distinguished. The most obvious are the large
deutoplasmic spherules which form a cap between the germinal
vesicle and the vegetative pole; this may be called the ‘spherule
cap’ (figs. 1, 9, 11, 18, sc.). The periphery of the oocyte is free
from these spherules. They are highly refractive in fresh mate-
rial and appear yellow by reflected light. They stain intensely
with Sudan mr and are lipoid in character. At the center of the
spherule cap, lying close to the germinal vesicle is a finely gran-

Reference letters

bld., blastodise N.A., nuclear axis

c.c., cloacal chalaza o.p., posterior margin of outer periblast
c.p., central protoplasm P.A., polar axis

c.st., cloacal end of stigma p.p., peripheral protoplasm

E.A., embryonic axis s.c., Spherule crescent

g.v., germinal vesicle st., stalk of follicle

L.A., long axis s.ov., surface of ovary

Lat., \atebra s.z., Spherule zone

y.n., yolk nucleus
------------ L.A., long axis
wer. eyc ete ee. P.A., polar axis
gh cise eel V.A. and #.A., nuclear and embryonic axes

BILATERALITY OF THE PIGEON’S EGG 279

weirs ae

S.C.
va

Diagram II Showing the development of the oocyte, and axial relations. A,
Primordial follicle of 50u in side view showing the relations of long and polar axes;
B, the same in polar view, showing the relation of the long (ooplasmic) axis to
the nuclear axis; and the relation of the germinal vesicle to the long axis; C, Oocyte
of 0.6 mm. Beginning of period of differentiation. Side view. The germinal
vesicle is near the center of the oocyte but nearest the animal pole and nearer one
end of the long axis than the other. The spherule zone lies between the central
and peripheral protoplasm. D, Oocyte of 3.0mm. End of period of differen-
tiation; side view. The germinal vesicle is at the periphery; the latebra has
appeared and is eccentric in position. The zone of peripheral protoplasm has
become narrow. H,mature ovarianegg of 20mm. Endof the final growth period.
Side view showing the long and polar axes and the eccentricity of the latebra.
F, Oviducal egg in polar view showing the relation between long axis and the
embryonic axis as indicated by the shorter axis of the blastodisc.

280 GEORGE W. BARTELMEZ

ular region (y.n.) which stains intensely with neutral red, intra
vitam; this is the yolk nucleus, part of which at least, is derived
from the germinal vesicle and may be considered as cytoplasmic
chromatin or ‘chromidial substance.’ Spherule cap and yolk
nucleus, by their position, emphasize the polar axis, which is
defined by the eccentricity of the germinal vesicle (fig. 1 and its
legend.) The mitochondria form the third kind of granulation
and they alone stain with the Janus green. They differ markedly
in shape and size from the mitochondria of the follicle cells and of
all other cells in the ovary, in being smaller and apparently spheri-
calin shape. Most of the workers on the bird’s ovary have either
noted or figured the structures that have been mentioned, except-
ing the mitochondria which are dissolved by all the ordinary fixing
fluids. No one however has discussed the question of polarity,
doubtless because it seems to be indeterminate (see p. 288).

Two authors have figured primordial follicles in the fresh con-
dition; Waldeyer (’70) and Coste (’47) in his atlas (second
‘Poule’ plate, fig. 3). This figure is somewhat diagrammatic but
it illustrates a constant feature of some interest; viz., the position
of the germinal vesicle, nearer one end of the long axis than the
other so that the polar axis, instead of bisecting the long axis,
cuts it nearer to one end (fig. 1). This means that morphologi-
cally the primordial follicle is bilaterally symmetrical; one end of
the long axis is different from the other with reference to the ger-
minal vesicle. This long axis is the same as the chalazal axis (p.
294) and it will be remembered that one end of the latter is dif-
ferent from the other with reference to the embryo. The extent
of this second eccentricity of the germinal vesicle varies somewhat
in different ovaries as may be seen in figure 1 and the photo-
graphs, but it is almost invariably distinct in direct side views
and polar views.

There is another feature of the primordial follicles which is clear
in many ovaries though not in all, but which is very suggestive.
The large germinal vesicle is not spherical but elliptical and is so in-
clined to the long axis of the oocyte that the polar view appears as
is shown in figure 1A and diagram IIB. It is obvious that this
nucleo-cytoplasmic relation is the same as that between the em-

BILATERALITY OF THE PIGEON’S EGG 281

bryo and the whole ovum in the incubated egg (diagram II B
and F). It is difficult to convince one’s self that this relation is
of general occurrence, for there are many technical difficulties
involved. No confidence can be placed in preserved material and
it is not easy to study the fresh oocytes from all angles under the
immersion lens; the difference between the two diameters of the
germinal vesicle is rarely more than one micron, and it is not
apparent except in direct polar views. Still it was possible to
observe some oocytes in this way, and in one ovary, where the
conditions were especially favorable, it appeared that the angle
between the shorter axis of the germinal vesicle and the long axis
of the oocyte was relatively constant. This matter is discussed in
Section VI, page 299, where it is shown that the same holds true
for the angle between the embryonic and chalazal axes. The
long axis is definitely related to the embryonic axis of symmetry
and it has been shown that it marks the axis of bilaterality of the
primordial follicle. If we look upon the elliptical shape of the
germinal vesicle as an expression of bilaterality also, we have a
basis here in this nucleo-ooplasmic relation of the remarkable
relation between the embryo and the ovum as a whole in the incu-
bated egg.

Tosumup: Two axes of symmetry can be distinguished in the
primordial follicle; (1) The polar axis, which is marked by the
eccentric germinal vesicle, the yolk nucleus and the spherule cap.
(2) The axis of bilaterality defined by the long axis of the oocyte
with the germinal vesicle nearer one end of it. The relation of
this bilaterality of the oocyte as a whole to the bilaterality of the
embryo is traced in the following sections. (3) There is some
evidence to justify the interpretation that at this stage the ger-
minal vesicle has an axis of bilateral symmetry which bears the
same relation to the long axis of the ovum as later the embryonic
axis bears to this same long axis.

B. The second growth period: (0.09 to 0.4 mm.)

Sonnenbrodt (’08) has emphasized the fact that the primordial
follicles of the adult ovary are in a quiescent state; indeed the

282 GEORGE W. BARTELMEZ

only evidences of activity between hatching and maturity are the
growth to a maximum of 0.09 mm., and the accumulation of the
deutoplasm of the spherule cap. When a primordial follicle
begins to grow, however, striking changes occur in the germinal
vesicle and the ooplasm. ‘The chromosomes lose their thickened
form, become more finely granular and longer, so that they stain
more lightly, and many nucleoli are formed apparently at their
expense (figs. 10, 13, 14, 15, 16, 18). In the ooplasm more spher-
ules are laid down so that the peripheral zone becomes narrower,
(figs. 13 and 17), and there is a marked increase in the yolk nucleus
material (chromidial substance?), due largely, as D’ Hollander (04)
and others have contended, to a transfusion of chromatin through
the nuclear membrane; the striking observations of Munson (’04)
are of especial interest-in this connection. Figure 14 shows the
increase of the yolk nucleus in an oocyte at the beginning of the
second growth period. Figure 17 shows the yolk spherules at
this stage; it was taken from a free hand section stained with
Sudan m1. The red stained spherules (black in the photograph)
are present throughout the ooplasm except in the narrow periph-
eral zone and in the region occupied by the yolk nucleus which has
increased greatly in amount and is beginning to spread. Figure
15 shows this spreading of the yolk nucleus material clearly.
The irregular blocks and bands that may extend into any part of
the ooplasm are the ‘Balken’ of Holl (90). Figures 16 and 18
show later stages of this process, only one of the numerous ‘ Bal-
ken’ appearing in the sections. In both, the individual basophile
granules may be séen scattering in all directions; eventually they
become evenly distributed thrgugh the ooplasm. Occasionally
the original center of the yolk nucleus may persist in later stages
(fig. 20), but usually the ooplasm now comes to appear homoge-
neous in paraffin sections. The study of material stained with
Sudan shows, however, that yolk spherules are still being laid
down at the periphery of oocytes of 0.3 to0.4 mm. Figure 19
shows an oocyte of 0.8 mm. in which the spherules are confined
for the most part to a zone, sc, just inside the peripheral pro-
toplasm. This clearing of the central part of the oocyte, defining
the central and peripheral protoplasm with the yolk zone between

BILATERALITY OF THE PIGEON’S EGG 283

them, marks the beginning of the period of differentiation (see
p. 286).

2. The origin of the stigma and the follicular blood supply. The
characteristic ‘stigma’ of the ovarian follicles in the bird has long
been known, but so far as I am aware nothing has been published
concerning its origin, relations or its exact réle in ovulation. Since
the long axis of the oocyte has not been understood either, no one
has noted that the stigma is almost invariably in the long axis.

Mechanical factors seem to play an important part in the dif-
ferentiation of the stigma and of the closely correlated blood supply
of the follicle; at the same time both are definitely related to the
bilateral organization of the oocyte.

The stigma does not necessarily arise at a definite stage in the
development of the oocyte; it is present in some primordial fol-
licles of 50 « and occasionally follicles of 500 » are found imbedded
in the stroma which show no trace of it. The reason for this is
that the essential feature of stigma formation is the intimate
association of the follicular epithelium of the oocyte with the
peritoneal (‘germinal’) epithelium, and this association is con-
ditioned by the oocyte’s reaching the surface of the ovary. How
early in the embryonic history this association of epithelia may
take place has yet to be studied, but it is known that neither
oocytes nor follicle cells develop directly from the superficial
layer of the ovary, ‘les cellules indiférentes superficielles’ of D’ Hol-
lander ('04), so it can hardly come about until the follicle is formed
and the oocyte begins to grow. It has already been said that the
oocytes are so oriented in the stroma that the long axis is approx-
imately parallel to the surface of the ovary. Accordingly, when
the follicle reaches the surface, the fusion first occurs in the great
circle of the long axis and an elongate area of fusion is soon
formed, the longer diameter of which coincides with the long axis
of the oocyte. This appears in figures 3 and 4 where the region
free from capillaries indicates roughly the area of stigma fusion.
Coincident with the apposition of the two epithelia there is a
flattening out of the germinal epithelium. A correlation is estab-
lished here, between the long axis, itself a manifestation of the
egg organization and the other elements of the ovarian follicle,

284 GEORGE W.-BARTELMEZ

to form the stigma, which, as will be seen, plays an important
part in the orientation of the ovum in the oviduct. Figure 11,
taken from a median longitudinal section of an oocyte 83, in
long axis, shows the close approximation of the follicle and epi-
thelial cells; in the whole stigmal area of this oocyte only a few
isloated stroma cells were to be found between the two epithelia.
The study of many ovaries shows that the stigma is formed accur-
ately in the long axis of the oocyte; the Y-shaped and branched
stigmas found in many of the follicles of certain ovaries have
their main axis in the long axis of the oocyte and actual devia-
tions from this condition are rare.

In the majority of cases the oocytes come into contact with the _
germinal epithelium during the second growth period, (0.15 to
-0.4 mm.) and the stigma is formed at the free pole. The latter
frequently coincides with the vegetative pole of the oocyte, as
in all of the figures, but, as has been said, this is not essential for
normal development; the polar axis may lie anywhere in a plane
perpendicular to the long axis and so be variously related to the
surface of the ovary. Nevertheless the condition indicated in
diagram 2 is the.typical one, for it is the one found normally in
mature oocytes whatever may have been the condition in the
earlier stages. The factors which bring about the ‘typical’ con-
ditions in all oocytes that mature will be discussed in the follow-
ing section (p. 288).

2. The blood supply of the follicle. The development of a
distinct blood supply, i.e., the differentiation of a theca vascu- .
_losa usually begins soon after the formation of the stigma. The
capillaries in the cortex of the ovary course about irregularly
between the primordial follicles, as may be seen in figure 2. As
the oocyte grows, it projects farther and farther from the surface
and owing to the outgrowth and anastomosis of fine branches
from the pre-existing vessels, the network becomes finer over the
bulging surface (figs. 3 and 4). At the same time there is a
proliferation of stroma tissue over the whole surface, except in the
region of the stigma, where such growth is inhibited as appears in
figure 15, which shows a cross section of a stigma (st.) Two layers
can now be distinguished in the connective tissue follicle: a very

BILATERALITY OF THE PIGEON’S EGG 285

delicate continuous layer, the theca folliculi, closely applied to the
follicular epithelium, and an‘outer loose theca vasculosa which
develops hand in hand with the capillary plexus. The theca
vasculosa is absent, naturally, in the region of stigma fusion, for
here the delicate outgrowths of the capillaries do not penetrate.
Figures 5, 7, and 34 show how sharply the stigma is defined in
injected material.

The character and development of the blood vessels in the
theea vasculosa is illustrated in figures 3 to 7 and 31 to 34, all of
which were taken from whole oocytes cleared in creasote. During
the second growth period the vascular network comes to invest
_ the entire oocyte except the stigma which extends along the free
pole and up toward the attached pole at either end of the long
axis, thus dividing the vascular theca into halves which are
symmetrically placed with reference to the long axis (figs. 4 A
and B, and 5.) So it comes about that the stigma, which dif-
ferentiated with reference to the long axis, in turn plays a part
in determining the bilateral arrangement of the blood supply.
During the succeeding period of differentiation the bilaterality -
of the blood supply of the follicle is accentuated by the appearance
of a median artery and vein on either side, which lie, roughly in
the plane of the polar axis, i.e., perpendicular to the long axis
(figs. 6 and 7, and the intermediate stages, figs. 31 to 34). The
arrangement of the blood vessels, illustrated in figures 7 and 33,
is typical of that found in the subsequent stages, although an outer
set of vessels, which develops during the final growth period,
makes the bilateral arrangement less obvious, as the larger oocyte
in figure 37 shows.

The bilaterality of the blood supply is of interest because it
plays a part in the preservation of the bilaterality of the ovum
as a Whole during the final growth period, when the living ooplasm
is confined to the region of the blastodise and to a delicate film
of protoplasm over the periphery of the oocyte.

To summarize: The long axis appears as an expression of the
bilateral organization of the oocyte. It determines the position
of the stigma and together they determine the bilaterality of
the blood supply which helps to insure a bilateral deposition of

286 GEORGE W. BARTELMEZ

food yolk. Thus one of the main features of the bilateral struc-
ture of the mature ovum is determined. The position of the
germinal vesicle nearer to one end of the long axis determines the
other, as will be shown below.

C. The period of differentiation: (0.4 to 5 mm.)

As the oocyte grows from three to four-tenths of a millimeter
and the yolk spherules are laid down only at the periphery, the
germinal vesicle remains relatively stationary in position, so that
it comes to lie nearer and nearer to the center of the cell (com-
pare figures 14 to 17, 19 and 20). During the ensuing stages this
process is continued as figures 21 to 23 show. Does it ever become
quite central in position? This is a matter of theoretical im-
portance and, since the evidence from paraffin sections cannot be
relied upon, it has been decided by a study of creasote prepara-
tions of entire oocytes and by the use of thick free hand sections.
Figure 6 is from an oocyte drawn in creasote and in this, as in all
oocytes studied, the germinal vesicle was nearer one pole, the ani-
malpole. This is the only stage at which there could be any ques-
tion as to the eccentricity of the germinal vesicle and the fact that
it is always nearer one pole, defining the polar axis perpendic-
ular to the long axis, completes the evidence that we are dealing
throughout development with the same polar axis. If the ger-
minal vesicle migrated about in response to external forces it
should be possible to find cases in which it is quite central. Such
cases are not found and this agrees with the statements of all
recent workers on the bird and reptile ovary.

The appearance of the oocyte during the earlier stages of the
period of differentiation is shown in figures 21 to 23. The first,
from an oocyte of 0.51 mm. stained with Sudan III, illustrates
the position of the sperule zone (sz) near the periphery. The
next period of growth does not involve the deposition of yolk
spherules and so the peripheral protoplasm becomes wider. Some-
what later, in oocytes of about one millimeter, another zone of
spherules is laid down and this is the first evidence of the periodic-
ity in yolk formation which characterizes the final growth period
(Riddle, ’11). In oocytes from 0.5 to 0.8 mm. (figs. 22 and 23),

BILATERALITY OF THE PIGEON’S EGG 287

two ooplasmic zones may be more or less clearly distinguished,
the central and peripheral protoplasm of authors. At this time
when the germinal vesicle is still near the center of the oocyte,
the first definitive yolk granules (white yolk) appear, as the photo-
graphs show.

Now as the oocyte continues to grow, the germinal vesicle
begins to migrate peripherally along the polar axis to the animal
pole (figs. 24 and 25). This is no wandering of the germinal
vesicle to the best nourished region of the cell as Sonnebrodt
seems inclined to think; the path to the periphery is predeter-
mined, for the polar axis is never changed. The migration begins
in oocytes of about 0.9 mm. and the germinal vesicle is usually
quite peripheral when they have reached a diameter of 1.5 mm.
Shortly after the beginning of the migration very fine yolk sphe-
rules appear in the central protoplasm so that they are again
present in all parts of the ooplasm except in the peripheral pro-
toplasm.

During this period characteristic changes occur within the
nucleus which have been best described by Sonnenbrodt (’08)
for the hen. So far as my examination goes, I can confirm his
descriptions for the pigeon in most particulars. It may not be
out of place to say that in every oocyte I have studied it was
possible to demonstrate the chromosomes or their morphological
equivalents.

At any given time one finds in an ovary about twelve, rarely
more, oocytes from 2 to 5.5 mm. in diameter. The growth at
this stage is due chiefly to an increase of the fluid content and a
periodic deposition of yolk granules in the central protoplasm,
while the peripheral protoplasm grows thinner and thinner (com-
pare figures 27 and 8 from oocytes of 2.5 and 3.2 mm. respectively).
In figures 26 and 27 the path of the germinal vesicle to the periphery
is marked by a trail of fine reticulum and in the latter the region
that surrounded the_germinal vesicle during its stay near the
center of the oocyte is clearly defined by the smaller yolk granules.
This whole flask-shaped area (lat) is the anlage of the latebra.
Its position was determined by that of the germinal vesicle and
accordingly it is near the animal pole and nearer one end of the

288 GEORGE W. BARTELMEZ

long axis. This eccentric position of the latebra plays an impor-
tant part in the orientation of the ovum in the oviduct as will
appear below (p. 292).

4. The zone radiata and the rotation of the oocyte. One of the
reasons why the polarity of the bird’s egg has not been understood
hitherto, is that in the period now under discussion all possible
relations are to be found between the follicle with its long axis
and stigma, and the polar axis of the oocyte. In the preceding
and the subsequent stages the relations are constant. Thus in
the former practically every follicle shows the polar axis perpen-
dicular to the long axis and many have the animal pole coinciding
with the attached pole (diagram II, p. 279). During the final
growth period also, the conditions shown in diagram 2, F are
found almost invariably; i.e., the long axis of the oocyte is identi-
cal with the stigmal axis of the follicle and both are perpendicular
to the polar axis, the animal pole being at the attached pole of
the follicle. How does this come about? It might be supposed
that only those follicles in which the ‘typical’ conditions exist,
enter upon the final growth period, but this is not so; for the
final growth of the oocyte is not dependant upon the relation
of the oocyte to its follicle. The facts are these: Oocytes less
than a millimeter in diameter usually show the ‘typical’ relations.
After the germinal vesicle has begun to migrate peripherally all
possible relations may be found between the oocyte and its fol-
licle, (fig. 7 shows an extreme case). During the final growth
period the ‘typical’ relations are again found, this time almost
invariably. The explanation was first suggested by some experi-
ments which showed that, in the later stages, the oocyte as a
whole is free to rotate within its follicle. A bird with growing
oocytes was tied down on its back for twenty-four hours; when
the abdominal cavity was opened it was found that the germinal
vesicles occupied the highest points in the follicles viz: the free
poles, and yet the structure of the eggs was normal. The polar
axis is therefore not influenced by gravity, as some authors have
implied; the oocyte as a whole simply orients itself with reference
to gravity. In applying these data to the stages in question the
following facts must be borne in mind: The oocytes are lying in

BILATERALITY OF THE PIGEON’S EGG _ 289

the ovary in all possible relations to the direction of the force ot
gravity and as the germinal vesicle migrates peripherally many of
them are subjected to an increasing strain by the force of gravity,
since the cell is growing larger and the animal pole is constantly
growing specifically lighter than the vegetative pole. If the strain
in these oocytes were to be released by the formation ofa lymph
space, so that they could rotate within their follicles, they would
naturally swing into various positions according to their locations
in space (fig. 7). At this very time a structure arises which may
be interpreted as such a lymph space; it is the zona radiata which
is formed apparently from the follicular epithelium. No method
of showing that the rotation takes place between the follicular
epithelium and the vitelline membrane was found: the evidence
rests solely upon the fact that rotation first appears after the zona
radiata has been formed, and upon the absence of any other
structural condition that would permit rotation. The striations
of the zona would, according to this interpretation, be intercell-
ular bridges extending from the follicle cells and they would offer
no obstacle to the freedom of rotation.

This freedom of rotation of the oocyte within its follicle is a
matter of importance, for whatever may have been the axial
relations in the earlier stages, those shown as typical in diagram 2
are eventually established through its agency. As soon as the
final growth period is initiated and yolk is accumulated, the fol-
licle becomes large enough to hang down freely into the body
cavity; then its highest point is the middle of the attached pole
where the blood vessels enter. The oocyte now orients itself
along lines of least resistance, the animal pole, i.e., the germinal
vesicle with the anlage of the blastodisc come to lie at the attached
pole of the follicle and the long axis of oocyte and follicle becomes
perpendicular to the polar axis. The laying down of yolk with
reference to the polar axis is controlled by the ooplasm, but the
follicle, in large measure, determines the long axis, though the
peripheral protoplasm may also play a directive role. This orien-
tation of the oocyte makes it possible for the original long axis
which determined the follicular long axis, to persist during this
period when the great bulk of the oocyte is made up of inert

290 GEORGE W. BARTELMEZ

yolk, in the case of those oocytes in which such conditions as are
shown in figure 7 existed during the later stages of the period of
differentiation. There are many oocytes in which the ‘typical’
conditions hold throughout the entire development.

D. The final rapid growth period

The ovary of an adult unmated pigeon has normally from six
to twelve follicles in the terminal stages of the period of different-
iation, from 3 to 5.5 mm. in long axis; never any larger ones.
That is to say, the process of rapid yolk secretion which charac-
terizes the final growth period is initiated by the stimulus of mating.
The same holds true for birds that have reared young; the stim-
ulus is received when sexual activity is resumed after the young
leave the nest. The only theory that will account for all the
observed facts is that the initial stimulus is psychic in character;
the bird may be mated with another female, with a bird in another
cage, or even with her caretaker, and yet lay eggs. This state-
ment is based on Professor Whitman’s extended observations in
breeding pigeons, and, needless to say, it is easy to tell when a
bird is mated from her behavior. Harper (04, p. 4 ss.)?

Corresponding to the wide experience that pigeons never lay
more than two eggs at a sitting, one usually finds in an active
ovary two follicles, differing slightly in size, which are distinctly
larger than the rest. When the stimulus is received these two
begin to grow, one always keeping a few millimeters larger than
the other. Occasionally, (in 6.2 per cent of a total of 261 cases),
it happens that three follicles mature at the same time, one of
them larger than the other two; a condition which is considered
on p. 294; it should be noted that an arrangement of follicles in
sets in the hen has been observed by Patterson (’10, p. 105).
In the course of about eight days the oocyte grows from 5 to 20
mm. in long axis, the definitive size being relatively constant for
the eggs of a given bird. The average is about 20 mm., long
axis, 18 mm., polar axis, and 18.5 mm. for the third axis perpen-
dicular to these two; but the eggs obtained from one bird averaged

? See also W. Craig, Oviposition induced by the male in pigeons. Jour. Morph.,
vol. 22, p. 299, 1911.

BILATERALITY OF THE PIGEON’S EGG 291

16.7 by 14.7 by 16.83 mm. The details of this. final period of
rapid yolk secretion have been described by Loyez (’05-6), Riddle
(11) and others; suffice it to say, that the yolk is laid down con-
centrically so that the long axis is preserved and the latebra retains
its eccentric position. Figure38shows the conditions in an oocyte
twenty-four hours before ovulation; the end of the long axis
which was directed toward the cloaca is toward the right and it
is obvious that the latebra is nearer the other, ‘infundibular’
end. In oviducal eggs the outlines are no longer so clear and often
the eccentricity of the latebra is only represented by an extension
toward the infundibular end of the long axis as is shown in figure
39 which is the cut median surface of an oviducal egg (see also
description of the figures). In laid eggs the softening of the
yolk and transfusions make the relations much less distinct than
in the earlier stages.

2. Correlations in the reproductive apparatus. As has been said
usually but two oocytes enter upon the final growth period
together. Shortly after the initial stimulus has been received
in the ovary, the oviduct becomes highly vascular and increases
insize. Inonebird that wasstudied the larger follicle was 9.1 mm..,
less than one-half the definitive size, the oviduct with its fim-
briated infundibulum was about one-fourth the maximal size and
both funnel and glandular oviduct were in peristalsis. In another
instance, where the larger follicle was 13.2 mm. (in long axis),
the second 8.2 mm. the oviduct was almost full size and the active
funnel was wrapped about the larger follicle. The correlation of
ovarian and oviducal activities is presumably due to the presence
of the internal secretion of the ovary in the circulation. It may
be said in this connection that the interstitial cells of the ovary
show much greater signs of activity in functioning ovaries than
they do in ovaries from birds that had not laid for a long time.

The whole reproductive mechanism is delicately balanced and
there are interesting physiological problems here still to be worked
out. How is it, for example, that the funnel is attracted to the
follicle, and that it, eventually, always clasps the larger one,
though in the early stages of the final growth period it may, rarely,
be found about the smaller? What determines that usually but
two follicles mature at a time and that never more than two

292 GEORGE W. BARTELMEZ

eggs are laid? How comes it that yolk secretion, and other
factors are so regulated that ovulation.is closely correlated with
the nuclear maturation phenomena?

To discuss ovulation and the method of orientation it will be
necessary first to consider the relation between mature follicle
and oviduct. If the reproductive apparatus of a bird be studied
two to twenty-four hours before the rupture of the first follicle
is due, the following conditions are found: Oviduct and funnel are
in active peristalsis, the latter is closely wrapped about the larger
follicle, endeavoring so to speak, to swallow it. Under such con-
ditions the follicle is obviously oriented along lines of least resist-
ance, and accordingly its long axis coincides with that of the
oviduct and approximately with antero-posterior axis of the bird.
One end of the long axis of the follicle is therefore directed
posteriorly, toward the cloaca, and this may be termed the
cloacal end of the egg. It will be ‘remembered that, in the
incubated egg, the cloacal end of the egg is definitely related to
the head of the embryo (diagram I, p. 275) and since the antero-
posterior axis of the embryo is determined before ovulation, it
follows, when the method of orientation and ovulation is taken
into account, that this cloacal end of the long axis is predeter-
mined in the ovary. Figure 36 is from a ventral view of a pair
of follicles about twenty-four hours before ovulation. In the
larger one the stigma appears in the long axis of the follicle, and
the long axis extends antero-posteriorly with reference to the
bird. The funnel (inf.) contracted in preservation and is seen
on the left side of the bird.

How comes it that one end of the long axis is found nearer the
cloaca? The explanation appears when a mature follicle is
removed from the bird and suspended at the center of the animal
pole in an albumen solution of the density of that which fills
the body cavity at the time of ovulation; it is found that one end
of the follicle is heavier than the other, and this is’due to the
eccentricity of the latebra (fig. 38) whose position nearer the
infundibular end of the long axis is due, as will be remembered,
to the corresponding position of the germinal vesicle in the early
stages. The cloacal end of the ovum, i.e., the end which goes

BILATERALITY OF THE PIGEON’S EGG 293

down the oviduct first, can therefore be traced back to the pri-
mordial follicle. The chief factor, then, in the orientation of
the follicle is the fact that the egg is heavier at one end of the
long axis and, in the normal position of the bird, this end gravi-
tates toward the cloaca. In some cases the position of the
ovarian stalk corresponding to that of the latebra may also help
in the orientation, but this not a constant feature. The orien-
tation of the follicle in the oviducal axis is due, in the pigeon,
primarily to the activity of the funnel; its pressure from without,
together with the ever increasing pressure from within, i.e., from
the continued yolk secretion, are the main factors in the ruptur-
ing of the follicle.

The pressure due to yolk secretion is considerable as may be
judged from the way in which the egg bulges out when the rup-
turing of the follicle begins and also from the observation that
the egg is over a millimeter in diameter greater after ovulation
than the whole follicle was before.

The funnel can exert considerable pressure, for its wall is mus-
cular and it is attached anteriorly by part of the dorsal oviducal
ligament to the left body wall and posteriorly by the ventral
ligament to the ventral margin of several coils of the oviduct.
Miss Curtis (10) has clearly brought out these latter relations
in the hen and made several illuminating suggestions on ovula-
tion. She describes a ‘pocket’ formed by the body wall, the left
abdominal air-sac and the intestine with its mesentery on the
right side, surrounding the ovary. The mouth of this pocket
is occupied by the funnel. These relations undoubtedly play a
part in the orientation of the follicle and ovulation, and decrease
the possibility of the egg escaping into the body cavity. In the
pigeon however, I believe that the peristaltic action of the funnel
and the yolk secretion are the principal factors in orientation
and ovulation, for all these other conditions may be modified by
keeping the bird on her back during ovulation and still the ovum
succeeds in entering the oviduct.

The orientation may take place in the course of two and a half
to three hours, which is the time that elapses between the laying
of the first egg and the rupture of the second from the ovary,

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

294 GEORGE W. BARTELMEZ

assuming that the funnel is inactive while an egg is in the ‘uterus.’
The evidence for this is that in a dozen or more cases studied in
which the first egg was in the uterus, the funnel was not found
about the second follicle and no peristaltic movements were
observed in it. This is supported by the fact that in the hen,
where there are normally several large follicles, Patterson (’10)
found the funnel inactive while there was an egg in the uterus.
As has been said above, three follicles occasionally mature in
the pigeon. In two such cases the following conditions were
found. The first egg was laid normally and that night the bird
was killed. The second egg was in the oviduct, but in addition
a third one was found just ruptured from its follicle and in the
base of the funnel. This indicates that the funnel remains active
for some time after ovulation, but considering that there is no
record of a pigeon having laid more than two eggs at a sitting,
it seems probable that after an egg enters the ‘uterus,’ an anti-
peristalsis ejects any other that may be in the oviduct. In one
of the cases referred to, the orientation of the follicle in the funnel
probably did not take over an hour.

Another important factor in the orientation of the ovum in
the oviduct is the way in which the follicle ruptures. The stigma,
it will be remembered, extends in the long axis along the free
pole of the follicle and so its cloacal end lies at the base of the
funnel where this passes over into the glandular part of the oviduct.
It will readily be seen that the cloacal end of the stigma is there-
fore the one part of the follicle where the pressure from within
is not balanced by any pressure from without. Now when the
various pressures become great enough, the rupture begins at
the cloacal end of the stigma (a statement based on the study
of over two hundred recently ruptured follicles); usually before
the tear has extended along more than 10 mm. of the stigma,
the ovum has been squeezed out of the follicle, mainly by the
rapidly contracting wall of the latter, and lies at the base of the
funnel. The escape of the ovum through an opening one-half
its own diameter—I have watched this several times—indicates
the elasticity of the extremely delicate vitelline membrane (chor-
ion) and the flexibility of the ovum. In spite of this distortion

BILATERALITY OF THE PIGEON’S EGG 295

the moment the ovum is free it assumes the elongate form it
had in the follicle. Eggs newly ruptured from the ovary clearly
show the long axis perpendicular to the polar axis and we are
dealing with the same long axis in both cases for it is possible
to remove a mature follicle from the bird, carefully tear the
cloacal end of the stigma, and, under the most favorable con-
ditions for observation, watch the process of ovulation. The long
or chalazal axis of the oviducal egg is not, therefore, the result
of pressure from the walls of the oviduct, but is the long axis
which has persisted from the primordial follicle stage.

As the ovum is entering the glandular portion of the oviduct
the walls of the latter attach to the cloacal end a small button
which is part of the chalaziferous albumen, and when the whole
ovum has entered, the infundibular chalaza is formed at the
opposite end of the long axis. Usually the first formed (‘cloacal’)
chalaza is the heavier and it is almost always firmly attached to
the egg membrane, while the infundibular chalaza is sometimes
represented only by the ‘button.’ The long axis is now also
the ‘chalazal axis.’

A word may be said here with reference to the relation of the
maturation phenomena to ovulation. No careful study of the
maturation spindles, sperm nuclei or pronuclei has here been
made, but, so far as I have gone, I have seen nothing except
confirmations of Harper’s excellent account of these cytological
details. The breaking down of the germinal vesicle occurs be-
tween six and eight hours before ovulation and usually the pro-
cesses continue up to the metaphase of the second maturation
division, while the oocyte is still within the ovary; they do not
proceed any further unless fertilization takes place. The evi-
dence for this is as follows: Three of the eggs obtained at the
moment of ovulation were sectioned and the second maturation
was found to be just at metaphase; further, eleven eggs taken
shortly after ovulation, i.e., eggs found at the beginning of the
glandular part of the oviduct were all in the final stages of the
second maturation division; finally, it happens occasionally that
a follicle fails to rupture within three or four hours of the usual
time of ovulation, (8 p.m.) and in these instances also, the equa-

296 GEORGE W. BARTELMEZ

torial plate stage of the second maturation spindle is found. It
may be said of the pigeon’s egg, then, as of every other well
established case in the vertebrates, that the egg does not proceed
beyond the metaphase of the second maturation unless it be
fertilized. It agrees also with most (perhaps all?) vertebrate
eggs in that normally the first polar body is given off in the ovary,
and ovulation takes place while the second spindle is in metaphase.

I have no explanation to offer for the definite relation between
the breaking down of the germinal vesicle and ovulation. ‘Taking
all the data into consideration, however, it would seem that the
maturation processes begin soon enough so that they may progress
as far as the middle of the second maturation and rest there until
the yolk secretion and the other factors have brought about
the rupture of the follicle. The latter may take place during the
first maturation, judging from the fertilized egg figured and
described by Harper, ’04, figures 6 and 6a, still I am convinced
that the account given above describes the usual course of events,
since Harper’s is the only similar case that has been found.

V. THE BILATERALITY OF THE BLASTODISC
(Summary of Part IT)

A. Origin

The only reference in the literature to the origin of the blasto-
disc in the bird’s egg is, so far as I know, Coste’s (’47) surmise
that it arises from ‘le contenu granuleux’ (the spherule cap, p.
277, ss.), but the observations of Agassiz and Clark (57) on the
turtle egg are of value in connection with the conditions described
here. It has been shown that the spherule cap of deutoplasmic
granules is used up during the period of differentiation of the
oocyte (p. 286). The first traces of a blastodise appear toward
the end of this period of development, when the oocyte is about
2 mm. long axis (in life), and when the germinal vesicle has begun
to flatten out against the follicular epithelium. Figure 26 shows
a longitudinal median section of an oocyte in this stage and figure
28 is from the animal pole of the same oocyte more highly magni-
fied. It will be noted that the reticulum, the spaces of which

BILATERALITY OF THE PIGEON’S EGG 297

represent vacuoles in life, shows a finer meshwork about the
germinal vesicle than elsewhere. These smaller vacuoles sur-
round the germinal vesicle (g. v.) symmetrically on all sides,
except where it adjoins the peripheral protoplasm and from this
area the blastodisc is differentiated. The same conditions may
be observed in figure 27 in which, however, the germinal vesicle
has been distorted by compression in sectioning. The develop-
ment of the blastodisc symmetrically about the germinal vesicle
and the coincident changes within the latter indicate that
the former differentiates in close association with the germinal
vesicle and lend support to the suggestion made above that the
nuclear axis of bilaterality is transmitted to the blastodisc. In
figure 27 the deeply staining yolk granules, which are present
throughout the cell except in the peripheral protoplasm and the
latebra, are small in the neighborhood of the germinal vesicle
(g. v. ); they remain as the characteristic granules of the blasto-
disc. Throughout development they remain connected with the
typical white-yolk spheres by all possible transitional forms and
are to be looked upon simply as deutoplasm in a form that can
be immediately assimilated by the protoplasm. This ‘digested’
yolk persists only about the germinal vesicle, characterizing the
blastodise. Fig. 8 is from an oocyte with a clearly defined anlage
of the blastodise surrounding the germinal vesicle, which in life
has the form of a biconvex lens. Centrally, and at the peri-
pheral margin, the blastodise merges with the bed of white yolk,
the intermediate zone being the anlage of the central and marginal
periblast of Blount (09), the finely granular region about the
germinal vesicle giving rise to the segmental disc (see figs. 35, 40
and 43 for the regions in the mature egg). By this time the
peripheral protoplasm has become narrower, less so over the
blastodise than over the rest of the periphery (fig. 29).

Oocytes at the end of the period of differentiation, i.e., in the
résting stage before the final rapid growth, have the blastodisc
well developed; all the definitive ooplasmic structures are laid
down. Fig. 7 shows the blastodise of such an oocyte in oblique
surface view, from which aspect the germinal vesicle appears
circular. A median but oblique section of the same oocyte (fig.

298 GEORGE W. BARTELMEZ

29) also shows the blastodise regions, and in adjoining sections
a broad connection can be traced between the central periblast
and the latebra. The photograph shows an additional feature:
just below the peripheral protoplasm is a layer of deeply staining
granules, (wdg.) extending from the marginal periblast to the
germinal vesicle (g. v.). This layer of granules appears wedge
shaped in sections of later stages and will be referred to as the
‘wedge’ (wdg.), although actually it forms a broad collar, thicker
peripherally, about the germinal vesicle. The granules of the
wedge are characterized by being somewhat larger than the neigh-
boring granules and by their strong affinity for basic dyes. At
its central margin the wedge is continuous with a layer of basic-
staining but much smaller granules which form a thin stratum
over the germinal vesicle except at its outer edge where they form
a thickened rim. From their subsequent relations these have
been termed the polar granules (fig. 29, p. g.). Both groups can
be traced through the succeeding stages, and during maturation
and fertilization they are shifted about by cytoplasmic currents
so as to form configurations which are among the clearest evi-
dences of the bilateral organization of the blastodise. These
granules take the violet in the neutral gentian stain as intensely
as do the nucleoli and the granulations within the yolk spheres,
differing in this respect from the rest of the granules of the
blastodise, as appears clearly in figures 29 and 30, 40 and 41.

B. The blastodisc during maturation and first cleavage

For lack of space at the present time the detailed description
of the succeeding stages must be reserved for a later publication.
The evidences of bilaterality in these stages are essential for the
general thesis here maintained and will be briefly summarized.
About two days before ovulation the periblastic zones which
characterize the mature egg are established, first posteriorly, then
anteriorly. A surface view of a blastodisc in this stage is shown
in figure 42. This is the first direct morphological evidence that
the blastodise of the oocyte is bilaterally organized. It shows
beyond question that the embryonic axis as such exists in the
ovarian egg and establishes the validity of the reasoning from the

BILATERALITY OF THE PIGEON’S EGG 299

evidence in the ovarian history which independently led to this
conclusion.

After ovulation and fertilization, i.e., between the second matur-
ation and the first cleavage, the whole blastodise gives evidence
of its bilateral organization by a progressive change in shape
which can be followed in the living egg. At the time of ovula-
tion, which usually takes place when the second maturation spindle
is at metaphase, the segmental disc and periblastic zones are
circular, so far as can be seen. After fertilization they gradually
become elliptical, so that the antero-posterior diameter is shorter
than the right-and-left, as may be seen in figures 35, 36 and 43.
A periblastic ring which appears at this time gives further evidence
of the bilateral organization of the blastodisc. The most strik-
ing manifestation of bilaterality, however, is within the segmen-
tal dise itself. The granules forming the ‘wedge’ are gradually
moved from the anterior side of the segmental disc and form a
crescent around its posterior margin. This movement can be
followed in the living egg and will appear clearly if figures 40
and 41.be compared. The changes that take place between
these two stages have been worked out in detail and will be pub-
lished shortly.

VI. AXIS ANGLES

The long (or chalazal) and embryonic axes

It has been seen that in the incubated egg, two axesof bilat-
erality can be distinguished:

1. The embryonic axis.

2. The axis of bilaterality of the ovum as a whole, which is
defined by: (a) the long axis of the ovum; (b) the position of
the latebra nearer one end of the long axis.

The relation between these axes is shown in diagram I, and
its development is illustrated in diagram II. The essential fea-
ture of this relation is that when a definite end of the long axis
is held to the right, the head of the embryo is away from the
observer. Hitherto this relation has been described as existing
between the embryonic and chalazal axes, but it was shown,
(p. 295), that the long axis is the primary one, since it determines

300 GEORGE W. BARTELMEZ

the chalazal axis when the egg is oriented in the funnel. A word
should be said here with reference to the chalazae. Normally
they are attached to either end of the long axis, but sometimes
this attachment is irregular. Patterson (’09, p. 68) found this to
be the case in 8 per cent of the eggs; however he included in this
category, cases in which only a button is attached to the in-
fundibular end of the egg, and no chalazal thickening appears.
There are other cases in which the chalazae are not attached to
the ends of the long axis, i.e., in which the long and chalazal axes
do not coincide. I observed six such out of 299 eggs, namely
2.2 per cent. Considering the mechanical factors involved in
the orientation in the oviduct, this much abnormality is to be
expected. Accordingly, the results described here on the relation
of embryonic to long axis may be compared with those on its
relation to the chalazal axis which are given by the previous work-
ers who overlooked the long axis.

Blount (’09) and Patterson (’09) are the only authors who have
made statements with regard to the relation between the en.bry-
onic and long axes in the pigeon. The latter author states that
the angle between them is 45 degrees in 90 per cent of the cases.
No estimate of the probable error was given in the method of
measurement that was used, which was to orient the egg under the
binocular so that the chalazal axis was parallel to the base line
of a square ruled micrometer scale in one ocular, then to note the
position of the embryo, and derive the angle. There are several
sources of error in such a method and the observations were fur-
ther hampered by the fact that many of the measurements were
made on primitive streak stages. (It should be noted that the
angles given by Patterson are the complements of those given
here.) The method used in this study was to remove the shell
in salt solution, noting the relations of the smaller end of the shell,
the heavier chalaza, and the air chamber, for the two former mark
the end of the egg which went down the oviduct first and are
invariably opposite the air chamber. After it had been noted
that the long and chalazal axes coincided, the base of a protractor .
was laid parallel to them, and without moving the head, a slip of
glass 3 mm. wide was moved over the protractor so as to coincide

BILATERALITY OF THE PIGEON’S EGG 301

with the embryonic axis and then the angle was read at the center
of the glass slip. The probable error in measuring was found to
be from two to three degrees and in the measurements the aver-
age of three determinations was taken.

The series of measurements so made showed that there is much
greater variation than has been supposed in the pigeon’s egg,
but that it is not so great as Rabaud (’08) found in the hen’s egg
(see footnote p. 276). This variation is not due to any change
of angle as the embryo develops, for in the eggs of a given bird

Diagram III A polar view of anincubated egg illustrating the extremes of varia-
tion in axis angles in the pigeon. Complete inversion (180 degrees) involves
other factors and is not a simple variation, so it is not considered here.

the average of angles measured after thirty-six to forty-eight
hours of incubation was the same as after eighteen to thirty-six
hours. The same conclusion was reached by Rabaud for the
hen’s egg after a series of direct observations. Most of the
observations recorded here were made on embryos from three to
twelve somites. All workers have mentioned a certain amount
of variation, even to 180 degrees. In the pigeon. only four such
cases of complete inversion have come to my attention; one
was observed by Dr. Patterson one by Dr. Blount and two by
myself; that is, four out of more than six hundred careful obser-
vations showed that that end of the long axis which is usually
the infundibular one, went down the oviduct first. Aside from
these, the extremes of variation observed were 8 degrees and

302 GEORGE W. BARTELMEZ

135 degrees. A curve, (diagram IV) based on all the eggs I
studied, viz., 299, taken from eighty-two different birds, shows
that these extremes are rare. Four modes appear in the curve,
at 50, 70, 80 and 90 degrees. This suggests the most inter-
esting condition that has appeared from this study of axis
angles; namely, that the relation between embryonic and long
axes is far more constant for the eggs of an individual than for
eggs obtained from different birds—a maximum variation of 50
against 127 degrees. Table 2 shows that most of the eggs laid
by a given bird have almost identical relations, when one re-
members that five degrees is a very slight variation under such
conditions as these:

TABLE 2
FEMALE NO. 3 FEMALE NO. 4 FEMALE NO. 2

Degrees Eggs | Degrees | Eggs Dezrees Eggs

90-95 8 35 1 75-80 3

96-100 14 61-65 | 3 81-85 1
101-105 5 66-70 5 | 86-90 5
106-110 6 71-75 | 7 | 91-95 5
111-115 3 76-80 | 7 96-100 0
120-125 10 81-85 | 8 101-105 3
130 1 86-90 | 5

It appears that there is one grouping of the eggs of no. 3, about
100 and another about 125 degrees. The fact that it is so excep-
tional to find a pigeon’s egg in which the axial angle is greater
than 90 degrees makes the case all the more striking. The
record of no. 2 shows that her eggs occasionally had the axis
angle greater than 90 degrees; among the eggs of the other
eighty birds studied, only twelve of the remaining 287 had angles
of 90 degrees or more. The variation in no. 4 is more typical.
In the curve, (diagram IV) 66 eggs obtained from the two birds
whose axis angles were constantly greater than 90 degrees, are
omitted; the dotted line at the right shows the curve with these
eggs included.

BILATERALITY OF THE PIGEON’S EGG 303

This constancy of the axis angles in the eggs of a given bird
gives further support to the thesis that has been maintained
throughout this paper, namely that this relation of axes is deter-
mined by factors which are themselves expressions of the bilateral
organization of the egg; that is to say, the organization expresses
itself in a most constant fashion in the eggs of a given bird. If

10° 50° (Ue a0° 125°

Diagram IV Two curves illustrating the variability in the relation between
the embryonic and long axes in the pigeon egg. ‘A’ was plotted from observations
on 299 eggs. They were grouped in 5 degree classes and the number of eggs in a
class is plotted on the ordinates, the angles on the abscissae. There are three
modes in the ‘normal’ curve (see text), at 50, 70 and 90 degrees. The broken line
at the right represents the angles of the eggs laid by two birds whose norm was
above 90 degrees. B (the heavy line) is a curve plotted in a similar way from obser-
vations on 59 eggs in stages previous to the third cleavage. The angle measured
was between the shorter axis of the blastodisc and the long axis of the ovum. The
similarity of the two curves is obvious.

this be true, the nucleo-cytoplasmic relation (diagram II B) in
the younger oocytes of a given ovary should correspond to the
axis angles in the eggs laid by the same bird. The ovaries of
only two of the birds in which a norm had been established, were
studied; one was no. 2 mentioned above, the other a bird whose

304 GEORGE W. BARTELMEZ

eges ranged about 70 degrees. In the ovary of the former many
oocytes were found in which the nucleo-cytoplasmic angle was
clearly 90 degrees or more, while the majority of the oocytes of
the other had an angle less than 90 degrees. This meager evidence
may be taken for what it is worth, for obviously the personal
factor might find expression through various channels in such
a study. Another possible line of evidence in this regard is the
study of the axis angles of the eggs laid by the offspring of a bird
previously studied. Experiments with this end in view are now
in progress. The heavy line B in diagram IV is a curve plotted
from the angles observed between the short axis of the blastodise
and the long axis of the whole ovum, in stages previous to the
third cleavage. The similarity between the two curves makes
it practically certain that the short axis of the blastodise is the
antero-posterior axis, especially when one bears in mind the large
body of evidence which goes to show that the anterior end of
the embryo is predetermined in the ovary.

Three conclusions may be drawn from these observations on
axis angles:

1. The essential feature of the relation between the embryo and
ovum as a whole is the fact that the embryo bears a different
relation to one end of the long axis than to the other.

2. The eggs of a given bird vary much less as to their relation
between embryonic and long axes than do the eggs of different
birds.

3. The short axis of the unsegmented blastodisc is identical
with the antero-posterior axis of the embryo.

VII. DISCUSSION AND SUMMARY
A. Bilaterality in vertebrate ova

The fact that a fundamental character like bilaterality appears
during the ovarian history in so highly specialized a form as the
pigeon, suggests the possibility that in the vertebrates as in the
insects the antero-posterior axis of the embryo is predetermined
in the ovary. The suggestion is strengthened by the striking
manifestations of bilaterality which Conklin (’05) describes in

BILATERALITY OF THE PIGEON’S EGG 305

the fertilized eggs of the ascidians. It will be of interest then
to consider briefly the evidences and indications of bilateral organ-
ization that have been found in vertebrate eggs.

Myzxinoids. The ovarian eggs of Bdellostoma have been de-
scribed as being bilaterally symmetrical, their shape correspond-
ing somewhat to that of many insect eggs. Dean (’99) denied
this for ovarian or newly laid eggs, but it must be remembered
that the question is complicated here, since the eggs are subjected
to a great and rapid change in pressure when they are collected.
The possibility exists that the antero-posterior axis of Bdell-
ostoma may arise in the ovary.

Selachians. The selachian egg presents many striking resem-
blances to that of the bird, as is to be expected, since these two,
together with the reptilian egg are the most extreme types of
meroblastic ova. The segmental disc in the selachians, as in the
birds, is surrounded by a periblast; in the latter Riickert (’92)
found evidences of bilateral symmetry while the egg was still
in the ovary. He described in Torpedo an extension of the
marginal periblast surrounding the blastodise and projecting down
into the yolk mass. This ‘Mantel’ showed considerable vari-
ation in form, but it was found to be constantly deeper and more
sharply defined at one side of the disc than the other. In some
eggs this structure could be traced from late stages of ovarian
eggs, through the fertilization and cleavage stages, and rarely
to the time of gastrulation when it appeared as if the deeper
side of the ‘Mantel’ were posterior. Rickert pointed out the
significance of this yolk configuration, if it were differentiated
along the antero-posterior axis, but he did not consider his evi-
dence as adequate to warrant a definite conclusion. I may say
that I have found similiar conditions in the eggs of Raja ocellata.
The variation in these axial relations is to be accounted for as
follows: The relations are only expressions of the bilateral organ-
ization, not the organization itself; and so the expression varies
in different eggs, but is relatively constant for the eggs of a given
bird. The relations may be modified by other factors in develop-
ment but the bilateral organization is nevertheless present; in

306 GEORGE W. BARTELMEZ

other words, these relations simply give evidence that such an
organization exists.

It may be then that the selachian and bird’s eggs agree not
only in that the axis of bilaterality appears in the ovary butalso
in that the first clear evidence of the bilateral organization of
the blastodise is found in the marginal periblast. This is the
only instance that has come to my attention in which definite
morphological evidence was found indicating the ovarian origin
of the axis of bilateral symmetry in a vertebrate.

Teleosts. Several authors have noted that the blastodise of
teleost eggs is thicker on one side than on the other; Oellacher’s
(’72) figure 17 indicates that this difference exists before cleavage
begins, i.e., that the thickening of the blastodise after the laying
of the egg is more rapid on one side. Agassiz and Whitman,
(85) and Kowalewsky (’86) consider this an antero-posterior
differentiation, the thicker being the posterior side, though nothing
is said of continuous observations to confirm this. It is quite
possible that here as in the Amphibia fertilization initiates the
appearance of the axis of bilaterality.

Amphibians. Many workers, notably Roux, Brachet (’05) and
Schultze (99), have shown that the amphibian egg is bilaterally
organized previous to cleavage. This organization expresses itself
in a definite movement of the superficial pigment granules so that
a gray crescent appears at the posterior side of the egg, the center
of the crescent marking the position of the blastopore in later
stages. It seems unlikely, from what is known of other eggs,
that the point of entrance of the spermatozoon determines the
axis of bilaterality, as Roux has maintained; bilaterality manifests
itself, independently, after the stimulus of fertilization has been
received. While the copulation path may determine the plane
of the first cleavage, both may be quite independent of the axis
of symmetry, as Brachet’s work has clearly shown. The arrange-
ment of,the different kinds of yolk seems to be radially symmetri-
cal, and how far the bilateral configurations of pigment granules
described by Van Bambeke (’76) in the toads’ egg are related to
the bilaterality of the embryo and how far they are due to the
mitotic forces, which here seem to be independent, remains to

BILATERALITY OF THE PIGEON’S EGG 307

be seen when these eggs are sectioned with reference to the gray
crescent.

Reptiles. The statement of Will (’93, p. 15) that in the gecko
ege the long axis of the embryo is approximately perpendicular
to the longest axis of the entire ovum is strong evidence that
conditions similiar to those described here exist in that egg. The
early appearance of the antero-posterior axis of reptiles is sug-
gested through the analogy of the bird’s egg, by the fact that most
workers on the early stages have found the blastodisec more or
less elliptical in shape.

Birds. Coste (’47), Kélliker (76), Duval (in his atlas) and
Patterson (’10) have figured surface views of the hen’s egg in
precleavage and early cleavage stages. In every case the segmen-
tal disc, the inner periblastic zone and the periblastic ring are
shown and KOlliker figures a narrow outer zone about the latter.
With one exception these zones are drawn circular, but Patterson’s
figure 13 shows them elliptical and the shorter axis of the ellipse
formed an angle of 90 degrees with the oviducal axis and coin-
cided therefore with the embryonic axis. It is probable that
in the hen’s egg the bilaterality is not so clearly manifested by
the change in shape of the blastodisc as is the case in the pigeon’s
egg. As has been said above, the orientation in the hen’s egg
seems to be less constant, but the matter needs further study.

Mammals. The definite orientation of the embryos of various
mammals with reference to the chorionic vesicle and the uterus
suggest that here too the antero-posterior axis may be traced
to an early stage of development.

B. Summary
See diagram II, p. 279

A definite relation exists in the pigeon’s egg between the axis
of the embryo and the long axis of the ovum.

Both the embryonic and the long axis are present in the ovarian
egg, that is, the antero-posterior axis of the pigeon is predeter-
mined in the ovary.

The ovarian history may be divided into four periods in each
of which the oocyte exhibits a characteristic organization.

308 GEORGE W. BARTELMEZ

(1) The first growth period, the final stages of which are pre-
sented by the youngest oocytes in the adult ovary—the primordial
follicles. The primordial follicle has a bilateral structure which
can be traced through the ovarian history and which is marked
primarily by the fact that the polar axis intersects the long axis
nearer to one end.

(2) The! second growth period, during which’ the bilateral sym-
metry of the oocyte is impressed upon the connective tissue fol-
licle which in turn plays a part in the preservation of the sym-
metry of the oocyte as a whole during the subsequent growth.
The germinal vesicle comes to lie nearly but not quite at the center
of the oocyte, its eccentricity marking the polar axis which is
constant throughout oogenesis. Two ooplasmic zones are differ-
entiated: central and peripheral protoplasm.

(3) During the period of differentiation the germinal vesicle
migrates peripherally, its path determining the position of the
latebra, which as a result of the corresponding position of the
germinal vesicle in the preceeding stages, is nearer one end of
the ooplasmic long axis. The Anlage of the blastodisc appears
about the germinal vesicle. The oocyte becomes free to rotate
within its follicle, probably as a result of the formation of the
zona radiata.

(4) The final growth period is initiated by the stimulus of
mating and in the course of it the great mass of yolk is laid down
in such form that the eccentricity of the latebra is preserved.

Data have been obtained on the character of the process of
ovulation and the orientation of the egg in the oviduct. The
position of the latebra nearer one end of the long axis determines
which end of the egg shall pass down the oviduct first and the
activity of the infundibulum and other factors orient the follicle
so that its long axis coincides with the oviducal axis.

The blastodise is formed symmetrically about the germinal
vesicle and the segmental dise and periblastic zones can be dis-
tinguished by the end of the period of differentiation. The embry-
onic axis first appears unmistakably in the formation of the
periblast in the mature ovarian egg. During the final stages of the
second maturation, at the time when fertilization has normally

BILATERALITY OF THE PIGEON’S EGG 309

taken place, the antero-posterior axis is expressed by the oval form
of the blastodise and by granule movements in periblast and
segmental disc. The most characteristic of the granule con-
figurations is a crescent formed about the posterior side of the
segmental disc by certain granules (‘the wedge’) which are
differentiated very early in the history of the blastodise. These
granules are deutoplasmic in character and their arrangement
with reference to the bilaterality of the embryo is due to activi-
ties within the living substratum, that is, the ground substance.

The essential feature of the relation between the embryo and
the long axis of the ovum is the fact that when that end of the
ovum which is predetermined in the ovary to pass down the
oviduct first, is held to the right, the head of the embryo is
directed away form the observer. The actual angle between
embryonic and long axes is subject to much greater variation
than is generally supposed, but it is relatively constant for the
eggs laid by a given bird.

&. Conclusions

If the evidence here offered that the structure of the primordial
follicle determines the end of the egg which shall pass down the
oviduct first, and that this end is definitely related to the embry-
onic axis of symmetry be accepted, the conclusion is warranted
that the structure of the primordial follicle is a manifestation
of the bilateral organization of the oocyte. In other words, the
antero-posterior axis of the pigeon is defined at least as early as
the stage of the primordial follicle.

The conclusion that the relation between the embryonic axis
and the long axis of the entire egg is an expression of the bilat-
eral organization of the ooplasm is also supported by the fact
that this relation is much more constant for the eggs of a given
bird, than for eggs obtained from different birds.

An explanation of the relation between the long axis of the
entire ovum and the axis of the embryo is suggested by a corre-
sponding relation found in many young oocytes between this
same long axis and the axis of the germinal vesicle, especially

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

310 GEORGE W. BARTELMEZ

when it is borne in mind that the blastodise from which the
embryo arises is differentiated under the direct influence of the
germinal vesicle.

The fact that in the pigeon the polar axis persists unchanged
throughout the growth period of the oocyte, and the fact that
there is a polar axis in the earliest stages of the germ cells of all
vertebrates which have been described, indicate that the polar
axis persists unmodified from generation to generation in the
vertebrates and is one of the fundamental features of the organ-
ization of the protoplasm. The facts here presented may be
taken as evidence that bilaterality appearing as early as it does
in development, is likewise an expression of a fundamental char-
acter of the protoplasm.

BIBLIOGRAPHY

Aqaassiz, A., AND WHITMAN, C. O. 1885 The development of osseus fishes.
Mem. Mus. Comp. Zool., vol. 14, p. 3.

Agassiz, L., AND Cuark, H. J. 1857 The embryology of the turtles. Contrib.
Nat. Hist. of U.S. A., vol. 25 p. 451.

AuLEN, B. M. 1906 The origin of the sex cells of Chrysemys. Anat. Anz’, vol.
29, p. 217.

Baurour, F. M. 1878a A monograph on the development of the elasmobranch
fishes. London.

1878 b On the structure and development of the vertebrate ovary.
Jour. Mic. Sc., vol. 18, p. 383.

BAMBEKE VAN ©. 1876 Recherche sur l’embryologie des Batraciens. Bull.
Acad. Roy. Se. Belgique, Ser. 2, tome, 41, p. 97.

1880 Nouvelles recherches sur l’embryologie des Batraciens. Arch.
Biol., vol. 1, p. 305.

1896 Sur une groupement des granules pigmentaires, etc. (in the toad’s
egg). Bull. Acad. Roy. Sc. Belgique, Ser. 3, tome 31, p. 29.

Bensuey, R. R. 1911 Studies on the pancreas of the guinea pig. Am. Jour.
Anat., vol. 12 [technique].

Buiount, Mary 1909 The early development of the pigeon’s egg, with especial
reference to polyspermy and the origin of the periblast nuclei. Jour.
Morph., vol. 20, p. 1.

Bracuet, A. 1905 Recherches experimentales sur l’oeuf de Rana fusca. Arch.
- Biol., tome 21, p. 108.

BILATERALITY OF THE PIGEON’S EGG 311

Conkuin, E. G. 1902 Karyokinesis and cytokinesis in the egg of Crepidula.
Jour. Acad. Nat. Sciences. Phil. Ser. 2, vol. 12, part 1.

1905 The organization and cell-lineage of the ascidian egg. Ibid.,
Ser. 2, vol. 13, part 1.

Costr, M. 1847 Histoire générale et particuliere du développement des corps
organisés. Tomel. Paris.

Curtis, M.R. 1910 The ligaments of the oviduct of the domestic fowl. Maine
Ag. Exp. Station, Bulletin 176.

Dean, B. 1899 Onthe embryology of Bdellostomastouti. Festsch. fiir v. Kupf-
fer, p. 221.

D’HoutanpeEr, F. 1904 Recherches sur l’oogenése et sur le structure et le sig-
nificance du noyau vitellin de Balbiani. Arch. Anat. Mic., tome 7,
TOs IM

Duvat, M. 1884” La formation du blastoderm dans l’oeuf des oiseaux. Ann. des
Sc. Nat. Ser. 6, tome 18, p. I.

E1GenMANN, C.H. 1892 On the precocious segregation of the sex-cells in Cyma-
togaster aggregatus. Jour. Morph., no. 5, p. 481.

1892 Sex differentiation in the viviparous teleost Cymatogaster. Ar.
f. Entwmech., Bd. 4, p. 125.

Evans, H. M. 1909 On the earliest blood vessels in the anterior limb buds of
birds and their relation to the primary subclavian artery [technique].
Am. Jour. Anat., vol. 9, p. 281.

Fiscuen, A. 1903. Entwickelungund Organdifferenzirung. Arch. f. Entwmech.
Bd. 15, p. 679.

Guyer, M.F. 1909a Thespermatogenesis of the domestic guinea. Anat. Anz.,
vol., 34, p. 502.

1909 b The spermatogenesis of the domestic chicken. Anat. Anz.,
vol. 34, p. 573.

Harper, E. H. 1904 The fertilization and early development of the pigeon’s
egg. Am. Jour. Anat., vol. 3, p. 349.

Harvey, B. C. H. 1907 A study of the structure of the gastric glands, etc.
[technique]. Am. Jour. Anat., vol. 6, p. 207.

Hertwiac. O. (Ed.) 1906 Handbuch der vergl. u. exp. Entwicklungsgesch..,
Bd. 1, no. 1, Jena.

Hott, M. 1890 Ueber die Reifung der Eizelle des Huhns. Sitzber. Akad. Wiss.
Wien., Bd. 99, p. 311.

Kionka, H. 1894 Die Furchung des Hihnereies. Anat. Hefte, Bd. 3, p. 429.

KGLLIKER, v., A. 1876 Die Entwicklungsgeschichte des Menschen und der
héheren Wirbeltiere. Leipzig.

KowaLeEwskI, v., M 1886 Ueber die erste Entwicklungsprozesse der Knochen-
fische. Zeit. f. Wiss. Zool., Bd. 48, p. 434.

alles GEORGE W. BARTELMEZ

Linus, F. R. 1901 The organization of the egg of Unio. Jour. Morph.,vol.
17, p. 227.
1906 Observations and experiments concerning the elementary phe-
nomena of embryonic development in Chaetopterus. Jour. Exp. Zool.,
VOl- 3, p. lds.

1908 The development of the chick. New York.

Loyrz, M. 1905-6 Recherches sur le développment ovarien des oeufs meroblas-
tique 4 vitellus nutritif abondant. Arch. Anat. Mic., tome, 8 p. 69.

Mertens, H. 1893 Recherches sur la signification du corps vitellin de Balbiani
dans l’ovule des mammiféres et des oiseaux. Arch. Biol., tome 13,
p. 389.

Morean, T. H. 1909 A biological and cytological study of sex determination in
Phylloxerans and Aphids. Journ. Exp. Zool., vol. 7, p. 239.

Munson, J. P. 1904 Researches on the oogenesis of the tortoise. Am. Jour.
Anat., vol. 3, p. 311.

Newman, H. H., anp Patrrerson, J. T. 1909 <A case of normal identical quad-
ruplets in the nine-banded armadillo and its bearing on the problems
of identical twins and of sex determination. Biol. Bull., vol. 17, p. 181.

Nicontas, A. 1900 Recherches sur l’embryologie des reptiles. Arch. Anat.
Mic. Tome 38, p. 457. -

OxrxtuacHER, J. 1872 Beitriige zur Entwickelungsgeschichte der Knochenfische
nach Beobachtungen am Bachforellenei. Zeit. f. Wiss. Zool., Bd. 22,

p. 373.

OrrEL, A. 1892 Die Befruchtung des Reptilieneies. Arch. f. mikr. Anat., Bd.
39, p. 215. |

Parrerson, J. T. 1909 Gastrulation in the pigeon’s egg. Jour. Morph., vol.
20, p. 65.

1910 Studies on the early development of the hen’s egg. Part 1.
Jour. Morph., vol. 21, p. 101.

Rasaup, EB. 1908 La position et Vorientation de l’embryon de poule sur le
jaune. Ar. Zool. Exp. et Gen. (Notes et Revue), tome, 9, p. 1.

Riwpie, O. 1911 White and yellow yolk in vertebrate ova. Jour. Morph.,
vol. 22, p. 455.

Ricxert, J. 1892 Die Entwicklungsgeschichte des Ovarialeies bei Selachiern.
Anat. Anz., vol. 7, p. 107.

1899 Die erste Entwickelung des Selachiereies. Festsch. fiir v. Kupf-
fer,. p. 581.

Scuuuze, O. 1899 Ueber das erste Auftreten der bilateralen Symmetrie im Ver-
laufe der Entwickelung. Arch. f. mikr. Anat. Bd. 55, p. 171.

BILATERALITY OF THE PIGEON’S. EGG oii

Somer pE, E. 1905 Les premiers stades de la vitellogenése dans l’ovule de la
poule. Ann. Soc. Med. de Gand., Tome 85, pp. 55-62.

SoNNENBRODT, 1908 Die Wachstumsperiode der Oocyte des Huhns. Arch. f.

mikr. Anat., Bd. 72, p. 415.

Toparo, F. 1895 Beobachtungen und Betrachtungen iiber die Furchung des
Hies und die Bildung der Keimblatter bei Seps chalcides. Untersuch.

z. Naturlehre. Bd. 15, p. 520.

VAN DER Srricut, O. 1902 Le noyau vitellin de Balbiani et les pseudochromo-
somes chez les oiseau. Verh. d. Anat. Ges., 16 Vers., p. 168.

Watpryer, 1870 Eierstock und Ei.

Leipzig.

Wuitman, C. O. 1878 The embryology of Clepsine. Quart. Jour. Mier. Sc.,

vol. 18, pp. 215.

1887 The kinetic phenomena of the egg during maturation and fecunda-
tion. Jour. Morph. vol., 1, p. 227.

1893 The inadequacy of the cell theory of development. Biol. Lec-
tures, Wood’s Hole. Boston, 1894.

Witt, L.
Zool. Jahrb., vol. 6, p. 1.

1893 Beitrage zur Entwicklungsgeschichte der Reptilien.

(Gecko).

‘“Witson, E. B. 1903 Experiments on cleavage and localization in the Nemertine
egg. Arch. f. Entwmech., vol. 16, p. 411.

1904 Experimental studies on germinallocalization. Jour. Exp. Zool.,

Vi0 lee ee lt:
Woops, F. A.

1902 The origin and migration of the germ cells in Acanthias.
Am. Jour. Anat., vol. 1, p. 307

REFERENCE LETTERS

a.o., anterior margin of outer periblast
bld., blastodise

bz., boundary zone of central protoplasm

cap., capillary

cl., cloaca

f.e., follicular epithelium

g.v., germinal vesicle

inf., infundibulum

i.pbl., inner periblastic zone
mg.dc., margin of segmental dise
lat., latebra

m.a., raedian artery

m.v., median vessel

o.a., anterior margin of outer periblast

0.p., posterior margin of outer peri-
blast

pbl., periblast

pbl.r., periblastie ring

p-g., polar granules

p.r., polar ring

p.p-, peripheral protoplasm

s.c., Spherule crescent

s.d., segmental disc

st., stigma

s.z., Spherule zone

ut., ‘uterus’

wdg., wedge

y.n., yolk nucleus

PLATE 1

EXPLANATION OF FIGURES

1 Camera drawings of living primordial follicles in median optical section;
Zeiss 3mm. apoch. imm. objective, ocular 6,480 diaméters. (In polar vaews the .
germinal vesicle is projected into the median plane.) Theconditionsillustrated in
diagram 2 are shown here; in each case the long axis was parallel to the surface of
the ovary and the germinal vesicle was nearer one end of it. A and B are from
ovarian cortex studied in warm salt solution. A. Oocyte, 57 x 43u, polar view,
showing the nucleo-ooplasmic relation. ‘The follicular epithelium (f.e.) surrounds
the oocyte. B.Oocyte, 70 x 63yu, side view. ‘The clear region central to the ger-
minal vesicle indicates the position of the yolk nucleus, which lies accurately in
the polar axis. C, D and # are from an ovary injected intra vitam with neutral
red and Janus green. The yolk nucleus granules were stained deep red and appear
solid blackin the figure. The mitochondria are not represented but they appeared
as very minute granules arranged in fine strings. In this ovary all the primordial
follicles had the long axis relatively much greater than the other two and the ger-
minal vesicle was markedly nearer one end of the long axis. On the other hand
the difference between the two axes of the germinal vesicle is hardly noticeable.
C and £ show a condition that is not uncommon, namely that the polar axis does
not bisect the yolk nucleus (y.n.) and asis usually the case there are more spherules
at one end of the long axis. C. Oocyte, 67 x 52u in side view showing the granules
of the yolk nucleus (y.n.). D. Oocyte, 78 x 58u polar view; g.v., germinal vesicle;
f.e., follicular epithelium; s.c., spherule crescent. H. Oocyte, 75 x 53u in polar
view; y.n., yolk nucleus.

2to7 Cameradrawingsillustrating the development of the follicular blood ves-
sels. From the outset the arrangement is bilateral with reference to the long
axis. All magnified 100 diameters. Zeiss 16.0, oc. 6. Studied and drawn in
creasote as solid objects.

2 Blood vessels of the cortex of the ovary around the primordial follicles.
Portion of an adult ovary injected with india ink, fixed insublimate-acetic-formol,
cleared and drawn in creasote. The heavy circles represent the follicular epithe-
lium of the oocytes, (f.e.) the fine wavy lines the walls of the capillaries, (cap.)
and the dotted lines, the germinal vesicles (g.v.) which are plasmolyzed.

3 Oocytes of 232u at the beginning of the second growth period viewed from the
free (stigmal) pole. The capillary network is spreading over the free hemisphere;
the stigmal area is still broad, but its longer dimension coincides with the long
axis of the oocyte.

4 Oocyte of 244. A,side view; B, polar view. The vascular network is spread-
ing symmetrically with reference to the long axis, on either side of the oocyte.
Other relations as in fig. 3; stigmal area (st).

314

THE BILATERALITY OF THE PIGEON’S EGG PLATE 1
GEORGE W. BARTELMEZ

PLATE 2
EXPLANATION OF FIGURES

5 Oocyte, 400 x 376, at the beginning of the period of differentiation; oblique
view from the stigmal pole. The vascular network completely invests the oocyte
except in the region of the stigma (st.) which extends in the long axis. A direct
view in the polar axis showed the germinal vesicle near, but not quite at, the center
of the long axis.

6 Oocyte, 902x829. Period of differentiation. Side view showing the differ-
entiation of the median vessel (m.v.). Note the germinal vesicle (g.v.) almost
central in position but slightly nearer to the animal pole, which is up, and to the
infundibular end of the long axis, to the left.

7 Oocyte, 5.47 x 5.14 mm. magnified 15 diameters, at the end of the period of
differentiation, showing the stigma (s¢.) in the long axis, and the median artery
in somewhat oblique side view. The oocyte is free to rotate within the follicle
during this period and the germinal vesicle (g.v.) surrounded by segmental disc
and periblast (pbl.) lies at one end of the long axis. This is an extreme case of
rotation; usually the rotation is around the long axis. Fig. 34 shows a stigmal
view of this oocyte.

8 Animal pole of an oocyte of 3.2 mm. in median longitudinalsection, 100
diameters, Zeiss 16.0, oc. 6. An early stage in the formation of the blastodise.
The germinal vesicle (g.v.) is surrounded by the fine granules of the segmental
dise, and this in turn by the periblast (pbl.). The peripheral protoplasm (p.p.)
is now anarrow zone. At the center of the germinal vesicle is the group of chro-
mosomes and nucleoli.

316

THE BILATERALITY OF THE PIGEON’S EGG PLATE 2
GEORGE W. BARTELMEZ

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

317

PLATE 3
EXPLANATION OF FIGURES

All the figures except 17 are from photographs of sections of ovaries fixed in
sublimate-acetic-formol mixtures, cut 10u thick in paraffin and stained either with
neutral gentian or with neutral safranin-siure violett. All were taken with a
Zeiss 4mm. apochromat, ocular 4 and magnified 350 D. The germinal vesicle is
plasmolyzed more or less in every case. In each oocyte the ‘typical’ relations of
diagram 2 appear but the critical evidence for these conclusions was obtained from
fresh and creasote preparations and from free-hand sections. 'The measurements
are taken from the sections; the dimensions in life are estimated from 5 to 10 per
cent greater.

9 Primordial follicle, 52x 44y, cut in the longand polar axes. y.n., yolk nucleus,
surrounded by the spherules which appear as vacuoles.

10 Primordial follicle, 78 x 624, Horizontal section cut perpendicular to the
polar axis, near the animal pole. Note the germinal vesicle nearer one end of the
long axis.

11 Primordial follicle, 83 x 604, cut in long and polar axes and showing the
stigma (st.) already formed in the long axis.

12. Primordial follicle, 88 x 72u, imbedded in the stroma so that the stigma has
not yet formed. Cut in long and polar axes.

13 Primordial follicle, 95 x 834 cut in long and polar axes showing the relations
of germinal vesicle (g.v.), yolk nucleus (y.n.) and spherule crescent (s.c.); f.e.
follicular epithelium.

14 Oocyte at the beginning of the second growth period, 103 x 100u, showing the
yolk nucleus material increasing in amount.

15 Oocyte 277u in long axis; second growth period. Cut in the polar axis,
transversely to the long axis, showing the stigma (sf.) in cross section and the
increase of the spherules (s.c.) The germinal vesicle is still peripheral and the
substance of the yolk nucleus (y.n.) has increased greatly in amount and is
spreading irregularly through the ooplasm.

16 Oocyte, 203 x 1664; second growth period. Cut in the long and obliquely
to the polar axis. A later stage than fig. 15; from another ovary. The germinal
vesicle is no longer peripheral and the chromidial substance is diffusing through
the ooplasm.

17 Oocyte, 198 x 1624, cut in the long and polar axes. Photographed from a
free hand section of an ovary fixed in formol-bichromate and stained with Sudan
11. The lipoid spherules, (s.c.) appear black, the lighter central region is the yolk
nucleus which is enlarging. The peripheral protoplasm (p.p.) is free from
spherules.

18 Oocyte 280 x 255; magnified 325 diameters and cut almost perpendicular to
the polar axis. The section passes through the center of the germinal vesicle,
which is distinctly nearer to one end of the long axis. <A cross section of one of
the ‘Balken’ appears at y.n. and shows clearly the diffusion of the chromidial
(?) granules.

318

PLATE 3

319

EW. BARTELMEZ

o
2)
a
m
%
: +
_
3 7
g ie
: 5
: g
cS)
»
é
5

GEO!

JOURNAL OF MORPHOLOGY,

Ke
THE BILATER

PLATE 4

EXPLANATION OF FIGURES

Photographs from the same series of sections used for plate 3.

19 Oocyte, 293 x 270u, cut in the long and polar axes. Zeiss 4 mm., oc. 4,
reduced to X 280. From a free hand section stained with Sudan m1. The spher-
ules are disappearing from the central protoplasm and one can begin to distinguish
a spherule zone (s.c.) within the peripheral protoplasm (p.p.). The size relations
between germinal vesicle (g.v.), and ooplasm are noteworthy in comparison with
the oocytes sectioned in paraffin e.g., fig. 20. ,

20 Oocyte, 354 x 324 4; magnification as in fig. 19. End of second growth period
Cut in long and polar axes. The persistence of the center of the yolk nucleus
(y.n.), at this stage is a condition found only occasionally.

Period of differentiation.

Figs. 21 to 25 were taken with a Zeiss 8 mm. apochromat, oc. 2, and reduced to
x 80.

21 Oocyte, 512 x 400u, cut near the animal pole. Photographed from a frozen
section, 50u thick, stained with Sudan. At this stage the spherules form a zone,
(s.z.), near the periphery, but with a higher power the clear peripheral protoplasm
ean be distinguished. The alveolar appearance is an artefact due to freezing.
Only the tip of the germinal vesicle (g.v.) appears.

22 Oocyte, 762 x 710u, cut perpendicular to the polar axis, showing the ger-
minal vesicle nearer one end (left) of the long axis. The central and peripheral
protoplasm (p.p.) can be distinguished with the spherule zone between them.

23 Oocyte, 786 x 762u, cut in the long and polar axes, in about the same stage
as fig. 6. At the right the section passes to one side of the stigma (si.). The
spherule zone (s.z.) can be seen and in addition many large definitive yolk granules
have appeared throughout the ooplasm.

24 Oocyte, 1.06 x 0.89 mm., cut in long and polar axes. The germinal vesicle
has migrated peripherally out of the central protoplasm, the boundary of which is
marked bz.

25 Oocyte, 1.17 in long axis, cut in the polar axis and obliquely to the long axis.
The photograph was taken from a section nearer to one end of the long axis. The
central and peripheral protoplasm and a single spherule zone may be distinguished
(more clearly on the right). The definitive yolk granules are especially numerous
peripherally. 6.z, boundary of the central protoplasm.

320

PLATE 4

THE BILATERALITY OF THE PIGEON'S EGG

GEORGE W. BARTELMEZ

JOURNAL OF MORPHOLOGY, VOL, 23, No 2

321

PLATE 5

EXPLANATION OF FIGURES

26 Oocyte, 1.74 x 1.44 mm., cut approximately in the long and polar axes, but
much distortedin sectioning. The germinal vesicle is peripheral. The location of
the future latebra is indicated by the finer meshwork at the center of the oocyte.
Taken with a one-half inch B. and L. objective, X 40. p.p., peripheral protoplasm.

27 Oocyte, 2.5mm. in long axis (life), x 40, Leitz obj. no. 2. Cut in polar axis,
perpendicular to the long axis and compressed from right to left in cutting. The
follicular epithelium has shrunken somewhat from the theca folliculi, and there is
a shrinkage tear near the germinal vesicle. The region of the latebra (lat.) is
clearly defined at the center and the layers of yolk granules indicate the rhythmical]
formation of yolk. The stigma is shown in cross section at sé.

322

THE BILATERALITY OF THE PIGEON’S EGG PLATE ;
GEORGE W. BARTELMEZ

ee

Ney & i. Ag -
*

a

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

323 ri

PLATE 6

EXPLANATION OF FIGURES

The early development of the blastodise

28 The animal pole of the 1.74 mm. oocyte shown in fig. 26. Zeiss apochromat
8.0 mm. ocular 2. X 100. The finer reticulum about the germinal vesicle (g.v.)
is the first trace of the blastodise which differentiates symmetrically about the
germinal vesicle (see fig. 8).

29 An oblique section of the blastodise of an oocyte of 5.47 mm. (see fig. 7. for
surface view). Leitz objective 2. x 55. Marginal (pbl.) and central periblast
can be distinguished about the finely granular segmental disc. The more deeply
staining wedge granules (wdg.) have differentiated at the periphery of the dise.
The guide line g.v. runs to the group of chromosomes and nucleoli at the center of
the germinal vesicle. P.g., polar granules. A narrow break on the left sepa-
rates the wedge granules from the peripheral protoplasm. Final stage of period
of differentiation.

30 The blastodise of an oocyte 8 mm. long axis in life. Beginning of the final
growth period. Magnification as in fig. 29. The wedge (wdg.) extends from the
periblast to the germinal vesicle and the group of polar granules at the edge of the
latter is well marked at p.g. Note the abrupt transition of the periblast into the
surrounding bed of white volk; there is only a trace of a marginal periblast in this
ege.

324

THE BILATERALITY OF THE PIGEON’S EGG PLATE 6
GEORGE W. BARTELMEZ

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

9O£
32%

LIANE Fi
EXPLANATION OF FIGURES

31 to 34 are from entire follicles photographed in creasote.

31 Follicle 1.1 mm. in long axis. X 15. The focus was changed during the
exposure so as to show both the upper surface and the median optical section.
The attached pole is up, the long axis extends right andleft. The median arteries
(m.a.) appear on either side of the long axis (the upper one only is in focus.)

32 View of the stigmal pole of an oocyte 2.3 mm. in long axis, showing the two
median arteries ending at the stigma, the boundaries of which do not appear
clearly. X< 5:

33 Side view of a follicle, 5.1.x 4.7x4.8mm. ™X 8. Focus on upper side and
on median plane, to show the bilateral arrangement of the median artery (m.a.)
with reference to the long axis.

34 Oblique stigmal pole view of the 5.47 mm. oocyte shown in fig. 7. There are
no capillaries in the region of the stigma (s/.)

35 Surface view of the blastodise of an ovum taken from the middle third of the
oviduct at 11 p.m. and photographed by reflected light in the fixative (sublimate-
acetic-formol. X 6.6. Fragmentation stage of the segmental dise (s.d.), i.e.,
an early stage in the copulation of the pronuclei. The disc and periblastic zones
are clearly elliptical. pbl.r., periblastic ring; o.p., posterior margin of outer peri-
blast.

36 Oblique surface view of the segmental disc of an egg taken from the middle
third of the oviduct at 11 p.m. and photographed by reflected light in the same
fixative as fig. 36, with one tube of a Zeiss binocular, obj. ao., oc. 2. X 15. The
egg was ina slightly later stage than that shown in fig. 35 and the higher magnifi-
cation brings out the fragmentation clearly.

37. Two follicles, a day before ovulation, in situ. A trifle larger than natural
size. The long axes of the follicles are oriented in the antero-posterior axis of the
bird. The median longitudinal section of the larger follicle is shown in fig. 37;
c.l, cloaca; ut., ‘uterus;’ ¢nf., infundibulum.

326

THE BILATERALITY OF THE PIGEON’S EGG PLATE 7
GEORGE W. BARTELMEZ

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

3y4if

PLATE 8

BXPLANATION OF FIGURES
38 The median surface of an ovarian egg twenty-four hours before ovulation,
cut in half through the long and polar axes, and photographed by reflected light.
«8. The central latebra (lat.) is distinetly farther from the right hand end of the
long axis, the end, namely, which was directed posteriorly (‘cloacally’) when the
bird was opened. ‘The neck of the latebra extends up towards the blastodise at
the periphery. Compare with diagram 2 # (in which the concentric layers of
yolk are not represented).

39 A photograph similar to the one shown in fig. 38, from an oviducal egg taken
twenty-six hours after fertilization, showing the common type of latebra of ovi-
ducal eggs. The section did not cut the blastodise exactly in the center and so
the infundibular extension of the latebra does not appear, but it is indicated by the
arrangement of the surrounding layers of yolk. When this is considered and
allowance is made for the cracking of the yolk at the infundibular end, the center
of the latebra is 2mm. nearer this end of the long axis. The hole in the yolk at
the right was made by the pin inserted to mark the cloacal end of the egg when it
was obtained.

40 A photograph of a parasagittal section of the blastodise of an oocyte, 17.8
x 16.9 x 16.9 mm., taken about 26 hours before ovulation, fixed in sublimate-
acetic and stained with neutral gentian. Leitz obj. 3, * 56. The ‘wedge’ gran-
ules (wdg.) form a collar around and tapering toward the germinal vesicle (g.v.),
which is immediately surrounded by the polar granules, (p.g.). 0.a. and o0.p. mark
the anterior and posterior margins of the outer periblastic zone of the blastodise.

11 A photograph of a parasagittal section of an egg with the first cleavage
spindle formed, taken from the entrance to the shell gland at 11 p.m., fixed in sub-
limate-acetic and stained with neutral gentian. Leitzobj. 8, X 48. The wedge
(wdg.) is clearly defined posteriorly and absent anteriorly. The periblastie ring,
(pbl.r.) is of the diffuse type.

42. Surface view of the blastodise of an oocyte removed from its follicle about
forty-five hours before ovulation would have occurred normally. A copy of a
drawing made from the living ovum. 10. At the center is the germinal vesicle
(g.v.), with the narrow opaque (white) zone of polar granules (cf. fig. 40, p.g.). The
segmental dise is circular and appears light; at its outer margin the transparent
(dark) inner periblastic zone (7.pbl.) is differentiating. At the outer margin of the
periblast are light spots which represent groups of periblastic granules. These
have not yet appeared anteriorly at a.o. The antero-posterior axis indicated by
this difference in the structure of the periblast formed an angle of 65 to 70 de-
grees to the long axis of the oocyte.

43° The surface view of the blastodise of a fertilized egg taken from the infun-
dibular third of the oviduet and drawn after one and one quarter hours of ineu-
bation. Magnified ten times. A copy of one of a series of drawings of the living
egg showing the formation of the periblastie ring (pbl.r.) (ef. fig. 35).

328

THE BILATERALITY OF THE PIGEON'S EGG

GEORGE W. BARTELMEZ

JOURNAL OF MORPHOLOGY, VOL.

ABGrecdain del

329

- MIG dC
-- poli

PLATE 8

I. A FURTHER STUDY OF THE CHROMOSOMES OF
THE REDUVIIDAE. II. THE NUCLEOLUS IN THE
YOUNG OOCYTES AND ORIGIN OF THE OVA IN
GELASTOCORIS

FERNANDUS PAYNE

From the Zoological Laboratory of Indiana University}
TEN FIGURES

I. A FURTHER STUDY OF THE CHROMOSOMES OF THE REDUVIIDAE

Since describing several irregularities of chromosome distribu-
tion in this family (’09 and ’10), I have been seeking further
information in regard to the origin of such irregularities. I
expressed the view (’09) that these irregularities probably arose
by the large idiochromosome breaking up into two, three, four, or
five elements. While I had no authority for believing it, I looked
upon these changes as recent ones, and hoped to find a species in
which such a change was taking place. During the past summer,
I made a collecting trip through the West and South. A study
of this material has thrown some doubt on this view. ! have
five species of Sinea (diadema, rileyi, confusa, complexa, spinipes)
and they are distributed from Massachusetts to California.
Four of them (spinipes, confusa, complexa, and diadema) show
the same number and size relations of the chromosomes and the
same type of distribution as Sinea diadema (Payne, 709). From
this fact it would seem that both the number of chromosomes
and the type of distribution existed before the genus split up into
these four species and before it became distributed over such a
wide area. Otherwise it is rather difficult to explain this con-
staney in number and behavior of the chromosomes in the four
species. On the other hand a study of the chromosomes of Sinea

1 Contribution No. 124.
331

JOURNAL OF MORPHOLOGY, VOL, 23, NO. 2

Son FERNANDUS PAYNE

rileyi, which presents a type of distribution similar to that of
Acholla multispinosa, might lead us to the opposite conclusion.
If there were present in the parent species a type of distribution
similar to that of Sinea diadema we would have to assume a fur-
ther splitting up of these chromosomes in order to reach the con-
dition found in Sinea rileyi. Such an assumption would only
complicate matters. In this case it would seem that this species
split off from the parent one before any breaking up of the single
large idiochromosome had occurred, assuming, of course, that
a single pair of idiochromosomes was the original condition. In
all probability then, the breaking up of the large idiochromosome
has occurred independently in the parent species and in Sinea
rileyi and, of course, may have occurred at any time since the
origin of the species.

As the type of chromosome distribution found in Sinea rileyi
has been described in only one other species it seems worth while
to describe it somewhat in detail. Unfortunately, I have a
small amount of material and it does not show spermatogonial
or oogonial divisions. However, first and second spermatocyte
divisions are in abundance and, I think, indicate clearly that the
type of distribution is similar to that of Acholla multispinosa
(Payne, 09 and 710). The first spermatocyte division (fig. 1,
A and B) shows eighteen chromosomes, six of which are smaller
than the remaining twelve. All divide in this division so that
each secondary spermatocyte receives eighteen chromosomes.
There is no definite arrangement of the chromosomes in the first
division but the characteristic regrouping, found in the remainder
of the Reduviidae, occurs in the second division. The twelve
larger chromosomes are arranged in an irregular ring with the
six smaller ones forming a hexad group in the middle. Five of
these six lie in one plane while the other one lies in a different
plane either above or below the five. Figure 1, D and £ are pole
views of this division showing the twelve chromosomes in the
ring and the five which lie in one plane in the middle. Figure 1, F
is the same as #7, but drawn at a different focus to show the one
chromosome in the middle instead of the five. Figure 1, G is a
slightly oblique view showing the hexad group. Figurel, H is a

CHROMOSOMES OF THE REDUVIIDAE ooo

side view and shows clearly the relative positions of the X and Y
elements. Only four of the five chromosomes show in this figure.
Although no anaphases are present in my material, I think we
may safely conclude, from what has been found in the rest of the
family, that the twelve chromosomes in the ring divide equally,
while the five members of the hexad group which le in one plane
pass to one pole undivided and the one below or above passes

&
. oo: @®@ Ps « or @
@. - e, @ ®& e
@ee, e° e. 838
e © ee ®B e@ e ; é oon ta 8 Code g
é
Og 8. Ger: G H

Fig. 1 Sinea rileyi Montd. A and B, metaphase plates of the first spermato-
cyte division showing eighteen chromosomes, six of which are noticeably smaller
than the remaining twelve; C,a side view of an anaphase of the first spermatocyte
division showing that the small chromosomes divide in this division; D and E,
metaphase plates of the second spermatocyte division with twelve chromosomes
in the ring and the five chromosomes (X element of Wilson) which lie in one plane
in the middle; F is the same as F except it is drawn at a different focus to show the
single chromosome (Y element); G is a slightly oblique view of the second division
to show the hexad group in the middle; H is a side view of the second division and
shows the relative positions of the X and Y elements (only four of the five chromo-
somes show in this figure). > 2275.

to the opposite pole undivided. This would give two classes of
spermatozoa, one with thirteen and one with seventeen chromo-
somes and the fertilization formula would be as follows:

SPERM EGG
13 ae 17 = 3007
17 =F 17 = 349

334 FERNANDUS PAYNE

The most striking difference between the chromosomes of Acholla
and Sinea is in the size relations of the idiochromosomes. In
Sinea all six of them are practically the same size while in Acholla
three of them are very small, two intermediate and one (the Y
element) very large.

A study of the chromosomes of Pnirontis modesta Banks has
revealed a type of distribution new to the family and similar to
that described for Gelastocoris oculatus (Payne,’09). Ihave only
one specimen, but I believe it shows sufficient stages to justify
the above conclusion. Figure 2, A, B, and C show the spermato-
gonial group with twenty-five chromosomes, three of which are
very small. In the first spermatocyte division (fig. 2, D and F),
there are fifteen chromosomes. Again three of these are very
small. Of the remaining twelve, ten are larger and two inter-
mediate between the large and small ones. There is no definite
arrangement of the chromosomes and all divide in this division
so that all the secondary spermatocytes receive fifteen chromo-
somes. In the second division there is the characteristic regroup-
ing found in all the Reduviidae. The ten large chromosomes form
a more or less irregular ring with the two intermediate and three
small ones forming a pentadgroup in the middle. This group was
somewhat difficult to analyze on account of the small size of the
chromosomes and the fact that they lie so close together. The
most favorable metaphases of the second division (fig. 2, / and
G) show, however, that four of these five, the three small and one
intermediate lie in one plane while the other intermediate one
lies below or above them. Figure 2, H, the best side view obtainable
of the second division, shows this one chromosome and its relative
position with respect to the other four, which in this figure are
massed together as one. Although no anaphases are present
showing the method of this division, I think there can be little
doubt, judging by the number relations in the spermatogonial
and first spermatocyte divisions and also from the analogy here
to the second spermatocyte divisions in the other Reduviidae,
that the ten chromosomes in the ring divide equally, while the
four members of the pentad group (X element), which lie in one
plane, pass to one pole undivided and the other one passes un-

CHROMOSOMES OF THE REDUVIIDAE 330

divided to the opposite pole. This gives two classes of sperma-
tozoa, one with fourteen chromosomes and the other with eleven,
and the fertilization formula would be as follows:

SPERM EGG
11 —- 14 = 250"
14 + 14 = 289

The chromosomes of Pselliodes cinctus Fabr. also deserve men-
tion. Asshown in figure 3, the type of distribution is similar to

ose e042 OD sires

§ 0.50 ise ce ty 0% Ae @°@

‘oe @ 0% °, o° 0, 9 @

A @® oe C @ >
.°,@
E o 3% e Se o°e
*o,® Soe a :¢
E F @ fe

Fig. 2 Pnirontis modesta Banks. A, B and C, spermatogonial divisions show-
ing twenty-five chromosomes; D and #, first spermatocyte divisions with fifteen
chromosomes (the five small ones are the idiochromosomes); /’, second spermato-
cyte division, showing ten chromosomes in the ring and the one idiochromosome
(Y element) which lies in a different plane from the other four; G, second sper-
matocyte division with ten chromosomes in the ring and the four idiochromo-
somes which lie in the same plane; H, side view of the second maturation division
showing the single idiochromosome below and the four above massed together as a
single body. X 2275.

that described for Prionidus cristatus and Sinea diadema (Payne,
09). The size relations of the idiochromosomes, however, are
different. In Sinea all four were practically the same size; in
Prionidus one was slightly larger than the other three; while
in Pselliodes the single chromosome, the homologue of the small
idiochromosome (Y element), is much larger than the others;

336 FERNANDUS PAYNE

so much larger, in fact, that it undoubtedly contains a larger
quantity of chromatin than the other three combined. In this
respect. Pselliodes resembles Acholla multispinosa (Payne, ’09
and 710).

The paper of Della Valle (09) again brought to the surface the
question of the variation of the number of chromosomes within
a given species and individual. This paper has been ably dis-

e@ , ae
eer ~ i
atees’ | @e
eo. .80 oe” Peo? © &

iz B ee c
GN ee ae
HII I, ® \\\\ -
é i a? : oo $ e® @ e@
Wine @ * eee
NYY D e@® E $e F

Fig. 3 Pselliodes cinctus Fabr. A, oogonial division with thirty chromosomes ;
B, spermatogonial division with twenty-eight chromosomes; C, first spermatocyte
division with sixteen chromosomes; D, side view of the second spermatocyte divi-
sion, showing the four idiochromosomes as a tetrad group in the middle and the
relative size of the four; # and F’, metaphase plates of the second spermatocyte
division, H showing the three small idiochromosomes in one plane in the middle,
and F showing only the single large idiochromosome in the middle. X 2275.

cussed by Montgomery (710) and Wilson (10). These authors,
while they admit Della Valle has done a good service in collecting
a large amount of data against the prevalent notion of chromo-
somal constancy, believe he has been unjust in his criticisms and
that his evidence is insufficient for his sweeping conclusions. It
is not my intention to discuss any of these papers, but to describe
a case of apparent variation and give what I think to be its
explanation.

CHROMOSOMES OF THE REDUVIIDAE oot

Apiomeris crassipes is one of the Reduviidae in which we find
a single slightly unequal pair of idiochromosomes. The spermato-
gonial number is twenty-four, the first spermatocyte thirteen and
the second spermatocyte twelve. I made twelve counts of the
first division and thirty-four of the second. These counts were
made from metaphase plates which were perfectly flat and in
which the chromosomes were well separated. The counts of the
first division showed a range from thirteen to sixteen; one with
thirteen, six with fourteen, four with fifteen, and one with six-
teen. Figure 4, A, B, C, D, EH and F are metaphase plates showing
these variations and I think that even the most optomistic will
admit that these figures seem to show a real variation. Things
look somewhat differently though when such figures are viewed
from the side (fig. 4, G,H and J). These figures show clearly
that some of the bodies, which in pole view appear to be chromo-
somes, are not really chromosomes at all but in all probability
are yolk granules. At any rate, they do not behave as chromo-
somes and, it seems to me, behavior is about the best test for a
chromosome. ‘To be sure, chromosomes differ in their behavior
and within the last few years several types of behavior have been
described, but these bodies do not behave like any of the described
types. The chromosomes are clearly bipartite. The granules
are spherical and never at any stage show signs of constriction.
They may or may not lie in the same plane as the chromosomes.

Counts of the second maturation division also show similar
variations. The normal number in this division is twelve. Out
of thirty-four counts, there were six with twelve, eleven with
thirteen, six with fourteen and eleven with fifteen bodies resem-
bling chromosomes. Some of these variations are shown in figure
4,J and L. Side views (fig.4, M and NV) of these metaphases again
show the granules as spherical unconstricted bodies.

I have no anaphases of Apiomeris showing conclusively that
these granules do not divide, but have made some counts of
the second maturation division in Conorhinus where anaphases
are present. ‘The normal number of chromosomes in this division
of Conorhinus is thirteen, but since one of the triad group in the
middle les below the two, only twelve show. One hundred

338 FERNANDUS PAYNE

®
28 ee | °e | Bae
Pes 6 @ *e
| e@
@ = ° o@ =.
. ©. 0@ ®
é Ceo See
rie a oO aE @ C
ee° » A ;
ee’ @ ». - \ 006%
® @ e® ® @
bs = fp a9 E ®. @ F

Fig. 4 Apiomeruscrassipes. A, B, D, E and F, pole views of the first sperma-
tocyte division showing an apparent variation in the number of chromosomes (the
bodies marked ‘g’ are probably yolk granules); C, first spermatocyte division
showing the normal number; G, H and J, side views of the first spermatocyte divi-
sion showing the chromosomes constricted and the spherical yolk granules; J, K
and L, pole views of the second spermatocyte division showing an apparent varia-
tion in number (K shows the normal number); M and N, side views of the second
spermatocyte division, showing the granules among the chromosomes. XX 2275.

CHROMOSOMES OF THE REDUVIIDAE 339

and eighty-five counts gave one hundred and fifty cases with the
normal number and thirty-five with thirteen and fourteen. Figure
5, A, B, C and D show polar views of such metaphases. In C and
D the granules are near the periphery of the cell and perhaps in
these cases would not be taken for chromosomes, /, Ff, G and H
show clearly that these granules may lie in any plane and that
they do not divide.

Fig. 5 Conorhinus sanguisugus Lec. A and B, metaphase plates of the second
spermatocyte division showing a single granule (g) lying just outside the ring of
chromosomes; C and D, metaphase plates of the second maturation division show-
ing two granules near the periphery of the cells; H, metaphase plate, side view
showing two small granules; Ff, Gand H, anaphases, side view, of the second matu-
ration division, showing that the granules do not divideand may pass into either
daughter cell. X 2275.

In these cases, then, what might be termed chromosomal
variation is not variation at all, but is due to the presence of
yolk granules which may happen to lie in the metaphase plate.
It is a significant fact that in these counts not a single one fell
below the normal number. ‘This is also illustrated in Reduvius
personatus where ninety-eight counts of the first maturation divi-
sion gave ninety-six with the normal number (twelve) and two
with thirteen. Eighty-four counts of the second division gave

340 FERNANDUS PAYNE

seventy-seven with the normal number (eleven) and seven with
twelve. If this bea real chromosomal variation why should the
variations always be greater than the normal number and never
less?

I have included two figures of Gelastocoris (fig. 6, A and B)
where a large number of undoubted yolk granules are present.
A is a prophase of the first maturation division before the nuclear
wall breaks down and B a metaphase plate showing how difficult it
is, looking at a single cell, to differentiate between granules and
chromosomes.

Fig. 6 Gelastocoris oculatus Fabr. A, prophase of the first spermatocyte divi-
sion showing many yolk granules outside the nucleus and their close resemblance
to chromosomes; B, metaphase plate, first spermatocyte division showing the chro-
mosomes in the middle and the granules near the periphery of the cell and also
the close similarity between the two. X 860.

II. THE NUCLEOLUS IN THE YOUNG OOCYTES.AND ORIGIN OF THE
OVA IN GELASTOCORIS

In arecent paper (711) Foot and Strobell describe a chromosome
nucleolus in the oogonial cells of Protenor and also describe the
origin of the two large idiochromosomes from it at the time of
mitotic cell division. At the time this paper appeared, I was
working on the nucleolus in the ovaries of Gelastocoris oculatus.
While the two forms have some things in common there are several
points of divergence. As stated, the nucleolus in Protenor is
confined to the oogonial cells. This is not the case in Gelastocoris.
Here it appears after the last oogonial division, in the early oocyte,
and persists until shortly after synapsis. Figure 7, A is an oogonial
cellin which there isno indication of anucleolus. Figure7, B shows
an early oocyte and the beginning of the formation of the nucleo-

CHROMOSOMES OF THE REDUVIIDAE 341

lus. In this early stage it seems to be pretty conclusive that the
nucleolus is formed by an aggregation of chromatin. This agrees
with Foot and Strobell’s interpretation of the formation of the
nucleolus in Protenor. Figure 7, C, D and E are later stages in
the development of this nucleolus. These figures show that it
increases in size rather rapidly and becomes large in proportion
to the size of the nucleus. This rapid growth along with the fact

Fig. 7 Gelastocoris oculatus Fabr. A, an oogonial nucleus showing no nucle-
olus is present; B, a young oocyte nucleus in which some of the chromatin is col-
lecting together to form the nucleolus (nwc.);C and D, young oocyte nuclei, show-
ing the growth of the nucleolus; HL, young oocyte nucleus showing the nucleolus at
approximately its maximum size (in this case the nucleolus is not stained so
intensely as in the others and eight darker bodies lie imbedded in it) ; 7, an oocyte
nucleus shortly after synapsis with the chromatin as doubly split threads and the
nucleolus much reduced in size. XX 2275.

that no chromatin is added after its beginning indicate that its
rapid growth is due to the addition of something other than chro-
matin. Figure 7, E also indicates the same thing since here the
nucleolus does not stain uniformly but darker bodies can be seen
within it. In this case there are eight of these darker bodies and,
while the evidence is not conclusive, it seems probable that they
are the idiochromosomes and the whole structure is a nucleolus
or plasmosome in which the idiochromosomes are imbedded.

342 FERNANDUS PAYNE

So striking is the resemblance to the condition described in the
growth period of the spermatocytes of Pironidus (Payne, ’09)
where beyond a doubt, the idiochromosomes are imbedded in a
plasmosome, that this conclusion seems very probable. At the
time of synapsis or shortly after (fig. 7 F), the nucleolus is much
reduced in size and later it completely disappears (fig. 8 B, C, D,
E and F, serial sections of a single young oocyte). Foot and
Strobell believe, on account of size relations, that the nucleolus
in Protenor represents something more than the two large idio-
chromosomes. Is it: not possible that here also, the nucleolus
is one in which the chromosomes are imbedded and that it dis-
appears at the time of cell division, leavong only the chromosomes?
The main difference then between the nucleolus in the ovaries of
Protenor and Gelastocoris is the period during which it persists.

In this same paper Foot and Strobell describe the terminal
chambers of the ovaries of Protenor as differentiated into three
distinct zones. They designate these zones as A, B and C, A
being the apex of the terminal chamber. B is the middle zone and
is characterized by large nuclei which vary in form and structure
and which stain intensely with chromatin stains. Zones A and C
are somewhat alike, the nuclei being smaller than in zone B and
staining less intensely. These same three zones have been
described by other workers (Will, ’85; Korschelt, ’86, and Preusse,
95). So far all these workers agree. They further agree that
the larger nuclei of zone B arise by a process of growth from the
nuclei of zone A. The difference of opinion arises as to the inter-
pretation of the nuclei of zone B and the origin of the nuclei of
zone C and hence the origin of the ova. Korschelt holds that the
nuclei of zone B are purely nourishing nuclei which disintegrate
to form food for the developing ova, and that the nuclei of zone
C' arise by a continuation of the nuclei unchanged from zone A.
On the other hand Foot and Strobell (’11) with Will (’85) claim
that the nuclei of zone B are not all nourishing nuclei and that
the nuclei of zone C arise in the main by fragmentation or amitotic
division from the large nuclei of zone B.

While the end chambers in the ovary of Gelastocoris are not
divided into three distinct zones, the material is very favorable

CHROMOSOMES OF THE REDUVIIDAE 343

Fig. 8 Gelastocoris oculatus Fabr. A, young oocyte just as it starts down the
tube and as the cytoplasm is beginning to collect about the nucleus (the chromatin
and nucleolus in this nucleus are in practically the same condition as in fig. 7, F);
B, C, D, E and F, serial sections of the same oocyte (a little older than A) showing
no nucleolus is present at this time. A, X 2275; B, C, D, EF and F, X 860.

A and B, longtitudinal sections through

Gelastocoris oculatus Fabr.
two terminal chambers of two tubes from the same ovary. These sections are

Fig. 9

They also show that

Xx 467.

taken from a young ovary and illustrate the gradual transition of the small nuclei

of the apex into the large food nuclei and into the oocytes.
the young oocytes (S) are scattered among the food nuclei and that there are not

three distinct zones of nuclei as in Protenor.

344

CHROMOSOMES OF THE REDUVIIDAE 345

for the study of the continuity of the ova from the tip of the end
chamber tomaturity. Figure9,A and Bare two drawings of theend
chambers from a young ovary. It will be seen that both show two
zones distinctly, the apex of small nuclei corresponding to zone
A and the rest of the end chamber composed principally of large
nuclei corresponding to zone B of Protenor. ‘There is no dis-
tinct region, however, which can be called zone C. A few small
nuclei (fig. 9,A and B) are found near the lower end of the region
which corresponds to zone B and it is possible that these nuclei
are the representatives of zone C, although they are not confined
to a definite region asin Protenor. As to the origin of these small
nuclei in Gelastocoris, I am not quite certain. I am willing to
grant that they may arise by migration, unchanged, from the apex
or that they may arise by fragmentation of the large nuclei. In
this case it is of little concern where they arise,as the evidence in
Gelastocoris proves conclusively, I think, that the ova do not arise
from these small nuclei. Figures 9, A and B, and 10 show a
number of nuclei in the synaptic stage (S). That these nuclei in
this stage are the true oocytes is demonstrated in figures 7, / and
8,A. Figure 8, A shows a young oocyte at the base of the terminal
chamber and as it starts down the tube (similar in position to the
oocyte ov in fig. 10). The cytoplasm is just beginning to form
around the nucleus and the chromatin is in practically the same
condition as in figure 7, /,where the double threads are coming out
of the contraction phase. This clearly demonstrates that the
oocyte starts down the tube shortly after the synaptic stage.
Hence my conclusion that all the nuclei in the synaptic stage are
young oocytes. Asis shown in figure 9, B, these nuclei in the syn-
aptic stage (S) are scattered among the large nourishing nuclei,
extending up as far as the border line between the large nuclei
and the small nuclei of the apex. This figure along with figure 7
shows that these nuclei undoubtedly arise by the growth of the
small oogonial nuclei at the apex and hence in this form there is
no break in the continuity of the cells from the oogonial stage to
the fully developed ova.

While the evidence in Gelastocoris warrants the above con-
clusion, I by no means wish to generalize and say that this must

yw
i)

A longitudinal section through the ter-

Fabr.
minal chamber of a tube from an ovary in which there were eggs approaching

maturity. This section shows young oocytes (S) in the synaptic stage and an
older oocyte (OV) as it starts down the tube.

Fig. 10 Gelastocoris oculatus
about it.

Note the small nuclei aggregated

x 467.

346

CHROMOSOMES OF THE REDUVIIDAE 347

be the case in all forms. However, it seems to me that Will, and
Foot and Strobell have failed to demonstrate conclusively that
the ova arise from the small nuclei of zone C. At least I do not
believe they have figured uninterrupted stages of transition from
one to the other and unless it be actually demonstrated that
these small nuclei give rise to developing ova, it matters not so
far as theories of heredity and chromosomal continuity are con-
cerned, whether they arise by mitosis or amitosis. Further, I
do not believe they have exhausted the possibility of the young
oocytes being present among the large nuclei of zone B.

BIBLIOGRAPHY

Deuua VALLE, P. 1909 L’organizzazione della cromatina studiata mediante il
numero dei cromosomi. Archivio Zoologica, tom. 4, no. 1.

_ Foot, K., AnD StrRoBELL, E.C. 1911 Amitosis in the ovary of Protenor belfragei
and a study of the chromatin nucleolus. Archiv fiir Zellforschung,
Bad: 7, no: 2:

KorscuEeut, E. 1886 Uber die Entstehung und Bedeutung der verschiedenen
Zellenelemente des Insektenoviariums. Zeitschr. fiir wiss. Zool., Bd.
43, no. 4.

Montacomery, T. H. 1910 On the dimegalous sperm and chromosomal varia-
tion of Euschistus, with reference to chromosomal continuity. Archiv
fiir Zellforschung, Bd. 5, no. 1.

Payne, F. 1909 Some new types of chromosome distribution and their relation
to sex. Biol. Bull., vol. 16, nos. 3 and 4.

1910 The chromosomes of Acholla multispinosa. Biol Bull., vol.
18, no. 4.

Preusse, F. 1895 Uber die amitotische Kernteilung in den Ovarien der Hemip-
teren. Zeitschr. fiir wiss. Zool., Bd. 54, no. 2.

Witson, E. B. 1910 Studies on chromosomes VI. A new type of chromosome
combination in Metapodius. Jour. Exp. Zool., vol. 9, no. 1.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 2

|
sty
“e
h
.
' ae
, ny
i ot |
ee ce wu
. , ‘
\y Py is
ye ‘
4 wee > ms
‘ yr} ab ft ‘ ;
R j
. y ¥ Pea) ¢
5 hg Al
‘ vies ty
)' ayy
oe 4h My 5 matt
; ad , .
Pa RAY | |
2 1. i.
q B\) ee ee om if i i) t Faken i |
pe ‘ y ‘ » oa
mee . lant. Tb bd ee a | Le fi in: ; mr ms uli, 4h) ewe j =
ve in Oe UTA 7.

lea
‘ , ida ee hos ihiryny b} “OER. hela oe

+ | . a : i oe
wut m7

i : 2 7 a “4 a ¥ %
; 4 4 . ‘ aii’ cal ie 7 ve Ven ts j ‘igh } ]
aan hee

ha py) POR a ar ‘ait

| Hatt ur 7 % aig
je ae ee -
; slight
un old
ATLL eee ce ar) iy
Ly ‘a
rs ‘ rip ll Ae ie

uy)
#

i) 7
f

Pa

hve vin beyial eo a hd al

Fla shane ce) 1 eee rt | . whet a
; ue _

ca) coe

hy it egies

reo

OBSERVATIONS ON THE LIFE-HISTORY OF TWO
RARE CILIATES, SPATHIDIUM SPATHULA
AND ACTINOBOLUS RADIANS

JULIA E. MOODY

From the Zoological Laboratory, Columbia University

SIXTY-SIX FIGURES

fees ps thidium spa bh en oe eee ee reise ee cet Pras < ef ape or Salts git a 349
ep OCUG HOLA = Stee Ee cetera Rte acter aie is Ya cchsichs Gatatsie 349
DaMaterial-andsmethodse-s sae een ee ee ele ais oe eA alae mens, ew ore 352
SeeMior phology: arid physiology tee eos Stee deel an) ee tele es ODE
Ar oMEWTOUUCTION) 5 /:-.sce eee Ieee A eT Sas Pcp oak ol gerne ee er OOO
Pye Ham CG CINCT by 3-235, eae Hare ee cp Pare teats cig ee 360
GwOpservations on theslute-histo nee emaar ee Cnr aereeeeee cee eee acre 361
(ERIE cYedsy aleve OO) ORAM Ws hiedrrtin ni aunt's: ys ater iclo™ clcioed eb orc ic eve se ai ici: 363
SPS ATTATIAELIS & =: cs nyc 8 Se Paha ol ee tcl aaa eve ete seek gees, cere eid 364

leeACHIMObOIIS-TAGIATIS: 2 sep oto eer cce ios Meee ke neta ee oy eae 365
NPIELISEOTY, cep S.00 serge eet Ee ie ote ete hect gee recess cramer reas 365
DaNiatertalrand methodSmeaor st. sone cme was itr aie tere) reckeveres 368
Ser NMonpholosy arid up hySlOlogys..2eS ots ae ech. = hia Sa -n actaehAe ice 369
GB TRG] TUCO TRH CODE eer enS Leth 6 Cen Ore Sa eam Dim omicnscla oretr 7) ye oen ow 374
EMMIS YUDTRIUIN EDT Vig oc. ons SR RUE. sone neck SERRE Oe Mais oi Sie, Soautge One oer aren 375

TiltliexGenenralaconsiderationses sete Vir. octets hs cn ee he ors tes Es een eens 377
A. The life-cycle, senescence and rejuvenescence ..................... 377
Be Nucleus*plasmia—nel a Glenn ty a. <n sia oe eect aes ent oa eee te esa a ata 381
Gs HOO: bats yates iro sce Sas eR PR te ae 2's ES Ee Ste 393
MBE AU e CLEC Ac Sec atkh eet ea tate 2c nec eR, cor ith Meme Mrs Lites e 396

I. SPATHIDIUM SPATHULA
1. Introduction

In his ‘‘Histoire Naturelle des Infusoires,’ Dujardin (41)
first described Spathidium as an organism cylindrical in form,
very transparent and marked with 20 to 27 delicate, longitudi-
nal striations; although he speaks of the oblique truncation of
the anterior end he evidently observed neither mouth nor cilia
in this region. Miller in 1786 had described a form, Leucophrys
spathula, which with Enchelys gigas of Ehrenberg, Stein and

349

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3
SEPTEMBER, 1912

390 JULIA ELEANOR MOODY

Entz, was probably synonymous with the form Spathidium dis-
covered by Dujardin.

In the main, Ehrenberg’s description tallies with that of Du-
jardin, differing from it only in the account of the anterior region
where he found the truncated edge bordered by distinct cilia.
Somewhat later, Perty, making a further study of this form,
discovered that these cilia surrounded a distinct mouth.

In “Sur la multiplication des Infusoires Ciliés,’”’ Maupas gave
a fuller description of this interesting organism, which he called
Spathidium spathula, placing it in the family Enchelinidae of
the holotrichous ciliates, although he considered it intermediate
between the Enchelinidae and the Trachelinidae, inasmuch as
it posesses a form typical of the first family but shows in the
region of the mouth an approach to the structures found in the
second family. .

In the more primitive forms represented by Holophrys, the
mouth, a simple passage in direct communication with the
endoplasm, is terminal, but many gradations exist between this
condition and that found in Dileptus where it is situated at
the posterior end of a long narrow lobe. Intermediate condi-
tions are to be found in Spathidium, Enchelys and Nassula;
the first two possessing a slit-like opening, subterminal in position
at the anterior end; while in Nassula the opening is about one-
third the length of the body from the anterior end. According
to Biitschli this change in position of the mouth has come about
by the gradual shifting of this organ toward a ventral side until
it has come to occupy a more or less central position.

In the flask-shaped body of Spathidium Maupas distinguishes
a rounded posterior portion, tapering anteriorly to form an elon-
gated neck, bearing at the truncated tip a slit-like mouth armed
with trichocysts. In common with the holotrichous ciliates,
the entire surface of the body is covered with cilia of equal length
except in the mouth region where they are somewhat longer.
In the long axis of the body, Maupas found a more or less sinuous
band-shaped nucleus accompanied by numerous micronuclei.

This infusorian was found in February, 1911, in a Paramoe-
cium culture which had been brought to the laboratory of the

LIFE HISTORY OF TWO RARE CILIATES aol

City College from Van Cortlandt Park. Through the kindness
of Dr.Goldfarb, some of the culture was sent to Professor Calk-
ins who identified the form as Spathidium spathula. Because
of its rarity, the peculiar and interesting nuclear changes shown,
Professor Calkins suggested that I make a study of its mor-
phological characteristics, its habits, reproductive phenomena,
response to stimuli and the processes occurring in its life-his-
tory. I wish to express my gratitude to Professor Calkins for
his interest in my work and for his helpful suggestions and _ criti-
cisms. .

Before the material came into my hands, Professor Calkins
had kept it under observation for a week or more, experimenting
with various food-media; he found that Spathidium was ex-
' tremely sensitive to old bacteria infusions. Although Maupas
states that Spathidium captures and eats all kinds of small
ciliates, as for example, Cyclidium and Glaucoma, during the
entire time I had it under observation, I never saw it’ paralyze
or eat any ciliate except Colpidium, in fact it seems impossible
to cultivate it without the aid of this particular form. From
time to time during the past five months, I have found in the
original culture containing no Colpidia, a few abnormally small
Spathidia. That the organisms were not in a healthy condition
was shown by their greatly reduced size and the perfectly trans-
parent condition of the protoplasm. I am at a loss to account
for the free-swimming individuals in the old culture; it may be,
however, that some slight change in the environment proved
sufficient to cause the emergence of the encysted forms.

In the rich cultures under observation during the first six
months of 1911, if for any reason Colpidium became reduced
in numbers, Spathidium encysted. It was quite possible to
recover them within four or five hours by adding to the medium
containing the cysts, fresh hay-infusion either with or without
an abundance of Colpidium. A jar containing the original cul-
ture was left tightly covered from January 1 to January 12,
1912, upon which date it was examined. Cysts were abundant
but no free-swimming forms. A small quantity of fresh hay-
infusion was added and the jar was left uncovered. On January

352 JULIA ELEANOR MOODY

13 large numbers of very small Spathidia were found but no
Colpidia. Change in environment, therefore, would seem to
account for the appearance of the small transparent individuals
which have been observed in the original culture from October,
1911, to the present date.

2. Material and methods

While details of Spathidium spathula can be seen only with
a compound microscope, I found the binocular much more con-
venient to use. From a watch glass containing daughter cells
of one individual, three were selected on February 24, 1911,
and each transferred by means of a capillary pipette to a sepa-
rate glass dish. These dishes are 4 cm. square, 8 mm. deep
and possess a central, circular, shallow cavity, the capacity of
which is about 13 ce., or 80 drops. I have found these small
dishes invaluable for this work because owing to the clearness
of the glass and the gradual slope of the depression it is impossi-
ble for the organism to get into any part of the dish which cannot
be brought into focus. They are also more convenient to handle
than depression slides. These glass dishes were kept in a moist
chamber, a large glass stender dish 10 inches in diameter and
3 inches deep.

After several trials with tap-water, pond-water and hay-
infusion, the latter was adopted as the best medium, prepared
according to the method of Calkins (’02) and used when twenty-
four hours old. In general, the hay-infusion was used undiluted,
but toward the end of the series a few drops of pond-water were
frequently added with satisfactory results.

Each morning the number of divisions during the preceding
twenty-four hours was recorded and one individual from each
line isolated according to the following method: into one of the
small glass dishes containing six drops of undiluted hay-infusion
and one drop of Colpidium culture, a single Spathidium was
transferred by means of a capillary pipette, care being taken to
carry over as little as possible of the old medium. This pre-
caution is necessary since the small quantity of fluid in the
depression dish soon becomes turbid with the accumulation of

LIFE HISTORY OF TWO RARE CILIATES 3538

bacteria and the waste products of Colpidium, both of which
are detrimental to the well-being of Spathidium. After isolating
one individual from each line, the rest were kept as reserve
stock. Such Spathidium stock was kept in Syracuse watch-
glasses and although many attempts were made to cultivate
it in larger dishes, none met with success, the free-swimming
forms encysting in large numbers within twenty-four hours.

All dishes used were carefully washed in hot water and dried.
The pipettes were also carefully cleaned, a special one being
reserved for the purpose of isolation.

Both living and prepared material were studied. Several
fixing agents were tried. A saturated aqueous solution of
corrosive sublimate with 5 per cent acetic acid gave the most
satisfactory results; the other fixatives led to shrinkage of the
cells. In making the total preparations, a quantity of Spathidia
were transferred with a capillary pipette to a watch-glass con-
taining a small quantity of the fixing fluid. They were then
carried over to 70 per cent alcohol and allowed to settle to the
bottom of the watch-glass. From this they were transferred
to a slide smeared with egg-albumen, the alcohol coagulating
the albumen and firmly attaching the ciliates. The slide was
next placed for half-an-hour in a strong aqueous solution of picro-
carmine and, after destaining with acid alcohol, was carried up
through the alcohols to xylol and mounted in balsam. Heiden-
hain’s iron haematoxylin and polychromatic methylen blue were
also used for staining but the results were unsatisfactory.

The material destined for sectioning was fixed in sublimate-
acetic and carried up to absolute alcohol from which it was
transferred to an absolute alcoholic solution of magenta. After
fifteen minutes in the stain, it was washed in absolute alcohol.
Xylol was added drop by drop to the last aleohol until a quantity
equal to the amount of aleohol had been used. The material
was then transferred to pure xylol. Considerable difficulty
was experienced in imbedding when the paraffin-oven was used,
due to overheating the material. In transferring from the first

to the second paraffin, which must necessarily be done under
the microscope, it is difficult to keep the paraffin at the melting

354 JULIA ELEANOR MOODY

point until the transfer is made. A method was finally suggested
which entirely removed this difficulty. A small quantity of
soft paraffin in a watch-glass was melted over the water bath.
By means of a capillary pipette, the specimens, either individu-
ally or en masse, were transferred from the xylol to the melted
paraffin. When a film had formed over the surface of the paraffin,
the dish was dropped into cold water. The imbedded Spathidia,
stained bright red, were readily located with the aid of the bino-
cular. Having melted with a warm needle the paraffin §sur-
rounding the specimen, it was transferred to a glass slide which
had been previously smeared with dilute glycerine. From a
pipette of large bore, melted hard paraffin, sufficient to form
a good sized drop or sphere, was squeezed upon it. The hot
paraffin melts at once the congealed soft paraffin adhering to
the specimen and thus embeds it in a homogeneous matrix.
The sections were cut 33 microns thick and stained with iron
haemotoxylin.

3. Morphology and physiology

Spathidium spathula, as Maupas describes it, is a slender
flask-shaped ciliate, the rounded distended posterior end of
which is drawn out anteriorly to form a long greatly compressed
neck, obliquely truncated at the tip (fig. 1). The records of
Dujardin and Maupas show a great variation in the size of the
organism, the former quoting 180 to 240u as the ordinary length,
while Maupas gives 160u as the outside limit. Of fifty indi-
viduals measured, I found the average length and width to be
110.54 and 35.2u respectively; the smallest measuring 73,5 in
length with a diameter of 31; the largest 157.5 in length and
57.54 in diameter. :

Maupas described the body as capable of great extension,
having often observed it stretched to five or six times its normal
length. In swimming, Spathidium sometimes becomes entangled
in the zooglea at the bottom of the culture dish and under such
conditions it often becomes greatly extended in its efforts to
escape, but I have never seen it more than double its length
and that only on one occasion.

LIFE HISTORY OF TWO RARE CILIATES 30D

Extending from end to end of'the body are many delicate
longitudinal striations, marking the lines of insertion of the cilia.
These organs are distributed uniformly over the body and are
of equal length except in the oral region where those bordering
the mouth are somewhat longer. The body is covered by a
thin but firm cuticle which gives it its permanency of form allow-
ing at the same time the greatest flexibility. In section the
cuticle is seen as a pale area, bounded by a definite line surround-
ing the more opaque central portion(fig. 2). The protoplasm
immediately underneath the cuticle, constituting the cortical
‘plasm, appears somewhat denser than that toward the center,
although there is not a distinct line of demarcation between
the two areas. The entire protoplasm is finely granular.

As the animal swims rapidly through the water, rotating on
its long axis, the anterior end of the body, which is extremely
flexible, is in constant motion, bending upward, downward and
from side to side as though feeling its way. Owing to this con-
stant change of position, this region continually presents new
aspects, as is shown in sketches of many of the total preparations
as well as in figures 13 and 17. The mouth, a narrow slit, sub-
terminal in position, is bounded by the thickened edges of the
truncated tip. Maupas thought the mouth-opening was limited
to the extreme posterior part of the oblique edge, but I have
found it to extend the entire length of the truncated end. The
nature and extent of the mouth can best be determined during
the act of feeding. Spathidium is a predatory form, swimming
actively about as though in search of prey. The moment the
anterior end of the body comes in contact with a Colpidium,
all motion in the latter ceases. Spathidium twists and bends
its flexible body until the smaller diameter of Colpidium is paral-
lel to the truncated edge; the mouth opens, the thickened lips
surround the small ciliate which passes slowly down into the
body of the captor, the entire process occupying not more than
five minutes.

Small spindle-shaped bodies, which in fixed specimens appear
as closely crowded fine lines, are imbedded in the thickened
rim of the mouth directly underneath and at right angles to

356 JULIA ELEANOR MOODY

the cuticle (figs. 8 and 12). Butschli mentions sixteen or more
club-shaped, contractile rods, which Maupas interprets as tri-
chocysts. Without doubt the trichocyst material is located in
this region, since Colpidium, paralyzed when touched by the
anterior end, suffers no evil consequences by contact with any
other part of the body. Three artificial methods were used to
explode the trichoeysts: exposure to osmic acid vapor, treatment
with a 2 per cent solution of acetic acid and a solution of methyl
green. After treatment with any one of these reagents, in fifty
individuals examined, the cilia were found fully extended, the
trichocysts being readily distinguished from the long oral cilia -
inasmuch as they were stiff and straight in appearance, whereas
the cilia showed a wavy outline. Mitrophanow in his contri-
bution, ‘‘Etude sur la structure, le développement et l’explosion
des trichocysts des Paramoécies”’ describes the presence of small
bodies in the region of the nucleus which take the nuclear stain.
Among these deeply staining masses he found a variety of forms,
ranging from spherical granules to rod-shaped bodies, similar
in structure to the typical trichocyst, and although he did not
actually observe the extrusion of these particles from the nucleus,
he thought it probable that the trichocyst material is of nuclear
origin, migrating from the interior of the cell toward the per-
iphery. Here by proper stimulation the cortex contracts, the
semi-fluid trichocyst material is forced out, solidifying as it
comes in contact with the water. I have seen in Spathidium
no deeply staining particles, originating in the region of the
nucleus, which could be interpreted as developing trichocysts;
neither have I found the greatly elongated trichocysts of com-
plicated form described by Meier, Schuberg and Schewiakoff
for Frontonia, Paramoecium and some other ciliates; owing,
however, to the paralyzing effect on Colpidium following con-
tact with the anterior end of the body, I think it safe to interpret,
as trichocysts, the short opaque rods so plainly visible in the
thickened rim of the mouth.

The mouth opens into a space, the pharynx, reinforced by-par-
allel thickenings of the cortex as shown in figure 18. Imbedded
in the protoplasm are numerous food vacuoles, the contents

LIFE HISTORY OF TWO RARE CILIATES 500

in various stages of digestion. In many of the total prepar-
ations the body of Colpidium is seen practically intact (figs.
23 and 25) while in others the cytoplasmic envelope has disap-
peared, leaving the macronucleus, with the micronucleus lying
in a depression on its surface, surrounded by the endoplasm
of Spathidium (fig. 30). Other deeply staining bodies, varying
greatly in size and number, are scattered throughout the pro-
toplasm. Doubtless they are the remains of food particles
which have been subjected during a longer period of time to
the action of the digestive ferments. According to Calkins,
the more conspicuous granules found in the protozoan cell are
formed by the breaking down of the food particles, some of
which are directly assimilated while others remain as reserve
nutriment. He notes the difference in appearance of the pro-
toplasm of a well-fed and a starved Paramoecium; that of the
former showing a typical granular structure, that of the latter,
the entire absence of granules. This contrast is very marked
in Spathidium, the protoplasm of a well-nourished individual
being densely granular, that of the starved forms appearing as
a clear and almost structureless substance.

At the posterior end of the body is a large single terminal
vacuole which, at room temperature, pulsates on an average
of once a minute. Miiller describes two vacuoles in Enchelys
spathula, one near the middle of the body, the other at the pos-
terior end. In the species under observation, two vacuoles
are sometimes seen for a time at the posterior end, but sooner
or later the two coalesce in one large terminal vacuole (fig. 27).
At the systole the contents of the vacuole are expelled through
a minute opening at the extreme tip of the body.

Although total preparations and sections of vegetative cells
as well as early and later division stages have’ been carefully
studied, the observations have yielded no positive evidence of
a differentiation of the nuclear material into two structures, a
micronucleus and a macronucleus. In a few instances, vesicular
bodies, varying from spherical to retort-shape containing deeply
staining granules, have been observed (figs. 10 and 14). The
contained granules are elongated and constricted at the center

358 JULIA ELEANOR MOODY

as though in process of division (figs. 11 and 15); but since
these bodies occur so rarely, they cannot be reasonably considered
permanent cell structures. The macronucleus, extending length-
wise of the cell, is a greatly elongated, rod-shaped body, circular
in section, measuring from 4p to 8» in diameter. It is coiled,
twisted and folded upon itself, often attaining a length two or
three times that of the body (figs. 1, 4 and 5). Although the
typical nucleus is long and intricately coiled, a few exceptions
were found, two of which are shown in figures 9 and 16. The
former might be interpreted as indicative of beginning division;
this however is not the case, since in the many division stages
studied, the nucleus could always be traced as a continuous
band from the anterior to the posterior cell. ‘The macronucleus
is bounded. by a very delicate membrane which may be readily
seen in total preparations which have been compressed by the
cover-glass. Here the macronucleus, free from the surrounding
protoplasm, retains its definite outline. The chromatin of the
macronucleus consists of a mass of minute granules, which take
an intense color in staining with iron-haemotoxylin. Surround-
ing the chromatin masses is a faintly staining substance which
in section appears somewhat more opaque than the protoplasm
of the cell. The granules in the late stages of division are ex-
ceedingly fine (figs. 21 and 22). In earlier stages, where an
elongation of the cell body has taken place, but as yet no con-
striction can be seen, the chromatin appears as deeply stained
fine threads which would seem to indicate a division of the larger
chromatin granules (figs. 19 and 20). Although it is generally
thought that the possession of the two types of nuclei is char-
acteristic of the ciliates, the present observations give no evidence
of this differentiation 1 in Spathidium.

4. Reproduction

Spathidium multiplies by simple fission, the rate of division»
according to Maupas, being one division in twenty-four hours:
The'cultivation of Spathidium during two hundred and eighteen
generations, showed considerable variation in the division rate.
During one ten-day period, the average daily rate of division

LIFE HISTORY OF TWO RARE CILIATES 309

was 3.1, the lowest rate of division for the same series for a simi-
lar period of time being 0.2 per day. Averaging the high and
low division rates of the three lines from February 24 to July 7,
1911, a division rate of 1.5 in twenty-four hours, was found.

Division of the cell and macronucleus is preceded by their
gradual elongation, this continuing until the cell has nearly
doubled its length, in fact in some cases the dividing cell exceeds
twice the length of the vegetative cell.

The variation in form of the macronucleus in division is inter-
esting. Sometimes it is simply folded upon itself, appearing
like a single strand except at the point where the two ends sepa-
rate (figs. 23, 24 and 25); again two distinct. pieces may be
distinguished, each one folded upon itself and the parts inter-
twined (fig. 26). Sometimes an exaggerated elongation of the
nucleus occurs as shown in figure 27, in which it is extremely
difficult to follow the coils and intertwinings.

After the elongation of the cell is accomplished a constriction
occurs half-way between the anterior and posterior ends, a new
vacuole appearing anterior to this infolding. A little to one
side of the point of attachment of the dividing cell, a new mouth
is formed in the posterior cell. The position of the new mouth
is best seen in the living organism immediately after division.
The constriction increases until the cell and the nucleus are
divided into approximately equal parts as shown in figures 29,
30 and 31, although an occasional exception is found where
the larger part of the nucleus lies in the anterior cell (fig. 28).
~ During the process of division Spathidium swims slowly about,
but during the later stages its activity increases. It not only
swims more rapidly but it twists the anterior end so energeti-
cally that the connection between the two cells is reduced to a
delicate strand, this condition continuing for an hour or more.
When the slender thread of tissue is severed, the anterior cell
swims rapidly away, leaving the posterior half rotating slowly.
The posterior individual is at first ovoid in form, the anterior
end marked by the new mouth placed somewhat obliquely,
the posterior end by the original vacuole. Soon the rotary
motion ceases; the new cell swims slowly about, increasing gradu-

360 JULIA ELEANOR MOODY

ally in length, while the anterior region is drawn out to form
the narrow, compressed neck truncated obliquely at the tip.

Although cultures of Spathidium were kept under observation
from February 24 to July 7 through two hundred and eighteen
generations, no conjugation was.observed. Attempts were made
to bring about this phenomenon by transferring numbers of
Spathidium to small dishes according to the method of Calkins
(04) but without success.

5. Encystment

Encystment takes place either for the purpose of protection
against conditions adverse to the well-being of the organism
or for the purpose of reproduction. During a period of one
month attempts were made daily to cultivate Spathidium in
dishes of greater capacity than the Syracuse watch-glasses.
Staining dishes 3 em. in diameter and 2 cm. in depth proved
fairly satisfactory for a short time, very few encystments occur-
ing during the first two or three days. But although given a
sufficient amount of hay-infusion and fed with a rich culture
of Colpidium at twenty-four-hour intervals, every individual
had encysted at the end of the fourth day. All attempts to
cultivate Spathidium in dishes larger than those described re-
sulted in encystment within twenty-four hours.

To show the process of encystment the following experiments
were tried. On March 24, 1911, stock in good condition was
transferred to four Syracuse watch-glasses, some of it being
kept as control. To all four dishes was added an equal amount
of hay-infusion prepared twenty-four hours previously. With
a pipette a small quantity of the Colpidium culture was trans-
ferred to watch-glass A. To watch-glass B a quantity of Col*
pidium equal in amount to the culture of Spathidium transferred
was added. In both cases cysts were formed within twenty-four
hours. To watch-glass C and Da medium quantity of Colpidium
was added. Culture C was left over night close to the window
where the temperature was considerably below room temperature.
In the morning every individual was encysted. Culture D
was placed on the edge of the table near the radiator at 8 o’clock

LIFE HISTORY OF TWO RARE CILIATES 361

in the morning. At noon one-half of the number had encysted.
The control, supplied with a few Colpidium remained in good
condition, no cysts being found.

The cause of the encystment of the individuals in A may
be traced to a too small quantity of Colpidium plus an excess
of hay-infusion, previous experience having shown that Spathi-
dium will not flourish in a large amount of fluid. Encystment
in B was due to an excess of the products of Colpidium metabo-
lism the same result having been observed in smaller cultures
of Spathidium under similar conditions. Unfavorable tempera-
ture conditions probably caused the encystment in C and D.

By the addition of fresh medium it is often possible to recover
Spathidium from the cysts. At 8 o’clock one morning, hay-
infusion and Colpidium stock were added to the medium con-
taining encysted forms. The cysts were semi-transparent, the
vacuole being distinctly visible. At 9.30 the contents of the
cysts began to rotate. After half-an-hour, the form of the
body was distinguishable. As the rotation continued, the wall
of the cyst became more and more transparent and at 12.45,
a portion of the protoplasmic contents was extruded beyond
the cyst wall. Five minutes elapsed from the time the first
portion of the cell-body appeared to the extrusion of the entire —
protoplasmic mass.

Upon first emerging from the cyst, Spathidium is weak and
shrunken in appearance. The individual under observation
remained quiet in the neighborhood of the cyst for several min-
utes; then swimming slowly about it paralyzed and swallowed
a Colpidium within ten minutes of its emergence from the cyst.
Figures 35, 36, 37 and 38 give a few of the stages described, the
individual being under constant observation from 8 A.M. until
1 P.M.

6. Observations on the life-history

From a culture of Spathidium started on February 24, 1911,
three lines, daughter-cells of one individual, were isolated. The
average number of divisions daily of the three lines during the
first ten-day period was 1.6. This was followed by a slight
increase in division rate (1.7) during the second ten-day period.

362 JULIA ELEANOR MOODY

In the interval between March 13 and March 28, there was a
sudden and marked increase, the number of divisions averaging
3.1 per day. Diagram 1 was made according to the method
of Calkins. The curve represents the general vitality of the
three lines through 218 generations, from February 24 to July
7, 1911. Woodruff defines a rhythm as “‘a minor periodic rise
and fall in fission-rate due to some unknown factor in cell metab-
olism from which recovery is autonomous’ and in summing
up his paper on ‘‘Rhythms in the reproductive activity of In-
fusoria’”’ he concludes ‘‘that it is not possible by constant envir-
onmental conditions to eliminate the rhythms and resolve the
graph of the multiplication into an approximately straight line.”
Spathidium, cultivated under constant environmental conditions
except for slight variations in the temperature, showed distinct
rhythmic fluctuations as indicated by the curve in diagram 1,
which result is in close accord with those obtained by Gregory
(09) in a study of the life-history of Tillina magna. The de-
crease in fission rate, followed by the death of the series; may
have been due to a difference in the salt content of the water.
_ From February 24 to June 6 Croton tap-water was used, but
from the latter date to July 7 spring water was substituted.
After March 23 there occurred a sudden decrease in the vitality
followed by high division rate from April 2 to April 30. From
this time on there was a gradual trend downward as indicated
in the diagram.

On June 6, the entire stock was in poor condition as indicated
by the low division rate. On this date the individuals were
transferred to fresh hay-infusion to which Colpidium in abun-
dance had been added. Spathidium was sluggish and although
it continued to feed, it showed no improvement. Stimulation
with beef-extract was tried. At the end of an hour, one out of
six had died, and those which survived showed little improve-
ment. At the end of the second hour these individuals showed
a return to their normal condition and were transferred from
the beef-extract to tap-water then to hay-infusion containing
a few Colpidia. Although the individuals were normal in ap-
pearance, the division rate remained low.

LIFE HISTORY OF TWO RARE CILIATES 363

On June 17, the cultures were carried in small phials from
New York to Maine. There was no opportunity to examine
them again until June 19 when it was found that all except
four individuals had encysted and all efforts to force them to
emerge from the cysts were fruitless. The four surviving indi-
viduals were normal in appearance and continued to divide.
On July 2, the culture was in a flourishing condition when without
any apparent cause, abnormal forms appeared, many died,
while among the surviving individuals the division rate was
practically nil. Various stimulants which previously had pro-
duced beneficial results, as for example, beef-extract, spring
water, 75, 130, revo Solutions of KCl had no effect. On July
6, six individuals only remained, all of which were abnormal
(figs 6, 7, 8). On July 7, three of these encysted, the remaining
three disintegrating while under observation. During the re-
mainder of the month many unsuccessful attempts were made
to recover the encysted individuals and the race died out in
the 218th generation.

7. Regeneration

A few experiments were made to test the regenerating power
of Spathidium, but as these were limited in number and scope,
no general conclusions can be drawn from them.

1. At 11.15 in the morning, half-an-hour after division, a
Spathidium was transferred to a depression slide, placed upon
the stage of the microscope. With a sharp scalpel an incision
was made at right angles to the long axis of the body as near
the middle line as possible (fig. 32). The anterior half formed
a new vacuole, swam slowly about and at 12.15 encysted. The
posterior fragment rounded up, gradually elongated and assumed
normal shape. It was transferred at 12.15 p.m. and although
it swam actively about, coming in contact constantly with its
prey, no paralysis of Colpidium resulted. It would appear
therefore that either the trichocysts were functionless or had
not yet been formed. If Mitrophanow’s theory of the nuclear
origin of these bodies be correct it seems resonable to believe
that sufficient time had not elapsed since the cutting for the

364 JULIA ELEANOR MOODY

development of trichocyst material. On the following day this
individual had divided once and continued to do so normally
during the week it was under observation.

2. The same experiment was tried ten minutes after division,
resulting in the disintegration of both fragments. The failure
to regenerate agrees with the results of Calkins in his experi-
ments on Uronychia where he found the regenerating power
low immediately after division.

3. Another individual was cut six hours after division, the
incision being made at right angles to the long axis of the body
but anterior to the middle (fig. 33). The anterior fragment
rounded up, swam actively about and at the end of twenty-four
hours had entirely regenerated, although it was unusually small.
By the following day it had reached normal size and divided twice.
The posterior fragment also regenerated and divided normally.

4. A fourth individual was cut across the long axis of the
body posterior to the middle, six hours after division (fig. 34).
The anterior fragment regenerated in twelve hours. The pos-
terior fragment, small and irregular in shape, gradually rounded
up, elongated and assumed normal form. After forty-eight
hours, it divided and continued to do so normally during the
week it was under observation.

&. Summary

(a) Spathidium spathula varies in length from 75.5 to 157.5y;
in diameter from 21, to 57.5u, the average length and diameter
of fifty individuals measured being 110.54 and 35.2u respectively.

(b) The mouth is an elongated slit, subterminal in position,
extending the entire length of the truncated tip.

(c) Imbedded in the thickened rim of the mouth, are minute
opaque rods, the trichocysts, which when artificially exploded,
are readily distinguishable from the long wavy oral cilia. The
trichocysts are used to paralyze Colpidium upon which the
organism feeds exclusively.

(d) At room temperature, the single, terminal vacuole pulsates
at an average of once a minute.

«

LIFE HISTORY OF TWO RARE CILIATES 365

(e) No evidence has been found in Spathidium of a differ-
entiation of nuclear material into a macronucleus and a micro-
nucleus. The macronucleus, a greatly elongated cylindrical
organ, sinuous in outline and intricately coiled, possesses a
delicate membrane within which le closely crowded chromatin
granules, surrounded by a faintly staining substance. Division
of the chromatin masses has been observed in sections of early
division stages.

(f) Division of the organism is preceded by an elongation
of the body and nucleus; a constriction appears half-way between
the anterior and posterior ends; a new vacuole forms anterior
to the constriction, a new mouth, posterior to it and a little to
one side of the median line of the dividing cell.

(g) Eneystment takes place under unfavorable environmental
food and temperature conditions. The free swimming forms
can be recovered by bringing about normal conditions of environ-
ment in respect to food, temperature and quantity of liquid.

(h) Conjugation was not observed, although numerous at-
tempts were made to bring it about.

(i) The descendants of a single individual of Spathidium
were followed through 218 generations extending from February
24, to July 7, 1911, variations in division rate being graphically
represented by a plotted curve. This curve, showing the divi-
sion rate of the protoplasm, illustrates the normal rhythms
described by Woodruff.

II. ACTINOBOLUS RADIANS
1. History

_ Stein observed this rare and interesting organism among the
filaments of fresh-water algae in standing water. His descrip-
tion, limited to a footnote in the second volume of ‘‘ Der Organ-
ismus der Infusionsthiere,”’ is as follows:

Diese neue Gattung beruht auf einem merkwiirdigen Thiere, welche
ich seit mehreren Jahren bei Niemegk ziemlich haufig in stehenden
Gewiissern zwischen der vielwurzeligen Wasserlinse beobachtete und

welches ich Actinobolus radians nennen will. Der Ko6rper ist fast
kugelig oder umgekehrt eif6rmig, am vorderen Pole mit einem kurzen

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

366 JULIA ELEANOR MOODY

zitzenformigen Fortsatz versehen, In dem die enge Mundoffnung
liegt, und ringsum mit gleichformigen Wimpern besetzt. Zwischen
den Wimpern stehen zahlreiche fadenférmigen Tentakeln zerstreut,
die sich, wie die Tentakeln der Acinetinen betrachtlich verlangern
und auch spurlos in den Korper zuriickziehen kénnen. Der After
und ein grosser kontraktiler Behalter liegen am hinteren Ko6rperpole.
Der ziemlich lange strangférmige Nucleus ist unregelmisig zusammen-
gekrimmt. Die Gegenwart von Mund und After schliesst unser
Thier entschieden von den Acinetinen aus, denen es auf den ersten An-
blick sehr &hnlich scheint.

The next recorded observations are those of Entz published
in April, 1888. Actinobolus, found in abundance by him in
June, had two weeks later entirely disappeared, not only from
his cultures, but also from the pond from which his material
originally came. To the facts recorded by Stein, Entz adds
the following details: The mouth, situated at the anterior
end of the body at the tip of a papilla-lke structure, leads into
a funnel-shaped gullet finely striated. These striations he in-
terprets as folds of the cortex, finding here no differentiation into
distinct solid rods, homologous to the basket-like structure
found in Chilodon. The tentacles, extending outward from the
entire surface of the body, he found to be of uniform thickness
throughout their length, ending generally bluntly, sometimes
terminating in a sharp point, but never knobbed. These organs
consisted of a homogeneous substance, capable of great exten-
sion and contraction, but different from the tentacles of Suctoria
inasmuch as in their contracted condition the close spiral ap-
pearance observed in the latter, was entirely lacking.

Stein ascribed no special function to the tentacles, but Entz,
although he states that he has never seen Actinobolus swallow
an infusorian, often observed the tentacles closely attached to
the filaments of Cladophora and other algae, the walls of which
upon careful examination, appeared to be ruptured at the point
of contact with the tentacles. Entz therefore concluded that
the tentacles of Actinobolus may assist the animal in securing
food, by the secretion of a substance from the tip of the organ,
thus dissolving the cell-wall of the plant and exposing the con-
tents to the action of the endoplasmic core of the tentacles.

LIFE HISTORY OF TWO RARE CILIATES 367

As a result of these observations he considered Actinobolus
an holophytice form.

Entz noted a difference between the nuclei found in old and
young cells; the nucleus of the former he described as an elon-
gated, cord-like organ, frequently broken up into spherical
segments, presenting the appearance of a mass of free nuclei;
the nucleus of the latter, a kidney-shaped, horse-shoe form or
spherical mass of chromatin. He found, scattered among the
filaments of the algae, many thin-walled cysts, within which
two, infrequently four, distinct masses of protoplasm were seen
rotating; He interpreted the cysts as division or reproductive
cysts; the protoplasmic masses as daughter cells. At the time
of encystment the following changes occurred: The organism
withdrew its tentacles, lost its cilia, the protoplasm became
very dense, the nucleus shortened and division of the cell fol-
lowed. After fission the new individuals emerged from the
eyst, taking on within a short period of time, the form of the
mature cell, developing tentacles and cilia, the protoplasm again
showing the characteristic foamy structure and the nucleus
elongating to form the cord-like organ typical of the adult cell.

Von Erlanger, working at Heidelberg under Biitschli, was the
next contributor to our knowledge of this ciliate, making a special
study of the structure and function of the tentacles. He found
these organs regularly arranged in the ciliary grooves extending
from anterior to posterior end of the body and could trace them
inside of the body in their contracted state, although he states
in a foot note that Biitschli was unable to confirm his observation
in this particular. In a fully extended tentacle, von Erlanger
distinguishes three regions; the proximal part, thick and spher-
ical in form, a long slender portion (both of these parts being
perfectly transparent); and a slender opaque rod terminated
by a small knob much more minute than the similar structure
found in the suctorian tentacle. Fully retracted tentacles
formed a peripheral row of minute rods appearing like tricho-
cysts found: in other ciliates. Tentacles treated with osmic
acid showed an extremely fine dart projecting beyond the knob.
Like Entz, von Erlanger never saw the tentacles used to seize

368 JULIA ELEANOR MOODY

prey, but because of the presence of trichocysts at their tips,
he looked upon them as organs of protection.

In 1901 in his paper dealing with some interesting protozoa
found in Van Cortlandt Park, Calkins records his observations
on the feeding habits of Actinobolus and the functions of the
tentacles. To quote: (’016, p. 50)

This remarkable organism possesses a coating of cilia and retractile
tentacles which may be elongated to a length equal to three times the
diameter of the body, or withdrawn completely into the body. The
ends of the tentacles are loaded with trichocysts (Entz ’83). When
at rest the mouth is directed downward and the tentacles are stretched
out in all directions forming a minute forest of plasmic processes, among
which smaller ciliates such as Urocentrum, Gastrostyla etc., or flagel-
lates of all kinds may become entangled without injury to themselves
and without disturbing the Actinobolus or drawing out the fatal darts.
When, however, an Halteria grandinella, with its quick jerky move-
ments, approaches the spot, the carnivore is not so peaceful. The
trichocysts are discharged with unerring aim and the Halteria whirls
around in a vigorous but vain effort to escape, then becomes quiet,
with cilia outstretched, perfectly paralyzed. The tentacle with the
the prey fast attached is then slowly contracted until the victim is
brought to the body, where by the action of the cilia, it is gradually
worked around to the mouth and swallowed with one gulp. Within
the short time of twenty minutes, I have seen Actinobolus capture and
swallow no less than ten Halterias.

Thirty-five years, therefore, after its discovery by Stein in
1867, Calkins solved the problem of the function of the ten-
tacles of Actinobolus, finding that, contrary to the conclusion
of Entz, it is a holozoic form, existing exclusively on the small
ciliate, Halteria.

2. Material and method

The pond water in which Actinobolus was found was brought
to the laboratory from Van Cortlandt Park. Entz’s obser-
vation, unverified, by Von Erlanger, that the ciliate is always
found associated with suctorian forms, was not found to be true
in this case, other ciliates and flagellates abounding, but no suc-
toria. Entz’s statement may be accounted for by the fact that
Actinobolus bears a superficial resemblance to the suctorian
Sphaerophrya and when at rest might easily be mistaken for
this form.

+ 7

LIFE HISTORY OF TWO RARE CILIATES 369

Maupas found bouillon, prepared by boiling a minute quan-
tity of flour in water, a satisfactory medium for the cultivation
of infusoria. A modification of this method was adopted for
the cultivation of Halteria, but owing to the rapid fermentation
of the flour solution it proved unsatisfactory. After experi-
menting with hay-infusions of various strengths, with tap-water
and with pond-water, the latter was chosen as the best medium
for the cultivation of Halteria stock. The Actinobolus stock
‘was kept in Syracuse watch-glasses containing pond-water to
which had been added a small quantity of Halteria.

Four lines were isolated on October 18, 1911, as subcultures
from Professor Calkins’ cultures. The same method was em-
ployed as for Spathidium culture, Halteria replacing the Col-
pidium as food. Both living and fixed material consisting of
total preparations and sections were studied. Schaudinn’s fluid,
a mixture of 80 parts corrosive sublimate and 20 parts absolute
alcohol, was found to be a most satisfactory fixing agent. The
method used in making preparations and sections was the same
as that described for Spathidium. Many attempts were made
to get satisfactory permanent preparations of the tentacles.
Narcotizing with chloral hydrate, killing in Schaudinn’s mixture
and staining for several hours in either picro-carmine or magenta
gave fairly good results. Useful, though not permanent, prepa-
rations were obtained by killing in osmic acid vapor and mount-
ing in glycerine. Living material, however, proved far more
satisfactory for the study of these organs than any of the fixed
preparations described.

3. Morphology and physiology

Actinobolus radians, a holotrichous ciliate, belonging to the
suborder Gymnostomina, family Enchelinidae, is a small, almost
spherical organism found in fresh water. The average length
and diameter of twenty-five individuals measured are 58.5u
and 46.8 respectively (fig. 39). Regarding the shape of the
body Entz and von Erlanger disagree, Entz describing the anter-
ior as the broader region, Erlanger stating that this portion

370 JULIA ELEANOR MOODY

of the body is tapering. The difference in width between the
anterior and posterior ends is slight, but becomes more marked
during division when, after the cell has elongated and the con-
striction appears, the future posterior cell is seen to be of some-
what smaller diameter than the anterior end (figs. 55 and 56).

The anterior region of the body is drawn out to form a minute
papilla-like structure, at the extreme tip of which les the mouth.
This projection is not always discernible, however, since this
region is extended and retracted at intervals, a slight depression’
often marking the position of the mouth.

In the living cell two clearly differentiated regions are dis-
tinguishable, a dense central region, the endoplasm, surrounded
by more transparent area, the ectoplasm. In common with
ciliates generally, the body is covered with an extremely thin
membrane, the cuticle, the surface of which is marked by 24
or 25 delicate lines extending spirally from the border of the
mouth to the posterior end of the body. These lines, which
in optical section are seen to be narrow ridges, indicate the
places of insertion of the retractile tentacles and the cilia, the
latter arranged in small groups at the base of each tentacle.
Directly underneath the cuticle is the cortical plasm, a semi-
transparent substance in which lie imbedded masses of highly
refractive bodies which are in constant circulation. Large
vacuoles form a characteristic peripheral border, while strands
of endoplasm extend between the vacuoles and grade insensibly
into the cortical plasm at the base of the tentacles (fig. 41).

The retractile tentacles, originating in the cortex and capable
of an extension equal to two or three times the diameter of the
body, are especially interesting. I have already referred to
von Erlanger’s work on the tentacles of Actinobolus in which
he distinguishes three distinct regions; a spherical basal portion
from which extends a rod-like body bearing at its distal end a
sharp dart terminated by a minute knob; the basal sphere and
middle portion being perfectly transparent, the dart quite opaque.
After treating the animal with osmic acid vapor, he noted an
extremely fine thread projecting beyond the knobbed tip of
the dart. Although a very careful study has been made of

LIFE HISTORY OF TWO RARE CILIATES Si

the organs in total preparations and sections, as well as in the
living animal examined under the oil immersion lens, I have
been unable to find the structures described by von Erlanger.

With partially retracted tentacles, Actinobolus, rotating on
its long axis, swims occasionally near the surface of the water,
coming to rest from time to time with the mouth downward,
on the bottom of the culture dish, moored by means of the oral
tentacles. Gradually the retracted tentacles elongate until, seen
with the low power, the animal is apparently surrounded by a
halo of short opaque rods at varying distances from the per-
iphery of the body. By careful focussing, the opaque rods
are seen to be the tips of the tentacles which can be readily
traced to the cortical layer of the body. The tentacles are
extremely slender organs of equal thickness throughout their
length, the tip being slightly rounded, never pointed or knobbed.
The opacity of the distal end is due to a mass of minute dark
granules, the trichocyst material, which can be seen in circu-
lation as the tentacle is extended.

As shown by Calkins, the tentacles are food-getting organs,
Actinobolus subsisting exclusively on Halteria grandinella; often
as many as four or five may be seen attached to these organs,
the little ciliates being carried around one by one to the mouth
and swallowed. It is a matter of conjecture as to the method
by which the capture of this prey is accomplished. Calkins
in his description of the action of the tentacles of Actinobolus
says, (Ola, p..657)::

The trichocyst which von Erlanger described, or its analogue (I
was unable to make it out as described by Erlanger), brings down the
prey, and the tentacle, by shortening, fetches it to the mouth. The
tentacle itself is inserted in Halteria, for it can be easily seen with the
greatest distinctness when the victim becomes quiet, and I believe
that the dart at the end as described is not discharged into the prey

but is driven into the soft body of the victim as a minute spear, with
shaft attached.

Since the tip of the tentacle is blunt and slightly rounded
it seems to me incapable of piercing the cuticle of Halteria.
In neither fixed nor living material have I found any evidence
of a dart-like structure at the tip. The tentacles appear to

ale JULIA ELEANOR MOODY

me to be fully extended soon after the animal settles down to
the bottom of the culture dish, and not shot out on the approach
of the prey. This conclusion is based upon many observations
of the living animal under the immersion lens, in which I have
seen no further elongation of the tentacles at the approach of
Halteria. When fully extended, the tentacles are turgid, the
ends packed with the dark granular trichocyst material. It
seems probable that the impact of the soft body of Halteria
against the denser protoplasm of the tentacle is sufficient to
drive the latter into the body of the small ciliate, the trichocyst
material, forced either through a minute pore or from the rup-
tured tip, paralyzing the prey. By contraction of the tentacle
the Halteria is gradually drawn close to the body of its captor
and worked into the mouth by the action of the cilia.

It sometimes happens that the tentacles, stretched out to
a great length, become entirely detached from the body. The
individual represented in figures 3, 4, 5 and 6 was under obser-
vation in a hanging drop for one hour and fifteen minutes. At
the beginning of the time it was normal in appearance, swimming
actively about; at the end of fifteen minutes it came to rest,
mouth downward. Under ordinary conditions the tentacles
are extended gradually from the entire surface of the body,
but in this case, only three or four were extended, these differing
from the normal tentacle inasmuch as they were of enormous
length and sinuous in outline (fig. 43). At this stage Actino-
bolus began to rotate very slowly, the extended tentacles, all
of great length, increasing in number; then swimming slowly
forward, it left in its wake a long trail of cast-off tentacles,
many of which showed a worm-like motion (fig. 44). Whether
this motion was due to the independent action of the tentacles
or to a motion transferred from the cilia to the trailing cluster
of detached organs, I am not prepared to say. After swimming
about for half an hour, Actinobolus, having freed itself of the.
tangle of liberated tentacles, came to rest. For some minutes
the ciliary action was very rapid, but this gradually became
slower and finally ceased. Within a few minutes a break ap-
peared on the periphery of the posterior end which was followed

LIFE HISTORY OF TWO RARE CILIATES 373

by complete and speedy disintegration of the cell. This phe-
nomenon was observed on three occasions (figs. 45, 46).

Von Erlanger’s statement that he was able to trace the par-
tially retracted tentacles for a considerable distance inside the
body, I am unable to verify, as I have found no evidence of
the presence of these organs beyond the cortical region. Sec-
tions of Actinobolus show, scattered throughout the endoplasm,
many faintly staining threads, some very slender, others comma-
shaped, the head of the comma taking the nuclear stain more
intensely (figs. 62, 63, 64, 65). Although the threads closely
resemble the axial filaments of Camptonema pictured by Schau-
dinn, I have in no case found any connection between them and
the nuclear membrane.

Maupas, looking upon the suctorian tentacle as a modified
pseudopodium, traced the origin of the ciliates from the rhizo-
pods through this group. Although bearing a resemblance to
the suctorian organs, the tentacles of Mesodinium, Ileonema
and Actinobolus differ radically from them in structure and
function, and, according to Biitschli, appear to be independent
modifications rather than organs of phylogenetic interest.

The cilia, arranged in clusters of from five to ten at the base
of each tentacle, are long and show a flagella-like motion, rather
than the synchronous beating of ordinary cilia.

At the posterior end of the body, situated somewhat eccen-
trically, is a large contractile vacuole, the pulsations of which
occur at intervals of one minute and a half. Once in some thir-
teen or fourteen beats a minute interval occurs.' Surrounding
the vacuole is a mass of minute dark granules, waste products
of metabolism.

The mouth of Actinobolus does not open directly into the
endoplasm but is separated from it by a funnel-shaped gullet
described by both Entz and von Erlanger, although they do
not agree with respect to its structure. Entz noted the striated
walls of this organ and attributed the ridged appearance to the
presence of folds in the cortex, finding no evidence whatever
of solid rods; von Erlanger, on the contrary, interpreted the
ridges as distinct straight rods forming a weakly built basket.

374 JULIA ELEANOR MOODY

The sketches of longitudinal sections through the mouth shown
in figures 40 and 42, are typical of the structure of this region.
I have found no straight rods; in every case the sections show
wavy lines, stained somewhat more deeply than the surrounding
protoplasm and having the appearance of cortical thickenings.

A rod-shaped macronucleus is imbedded in the endoplasm
which, in different individuals, shows great variety in form,
length and position; sometimes lying fully extended and fol-
lowing the circumference of the cell; again straight or loosely
coiled in the center of the body; sometimes so tightly twisted
as to appear made up of closely united segments (figs. 49, 50,
51, 52). In may have been some such closely twisted form as
the latter which Entz described as composed of separate spheri-
‘al segments, appearing like a mass of free nuclei.

In section the nucleus is seen to be covered with a delicate
but distinct membrane, within which, imbedded in an achro-
matic substance, lie coarse, deeply staining bodies (figs. 42, 47,
53, 65). Seattered throughout the endoplasm are bodies, some-
times spherical, sometimes elongated, many of which are solid
masses staining evenly throughout, while others show a differ-
entiation into several intensely staining granules, surrounded
by a pale area of definite outline. In one section a comma-
shaped body was found made up of distinct granules which
appeared to be in process of division (figs. 47 and 48). Similar
bodies are shown in figures 54 and 60, sketches of total mounts.
It seems impossible from the present observations to determine
the nature of these bodies. Although some of them may be
micronuclei, I am inclined to interpret them as nuclei of partly
digested Halteria, owing to the fact that there is no constancy
in number and that none of them has been observed in process
of division.

4. Reproduction

Actinobolus radians reproduces by transverse fission, the
number of daily divisions showing a wide variation due to the
extreme sensitiveness of the organism to food and environmental
conditions. During division Actinobolus swims actively about,

LIFE HISTORY OF TWO RARE CILIATES 375

settling down from time to time to feed; this process which
occupies about one hour and a half is preceded by an elonga-
tion of the body and nucleus, the difference in width between
the anterior and posterior regions becoming more marked as
the constriction deepens between the two ends (figs. 54 and 60).
The difference in size between the anterior and posterior body
is especially well seen when the animal comes to rest mouth
downward, the posterior cell resting on top of the anterior one,
the peripheries, in optical section, appearing as concentric circles.
About half-an-hour after the animal begins to divide, a vacuole
is formed anterior to the constriction and a little to one side
of the median plane of the body. It is difficult to determine
the exact location of the new mouth in the dividing cell, owing
to the fact that while the organism is in motion it is impossible
to keep it in focus long enough to find the mouth, while, when
it comes to rest, this part of the body is directed downward
and cannot be seen. When the two cells separate, however,
the mouth is readily found at the extreme anterior tip of the
body Gigs. 55, 56, 57, 58, 59; 61).

5. Summary

(a) Actinobolus radians is an almost spherical organism vary-
ing in length from 37.5, to 77u, in width from 22.54 to 66.5y; the
average length and width of 25 individuals measured was 58.5y and
46.8u respectively. It feeds exclusively on Halteria grandinella.

(b) Two clearly differentiated regions are distinguishable in
the living cell, a dense central mass, the endoplasm, surrounded
by a semitransparent envelope, the cortex. Large vacuoles
form a characteristic peripheral border.

(c) The cuticle is marked by twenty four or five delicate
lines extending spirally from the borders of the mouth to the
posterior end of the cell and marking the places of insertion of
the retractile tentacles and the cilia.

(d) The cilia are arranged in clusters of from five to ten at
the base of each tentacle and show a flagellum-like motion,
rather than the synchronous beating of ordinary cilia.

376 JULIA ELEANOR MOODY

(e) The tentacles, extremely slender organs, are of equal
thickness throughout their length; the tip, which is slightly
rounded, never pointed or knobbed, is quite opaque, owing to
the presence of a mass of minute dark granules, the trichocyst
material. The tentacles, greatly extended, are sometimes en-
tirely detached from the body; this phenomenon, in the three
eases observed, was followed by speedy disintegration of the
organism. ‘These organs, readily traced into the cortical region,
have not been observed in the endoplasm.

({) The single terminal vacuole pulsates at intervals of one
minute and a half, a one minute interval occurring once in
some thirteen or fourteen pulsations.

(¢) The mouth opens into a funnel shaped gullet, the walls
of which are strengthened by longitudinal folds of the cortex.

(h) Great variation is seen in form, length and position of
the rod-shaped macronucleus; sometimes it les, fully extended,
parallel to the circumference of the body; again, straight, loosely
coiled, or tightly twisted in the centre of the cell. The nucleus
is covered with a delicate membrane within which, imbedded.
in an achromatic substance, lie coarse, deeply staining chromatin
granules which in division stages are greatly elongated.

(i) Spherical and comma-shaped bodies composed of deeply
staining granules imbedded in an achromatic medium, have been
observed in many sections and total preparations, which may
possibly be micronuclei, but owing to the fact that they have
not been observed in process of division, it seems more probable
that they are the nuclei of partially digested Halteriae.

(j) Actinobolus reproduces by transverse division, the entire
process from the first elongation of the cell to the separation
of the two individuals, occupying about one hour and a half.
The division is unequal, the posterior cell being somewhat smaller
than the anterior cell. About half-an-hour after the beginning
of division, anew vacuole is formed anterior to the constriction
and a little to one side of the median plane of the body. The
number of daily divisions shows considerable variation owing
to the fact that the organism is extremely sensitive to food and
environmental conditions. Eneystment was not observed.

LIFE HISTORY OF TWO RARE CILIATES 317

Ill. GENERAL CONSIDERATIONS
A. The life-cycle, senescence and rejuvenescence

In 1838 Ehrenberg as a result of his studies of the infusoria
concluded that because of their simple structure and method
of reproduction, natural death was impossible. This view was
opposed by Dujardin who maintained that the life-history shows
definite cycles, indicative of successive periods of protoplasmic
vitality which eventually end in death.

If the succession of cells formed between two consecutive
periods of conjugation is comparable to a metazoan, the study
of the cell aggregate, represented in the life-cycle is essential
to complete knowledge of the morphological conditions of the
species. It often happens in the course of the life-cycle that
morphological variations occur, the indications of marked differ-
ences in protoplasmic vitality, which, unless the life history has
been carefully followed, may lead to confusion in regard to the
identification of the forms under observation. For this reason
many observers within the last half century have made the
life-history of various infusoria the subject of careful study,
thereby adding much to our knowledge of these forms.

Among the earlier workers were Biitschli (’76) and Englemann
(78) who were follwed by Maupas (’88) the first to make an
exhaustive study of the life-cycle. Schaudinn, by his valuable
work on Coccidium schubergi, emphasized the importance of
an intimate knowledge of the life-cycle of every species.

Interesting observations have been made since 1900 on Para-
moecium caudatum by Calkins (’02 and ’04); on Oxytricha
fallax by Woodruff (’05); on Gastrostyla mytilus by Popoff
(07); on Tillina magna by Gregory (’08), and on Paramoecium
aurelia by Woodruff (09 and ’11). With the exception of
Woodruff’s last work, these infusoria were bred in a more or
less constant food medium. As a result of this method of treat-
ment the life-history showed a division into cycles of varying
vitality, measured in terms of the division rate. Enriques in
his paper of 1908 criticized the results of these experiments and
the conclusions based upon them. He denied the existence

378 JULIA ELEANOR MOODY

of senile degeneration followed by physiological death, claiming
that the results obtained by Calkins and Popoff were due to
faulty technique resulting in bacterial invasion of the cultures.
Calkins’ methods, fully given in his paper of 1902, show that
a careful technique was followed throughout his work; the cycles
which he observed during the 742 generations of Paramoecium
must therefore have been due to a cause other than the one
suggested by Enriques.

Spathidium shows a rhythmic rise and fall in division-rate
corresponding to the vitality of the cell and in this respect agrees
with the results obtained by Calkins, Woodruff, Popoff and
Gregory for Paramoecium, Oxytricha, Gastrostyla and Tillina.

Reference to diagram 1 will show that Spathidium responded
but slightly to treatment with salts. During the period when
the division rate fell suddenly from 3.1 to 1.6, the cultures were
treated with beef extract. The curve shows only a temporary
response, the division-rate gradually Crone to 1.3 during
the eighth period.

Artificial stimulation at this time increased somewhat the
general vitality, which is indicated by the higher division-rate.
That the effect was, however, temporary is shown by the gradual
downward trend of the curve, ending in the death of the series.
In this respect the life-cycle of Spathidium spathula closely
agrees with the results of Gregory for Tillina magna concerning
which she says:

Unlike Paramoecium and Oxytricha, the division rate of Tillina
does not indicate a definite response to treatment with salts. Such
substances, apparently successful in other forms, seem to have been
effective only in raising the vitality slightly above the normal and
increasing it sufficiently to carry the protoplasm through periods of
weakness.

Woodruff defines a cycle as ‘‘a periodic rise and fall of the
fission rate extending over a varying number of rhythms and
ending in extinction of the race unless it is ‘rejuvenated’ by
conjugation or changed environment.” The variations in divis-
ion-rate represented in the life-cycle of Spathidium spathula
may all be interpreted as normal ‘rhythms,’ the entire curve
covering 218 generations, representing but one cycle.

LIFE HISTORY OF TWO RARE CILIATES 379

If rejuvenescence mean a renewal of the vital activities of
the organisms, this process cannot be said to have taken place
in Spathidium. In every case artificial stimulation was fol-
lowed by a slight reaction on the part of the protoplasm, this
effect lasting for a short time only, when the protoplasm either
lapsed into its original condition or showed a somewhat lower

Diagram 1 Complete history of Spathidium spathula, Culture A, from start
(February 24, 1911) to finish (July 7, 1911) in the 218th generation. The rate
of division is averaged for ten-day periods. The ordinates represent the average
daily rate of division for three individuals. The abscissae represent the number
of ten-day periods.

vitality than before. This response must be interpreted, not
as rejuvenescence, in the sense indicated above, but as a tem-
porary response to stimulation, the chemical acting as a spur
to the waning activity of the cell. If however we define re-
juvenescence as a certain power given to protoplasm by means
of which its activities are sustained for a longer period of time

380 : JULIA ELEANOR MOODY

than they would otherwise have been, then to a certain extent
Spathidium was rejuvenated by the beef-juice, by the treatment
of salts, and by the slight change in environment which the
cultures experienced when transferred from hay infusion to tap
water and later into spring water.

Considerable difference was noted in the amount of reaction
manifested by the individuals of the same culture. In some
cases there was absolutely no response to treatment with beef

extract or salts, while other individuals showed varying degrees
of sensitiveness.

If the individuality of the protoplasm of a culture subjected
to the same environmental conditions during a long period of
time show this marked difference in response to stimuli, it seems
reasonable to interpret the effects noted in Paramoecium, Oxy-
tricha, Tillina and Spathidium as identical in nature, differing
only in degree.

Except during the last period represented in the curve of
the life-cycle, Spathidium showed no morphological changes
coincident with the lowered division rate. Through the first
twelve periods of the life history the appearance of the proto-
plasm was normal; Colpidium were consumed and digested in
large numbers. It was only at the very end of the series that
abnormal forms appeared, this abnormality being accompanied
by an extremely dense condition of the protoplasm. Without
doubt it would have been impossible to carry the cultures as
far as the 218th generation hdd not the use of beef extract, salts,
etc., been resorted to. Since Spathidium subsists exclusively
on Colpidium colpoda, of which it had an abundance, the grad-
ual degeneration of the cultures resulting in physiological death
must be traced to a lack in the food medium. The addition
of beef extract and salts of various dilutions changed to some
extent the chemical composition of the medium and as a result
the protoplasm showed a slight response. Owing to the fact
that the cultures, aside from a low division energy, showed no
signs of depression until the last ten-day period, I am inclined
to think the death of the series was due, not to senile degeneration,
but to abnormal conditions of environment and that death

LIFE HISTORY OF TWO RARE CILIATES 381

might have been avoided or at least averted, had the proper
conditions, essential to the well-being of the organism at this
crisis, been found.

B. Nucleus -plasma-relation

There has been a growing tendency in recent years to regard
the protozoan nucleus as composed of two distinct substances,
similar in structure but differing in function; one a trophonu-
clear matter controlling the vegetative functions of the cell;
the other a substance concerned primarily with the reproduc-
tive activities of the organism.

The conception of the double nature of the nucleus is due
to R. Hertwig, who in 1887 discovered a mass of extranuclear
substance in Arcella, forming a band between the two vegeta-
tive nuclei. In 1889 he observed the formation of secondary
nuclei from this extra-nuclear band to which in 1902 he applied
the name ‘Chromidial-netz.’

We are indebted to Schaudinn (’03) however, for the true
interpretation of this chromatin mass. He found that from
this extra-nuclear substance, in the case of Polystomella, Cen-
tropyxis and Chlamydiophrys minute nuclei are formed which
become the nuclei of conjugating gametes. He concluded, there-
fore, that in these forms the chromidial-netz is sexual chromatin
existing in combination with the trophic chromatin during the
vegetative condition.

The work of Hertwig and Schaudinn has been greatly ex-
tended by Brandt (02 and ’05), Prowazek (’04), Goldschmidt
(05), Lister (06), Doflein (07), Calkins (’07), and others. Gold-
schmidt concluded, as a result of his work on nematodes and
protozoa, that every animal cell is binucleate, possessing both
trophic and sexual chromatin. These substances are usually
combined in one body, the amphinucleus, but the separation
of the two elements may be more or less complete.

We may distinguish three distinct degrees of nuclear differ-
entiation among the protozoa; first, a nucleus in which both
functions, vegetative and reproductive are combined, the so-

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

382 JULIA ELEANOR MOODY

called amphinucleus of Goldschmidt; second, a type, illustrated
by Arcella, Centropyxis and others, in which chromatin material
is extruded from the nucleus to form the chromidial-netz of
Hertwig, or the idiochromidia of Mesnil (’05); third, a type in
which there is a complete separation of the two elements into
distinct masses, represented by the macro- and micro-nucleus
of many infusoria.

Although this complete separation of nuclear material is of
common occurrence among the infusoria, it is by no means
universal. An examination of the family Enchelinidae in Biitsch-
li’s ‘“‘Protozoa”’ shows that in the description of fourteen genera,
the presence of a micronucleus is positively stated to occur
only in the case of Didinium. Among the others, micronuclei
have either not been observed or have been little studied. Gold-
schmidt’s opinion that the infusoria show a primitive nuclear
condition is opposed by Dobell, who looks upon the complete
separation of chromatin substance as another indication of the
high specialization shown by these organisms, rather than a
simplification as advocated by Goldschmidt. According to
Dobell, all functions, somatic and propagative, are resident in
the same living nuclear molecule, one or the other predominating
in the course of cell differentiation.

It sometimes happens that the separation of the two mate-
rials is but temporary. Neresheimer (’08) describes the inter-
esting case of Ichthyophthirius, in which the nucleus buds off
a smaller nucleus, which divides, each part undergoing two
reduction divisions. Three of the resulting micronuclei degen-
erate. The fourth divides again forming two micronuclei which
fuse. The zygote thus formed re-enters the original nucleus
and fuses with it. We have here evidently a modification of
the first type described, in which both functions are combined
in the same body.

Considerable confusion has resulted from the use of the term
‘chromidia’ applied by Hertwig to the nuclear substance ex-
truded by Actinosphaerium when starved or overfed. The
chromatin in this case was the direct result of nuclear fragmen-
tation brought about by abnormal conditions in the environment

.

LIFE HISTORY OF TWO RARE CILIATES 383

and was in no sense related to the ‘chromidia’ of Schaudinn,
or the idiochromidia of Mesnil.

I did not find in either Spathidium or Actinobolus any struc-
tures which I felt justified in interpreting as micronuclei. The
only bodies which in any way resembled these organs in struc-
ture were much too large to be so interpreted. In these organ-
isms both vegetative and propagative chromatin are combined
in the long cord-like nucleus, the form and extent of which vary
to a remarkable degree. According to Biitschli, the primitive
nucleus of the ciliate cell was doubtless spherical, a condition
which is found in almost all small ciliates. Using such a type
as a starting point it is interesting to follow the evolution of
this organ through the ellipsoidal condition, to the short band
or rod-shape, on to a greatly elongated nucleus twisted into
a complicated coil or divided into numerous ellipsoidal masses
enclosed in a common membrane, the increase in length in most
cases being coincident with the elongation of the cell body.
In 1903 Richard Hertwig first discussed his theory of the nu-
cleus-plasm-relation, maintaining that for each kind of cell in
a normal condition, there exists a definite relation between
nucleus and protoplasm, upon which the vitality of the cell
depends. He called attention to the fact that this normal cor-
relation between cell-size and nuclear-size is destroyed by changes
in environment, as for-example, an increase or decrease in tem-
perature, starvation or overfeeding, which affect the metabolic
activities of the cell controlling the exchanges of nuclear and
cytoplasmic material.. As a result of this increase of nuclear
material, the cell shows symptoms of senile degeneration which
end in death unless the normal cell-relations are restored, either
by direct elimination of the nuclear material, or by conjugation. .
In support of his own results, based upon the study of Actino-
sphaerium and Dileptus under different environmental condi- .
tions, he cited the conclusions reached by Boveri in the case
of the sea-urchin larvae and by Gerassimow in his work on
Spirogyra.

In 1902 and 1905 Boveri studied the relation of cells and
nuclei in sea-urchin larvae containing X, 2 X, and 4 X numbers —

384 JULIA ELEANOR MOODY

of chromosomes, the X larvae, the product of artificial par-
thenogenesis or the fertilization of an enucleated egg-fragment
and the 4 X larvae obtained by shaking the eggs soon after
fertilization. As a result of the investigation, Boveri formulated
two laws based upon the comparative measurements of these
larvae which are as follows: ‘‘The surfaces of the nuclei are
directly proportional to the chromosome number and _ hence
to the chromatin mass,” and ‘‘The size of the cell in the sea
urchin larvae is directly proportional to the chromosome number.”’
Gerassimow (’01 and ’02) found that, if in cell-division of Spiro-
gyra, the daughter nuclei were caused to remain in one of the
new cells, this cell in consequence grew abnormally large, and
he concluded, therefore, that the size of the cell is dependent
upon the size of the nucleus.

Summing up the results of the work of Gerassimow (01, ’02),
of Boveri (05) and of Popoff (07), Hertwig in 1908 says:

Das Neue, welche in der Lehre von der Kernplasma-Relation ge-
geben ist, ist der Gedanke, dass der Massenverhaltnis von Kern zu
Protoplasma, der Quotient ,, d. h., Masse der Kernsubstanz dividiert
durch Masse des Protoplasma, ein gesetzmissig regulierter Faktor
ist, dessen Grosse fiir alle von Kerne beeinflussten Lebensvorginge
der Zelle, von fundamentaler Bedeutung ist.

In this recent discussion of the kernplasma relation he dis-
tinguishes two definite periods of nuclear growth occurring in
the interval between consecutive conjugations; first, a ‘funk-
tionelle-Wachstum,’ a period of extremely slow nuclear growth,
accompanied by a rapid increase in volume of the plasma, re-
sulting in a disturbance of the normal kernplasma-relation:
Second, a ‘theilungs-Wachstum’, during which there is a rapid
growth of nuclear material by which the normal relation between
nucleus and plasma is restored, followed by cell division.

Hertwig’s conclusions find support in the recent work of
Popoff (’09) who, from a series of measurements made of Para-
moecium, found that the increase in size of the nucleus and
plasma between two consecutive divisions, measured at intervals
of one hour, did not follow a parallel course. During the first

LIFE HISTORY OF TWO RARE CILIATES 385

hours of the growth period there was a rapid increase of proto-
plasm Accompanied by a very slow growth of the nucleus, this
condition persisting up to within one hour and a half of the
next division, when the sudden and rapid growth of nuclear
material re-established the normal kernplasma-relation.

Hertwig, in describing the relation of nucleus and plasma
in young cells says: ‘“‘Ich werde im folgenden diesen Zustand
der Kernplasma-Relation, mit welchem die Zelle in eine neue
Phase ihren Existenz eintritt, die Kernplasma-Norm nennen.”’
He designates by the term ‘Kernplasma-Spannung”’ that period
in the life of the cell when the volume of the nucleus departs
from the Kernplasma-Norm. With a view to testing the valid-
ity of Hertwig’s theory, careful measurements have recently
been made of nucleus and plasma in Tillina magna by Gregory
(08); in sea-urchin larvae by Erdmann (’08); in Oenothera
lamarckiana and Oenothera gigas by Gates (’09); and in Crepi-
dula and Fulgur by Conklin(’11). Gregory concludes:

These facts seem to prove that there is no relation between the amount
of nuclear material in the cell and the general vitality of the proto-
plasm. In other words, the periods of weakness are not caused by
an excess of nuclear material. The nucleus may or may not increase
in size during periods of low activity; if an increase does take place,

it is generally found that the cytoplasmic material has increased also,
and the ratio between the two is the same as in periods of high activity.

Erdmann studied the effect of temperature on the cell size,
finding the chromatin volume varied with the temperature and
concluding that the chromatin mass and not the number of
chromosomes was concerned, and that the cell-size was approxi-
mately proportional to this mass. Gates, on the other hand,
decided that the larger sized cells of Oenothera gigas result
from a doubling in the number of chromosomes and not merely
from an increase of chromotin mass. Conklin, in summing
up the results of his observations on Crepidula and Fulgiir says;

In different eggs, corresponding blastomeres have approximately
the same kernplasma-relation; but in different blastomeres of the same
egg or of different eggs the kernplasma-relation is neither a constant
nor a self regulating ratio. It appears to be a result rather than a

cause of the rate of cell-division and consequently a variable rather
than a constant factor.

386 JULIA ELEANOR MOODY

In order to discover and analyze the nucleus-plasma relations
existing at different periods of the life-history of Spathidtum and
Actinobolus, I made measurements of many cells, in different
stages of cell growth, which had been fixed, mounted and stained
at different times during the life-history of the organisms. My
method consisted in drawing careful outlines of cells and nuclei
with the aid of a camera lucida at a magnification of 630 di-
ameters, and then making measurements of the projections.
The measurements of lengths and diameters used in the tables
are given in microns. In order to obtain a volume of Spathidium
which was approximately accurate, I used as diameter the aver-

TABLE 1
New cells of Spathidium spathula

; |
| DIAMETER LENGTH |\VOLUME OF CELL

[Por can | or cen. | eg, | wottaus | OF UcuBUs | MENUS VOLUME) “aeramrow

eae ae | a ie. te |
Ti; 9 odes Aen 6 104 2940'08 | 20612.4 | 1:6.1
Dae com 92 5 64 | 1192.96 | 30672.16 | 1:25.6
ea ie Oped es 79 1472.56 | 28476.34 1:10.9
4 22 60°) <6 90 2544.3 20200.7 "| 038
By) wee | 5 98 | 1826.72 | 27625.48 Isis
Gal 27 GON | Feo) mabe (1025.22 ey yeas oat ibs
Tar 28M, NN ZO 5 nase 3541.6 | 38401.87 1:10
Bilavreag 103 5 158 2945.12 | 36092.0 es
9). 27 71 6, | Sebag 3364.13 | 37283.4 nee
10| 30 79 6 | 158. | 4466.66 | 52230.6 1:11.6
lt | 29 63 5 186 3467.04 | 36117.12 1:10.4
12S ORD aL 5 143 2665.5 | 37982.0 1:14.2
13 a7) | 68 6 222 6276.9 61461.3 1:9.7
14]. + B86 ¢) 269 6 127 3590.2 | 55425.5 1:15.4
15 20 %|. 98 5 87 1621.68 | 29166.0 | 1:17.9
16 20 1200 5 79 1472.56 | 26801.84 | 1:18.2
17 7) es 5 Ae 58a 1565.76 | 35498.24 | 1:22
13 |): + 20". 9, -48 5 111 2069.04 | 19293.84 | 1:9.3
19 27 | 36 5 95 1770.8 |, 29716.7 | 1116.6
208) 5 27-0 ye! Baers 95 1770.8 | 32579.2 | 1:18.38
Divi 2rs ui tax Oak Mee 8 127 897.64 | 35169.86 | 1:39
22; 31 | 60 | 6 174 4918.98 | 41367.22 | 1:8.4

Average Average Average |Average | Average Average | Average

2545) 72 5a, 12414 | 2611.19 | 34795.85 | 1:13.3

Average coefficient = 112

LIFE HISTORY OF TWO RARE CILIATES 387
age of three measurements taken, one through the thickest
region of the cell and two near the extremities. The volume
was ascertained by treating both cell and nucleus as cylinders.
The formula ; 7 7 * was used in obtaining the volume of Actino-
bolus, the cell being treated as a sphere. All volumes are ex-
pressed in cubie microns.

In table 1 are given the measurements of twenty-two cells
fixed stained and mounted immediately after division. In these
new cells which have just begun to nourish themselves and to
grow we find, according to Hertwig, in the relation existing
between nuclear and protoplasmic volume, the Kernplasma-
Norm. An examination of these ratios given in the last column
of the table will show a wide variation in the nuclear-plasma
relation, ranging from 11:6 to 1:39, giving an average norm of
1:13.83. In table 2 are recorded the measurements of sixteen

TABLE 2
Vegetative-cells of Spathidium spathula

DIAMETER LENGTH , VOLUME OF CELL)
roresut | oFeeut | ug | wpeens | OF NUcUmUS [MINUS VoLUME| SEENTAss
M Me Me Me

1 35 127 6 157 4438 .4 93371 .27 1213
2 25 111 5 190 3541.6 50945 .97 £143
3 27 127 1.6 286 575.0 132775 .0 1:230.9
4 27 132 6 175 4947 25 70623 .75 1:14.2
5 24 135 3 292 2064.0 58956 .0 12825
6 28) j)-s130 5 282 5256.48 74791 .02 1:14.1
a 32 114 6 222 62.7594 85407.42 11326
“eS 19 106 3 95 671.46 29381 . 66 1:45.7
9 20 132 2 325 1021.02 40448 .1 1300
10 41 139 3 301 2127.47 172545 .49 1:85.8
11 20 122 2 325 1020.5 37307 .02 1:336:5
12 23 DUI 5 1a 2069 . 04 44048 .13 PALA
13 29 122 5 238 4436.32 72218 .72 PGES
14 32 103 3 174 1229.8 81607 .92 1:66.3
15 35 99 8 95 4774.7 90474 .19 1:18.9
16 29 106 5 214 3989 .96 62601 .96 131526
Average |Average |Average Average | Average Average Average
27.5 119.3 4.1 217.8 3027 .37 74810. 23 1:24.6

Average coefficient

82

388 JULIA ELEANOR MOODY

vegetative cells taken at random from the cultures. Although
a comparison of tables 1 and 2 shows an increase in length and
diameter of both nucleus and cell, an examination of the ratios
given in the last columns of these tables, indicates that the
growth of nucleus and plasma in the vegetative cells has been
unequal: that the normal kernplasma-relation has been dis-
turbed in favor of the plasma giving an average ratio of these
cells of 1:24.6. A comparison of the average coefficients at
these two stages of growth is of interest. This result was obtained
by expressing a ratio between the length and diameter of the
cell and the length and diameter of the nucleus as for example

— in which Z and D indicate the dimensions of the cell and

l and d those of the nucleus. An increase in size of this resul-
tant indicates an increase in size of the nucleus, a decrease in
the value of the coefficient, a corresponding decrease in the size
of the nucleus. Comparing tables 1 and 2, we find a decrease
in the average coefficient corresponding to the increased kern-

TABLE 3

Division-stages of Spathidium spathula

DIAMETER LENGTH |VOLUME OF CELL, |.

POT | eS Tron octnus | oF teens | MRUR YOLEN) “tron

[ea row B MK |
1 24 154— || 4.7 203 3369.8 75238 .2 12253
2 DA We MST 5 209 3895.76 67068 . 24 WEI Ffais)
3 Dost We Mas 6.3 219 | 66012.8 | 47594.2 17,2
4 22: | 1438 5 V7iGe |) o280),0 50916. 4 albeo
5 a0) © =) 160 8 277. | 13922 99175 .6 1:7.9
6 29 | 138 5 206 3839.4 | 82868.0 12S
7 29 | 149 5 241 | 4492.24 | 89127.44 1:19.8
Beichs ay 143 5 Berni). 5237 Bal, pe 7GR29%7 1:14.5
9 | 24 162 5 582 10848 .48 62375.5 53 7/
10 | art Aa. 162 3 230) | 11625278 65680 .37 1:40.3
11 | 24 | 169 3 346 | 2445.52 | 28742.5 ete7
12 24 | 171 5 222 | 4027.08 | 73265 .0 1:18.1

| = | —

Average eens Average Average | Average | Average Average

| 255 154} 5 266 5298.97 | © 68223.43 11258

|

Average coefficient = 112

LIFE HISTORY OF TWO RARE CILIATES 389

plasma-relation on favor of the protoplasm. In table 3 are
given the measurements of twelve cells in division stage, the
average kernplasma-relation 1:12.8 showing a return to the
-kernplasma-norm, the average coefficients recorded in tables 1
and 3 being identical. An analysis of the relations existing
between nucleus and protoplasm in new cells, vegetative cells,
and division-stages show, first, that the nucleusplasma-norm
expressed by the ratio 1:13.3 is destroyed during the period of
time immediately following division, the ratio between the nu-
cleus and protoplasm being almost doubled in favor of the latter:
second, that in the division stages the norm is re-established.
These conclusions based on average measurements, agree in
the main with the results of Hertwig and Popoff.

- Popoff (08), in his study of Frontonia leucas, in discussing
the relation between the depression periods and size variations
says, ‘The animals show during this period a marked decrease
in cell size and a disturbed kernplasma-relation in favor of the
nucleus.’’ And again, ‘‘the researches of R. Hertwig in Actinos-
phaerium, Dileptus and Paramoecium, those of Calkins on Para-
moecium and of Woodruff on various hypotrichous ciliates, and
my own work in Stylonychia proves that the protozoa, through
a period of uninterrupted activity, come into a condition in
which the nucleus shows an abnormal growth.’ In regard to
Calkins’ work, Popoff took for granted that the greatly enlarged
nucleus was the cause of the depression. Since Calkins, in
his paper of 1904, makes no reference to the kernplasma-relation,
Popoff must have drawn his conclusions from the photographs
of the cells and did not take into consideration the fact that
the cells had not divided for several days and were therefore
in an abnormal condition.

In table 4 I have recorded the length, diameter and volume
of protoplasm and nucleus of newly divided cells, vegetative
cells and division stages chosen from two periods in the life-
history of Spathidium. The cells represented in the first column
were taken from the cultures during a ten-day period in March
when the division energy was high, averaging for the period
3.1. The figures of the second column are measurements of

390 JULIA ELEANOR MOODY

cells selected during a ten-day period in May when the division-
rate had fallen to 1.3 for this interval. If the division energy
be an index of protoplasmic vitality, the cultures in May must
have been in a less healthy condition than in March. From
May to the death of the series, there was a general decrease in
division-rate, averaging but 0.4 during the last ten-day interval
in July. Reference to table 4 shows that in May, when the
division energy was low and the cultures had already entered
on a period of depression, both the coefficient and kernplasma
relation indicate a decrease in nuclear volume, rather than the
abnormal nuclear growth claimed by Popoff in his paper of 1908.

TABLE 4

Spathidium spathula

NEW CELLS RESTING CELLS | DIVISION STAGES
| March | May March May | March May
Average diameter of be b pb ea ros Se est
NUCLEUS Hera eT er 55 4t 4.3 4.3 | 5t 4
Average diameter of ; |
Celia reenoee octamer 23 243 28 38 23 24
Average length of nu- | | |
CLUS et eth manera 812 107 |W 220 216 | 202 345
Average lenght of cell | 693 72 W126 117 | 149 | 166
Cocfiicient: 2.02.04. 4 198 134 83 §1 | 403 Meee.
Kernplasma-relation.!.| 1:14.6 | 1:15.4 | 1:24.4 |143 | 1114.4 | 1:16
| | |

Division rate in March = 3.1
Division rate in May = 1.3

Hertwig maintained that an increase in nuclear mass led to a
slowing of the division rate. Table 4 shows that this 1s not
true of Spathidium, where a slow division-rate is coincident
with a decrease of nuclear material.

A careful study and comparison of the nucleus plasma-relation
of individual cells shows a wide variation at all periods of the
life-history; for example, the measurements of one cell, fixed
during March, when the division-rate for the period was 3.1,
give a kernplasma-ratio of 1:230 in favor of the protoplasm, |
also the average of ratios during a period of low division energy

LIFE HISTORY OF TWO RARE CILIATES 391

indicates a reduced nuclear volume. On the other hand, cells
with a ratio very close to the kernplasma-norm were abundant
during periods of reduced division energy. The variations
noted in Spathidium are seen by consultation of table 5, to be
also true for Actinobolus. These facts therefore, seem to indi-
eate that the kernplasma-relation is not a constant quantity
and bears no definite relation to the division-rate. In spite
of the wide departure from the kernplasma-norm shown in the
_vagetative cells recorded in table 2, this divergence, resulting
in a disturbed kernplasma-relation has not resulted in division
of the cell. It seems therefore that division in Spathidium

TABLE 5

Vegetative-cells of Actinobolus radians

| = VOLUME OF CELL] _
So eae DE Oa) Cee | amipse vor usee ONE aa
u Me B ‘

1 47 6 143 50965 46922 1E1I6

2 47 6 119 50965 47601 1:14.1

St ao 6 135 28730 24914 INS. (a0)

4 55 6 87 $2448 79989 1:32

5 47 6 148 50965 46922 1:11-6

6 39 9.5 119 28730 26112 1:10

7 39 4.7 150 28730 26075 IEORS

8 | 39 4.7 142 28730 26501 IEE fs}

9 67 6 174 150532 145617 1:29.5
10 | 55 3.9 174 $2448 78620 1:20.5
11 3l 3 87 14137 13528 1:22
12 47 6 ee 9D 50965 48279 elie
13 43 4.7 103 38792 37175 - 1:22.9
14 al 4.7 174 179595 176863 1:64.7
15 55 6 190 82448 T7077 1:14
16 39 6 95 28730 26049 1:9..7
Wz 63 6 206 124789 11930 PME AL 20. 7
18 55 4.7 135 82448 80329 | 1:37.7
19 39 4.7 135 28730, PGi sens 1312
20 | 63 4.7 135 124789 | 122670 | 1:57.8
21 | 39 4.7 103 28730 27113 16.7
22 47 4.5 158 50965 #7489" |) 1313.6
23 30 4.5 111 20580 LSISSH sz
24 By) ead 166 $2448 Magan = Ve1GES
25 47 | 4.5 166 50965 C78 1 a Mi ler

= + ! >

Average kern-plasma relation = 1:20.

392 JULIA ELEANOR MOODY

cannot be traced to a disturbed relation between nucleus and
protoplasm. These varying ratios found in cells of the same
period in the life history of the organism, may be explained
on the basis of differences in cell metabolism. Cultures sub-
jected to comparatively constant environment showed wide
individual differences in response to artificial stimulation and
it seems probable that the variations in the kernplasma-relation
are individual variations due to a difference in response to en-
vironmental conditions, since the interchange of nucleus and
protoplasmic material is dependent upon the metabolic activity
of the cell. Conklin states in his paper on “Cell size and nuclear
size” that ““As we have seen, the kernplasma-relation varies
widely in different blastomeres of Crepidula and Fulgur. In
these cases wide departures from the kernspasma-norm have
not brought on cell division and if the kernspannung is a cause
of cell-division, it must be a minor factor in the case.’”’ From
the evidence at hand therefore it seems impossible to trace
the phenomena of cell division to any of the causes thus far
suggested by Spencer, Strasburger, Boveri or Hertwig. Cell-
division is the index of protoplasmic vitality which is directly
dependent upon the metabolic activity of the cell. If therefore,
we accept Child’s interpretation of senescence and rejuvenes-
ence, we must attribute the varying relations between nucleus
and protoplasm to individual differences in metabolic activity
expressed in the division-rate.

Strasburger (’93) accounts for cell division on the basis of
the ‘working-sphere’ of the nucleus. The experimental work
of Brandt (’77), Nussbaum (’84), Gruber (’85), Lillie (’96),
Balbiani (’89 and 791) and Verworn (’89) all go to show that
the nucleus is indispensable to the formative energy of the cell;
that although the processes of destructive metabolism may
continue for a long time in an enucleated cell, that constructive
metabolism is only possible in the presence of the nucleus.

Considering these facts in connection with Strasburger’s theory
of the ‘working-sphere’ of the nucleus, the ratio existing between
the nuclear surface and the protoplasmic mass must be of great
significance in the interchange of nuclear and protoplasmic

LIFE HISTORY OF TWO RARE CILIATES , 32398

material. The maintenance of a certain ratio between the two
masses, making possible a normal interchange between them
of material, seems to me to explain the greatly elongated nuclei
found in Spathidium and Actinobolus.

C. Food habits

Frequent observations of the feeding habits of Actinobolus
and Spathidium are suggestive of many interesting questions.
In these two forms we have organisms subsisting exclusively
so far as we know, upon a special type of ciliate. Actinobolus,
moored by means of its oral tentacles, awaits the coming of
its victim, Halteria grandinella, before making use of its weapons
of offense, Spathidium, a predatory form, swims actively through
the water, its anterior end in constant motion, passing with
seeming indifference all food material except the little ciliate,
Colpidium colpoda. This behavior on the part of the organism
would seem to indicate an apparent exercise of choice, but when
one comes to a careful analysis of the basis of this choice, one
finds himself becoming involved in questions upon which the
observations of many other workers have thrown but little
light.

Interest in the feeding-habits of unicellular organisms, result-
ing in practical experimentation dates back to the time of Gleichen
in the latter half of the eighteenth century. The question
which arose in the minds of these early observers in the field
of protozoology was one which has been discussed by almost
every worker on the protozoa since that time, namely: Have
these simple forms of animal life the power of selection in the
kind of food-material upon which they live,—if so, upon what is
the choice based? Many of the experiments of Gleichen show-
ing that large quantities of finely powdered carmine or indigo
mixed with the water, were taken into the body of these uni-
cellular forms, were repeated and confirmed by Ehrenberg in
1838. Stein, Entz and many other workers, however, either
forgot or ignored these early results, since they expressed the
concensus of opinion at that time that infusoria showed a de-
cided power of selection, inasmuch as from the mass of sub-

394 JULIA ELEANOR MOODY

stance swept into the pharynx, food-particles were passed into
the endoplasm while all foreign substances were rejected by a
reversal of ciliary motion. This question was again taken up
by Verworn in 1889 who, after experimenting on Vorticella
with chalk erystals, carmine and indigo concluded ‘‘wenn man
die Bezeichnung Auswahl fiir einzelne dieser Erscheinungen
beibehalten will, sich klar machen miissen, dass darunter keine
bewiisste Auswahl in einer bestimmten Absicht zu verstehen
ist, sondern ein vollig unbewiisster Vorgang, dhnlich der natiir-
lichen Auswahl, der Selection, die der Kampf um’s Dasein hervor-
bringt.”’

Hodge and Aiken in 1893 published a paper on the “ Daily
Life of a Protozoan” with the significant sub-title ‘““A Study
in Comparative Psycho-Physiology.” Here again Vorticella
was the subject of the experiment, chosen because it is easily
kept in the microscopic field during long periods of time. The
fact that the same animal subjected to similar conditions of
experiment shows different reactions at the hands of two skill-
ful observers, leading to diametrically opposed conclusions is
of interest. Hodge classed the action of the cilia under psycho-
reflex movements, assigning to them a sorting or discriminating
function, dependent apparently upon a touch sensation and
concluded that this process indicated a no less conscious action
on the part of Vorticella than the seeking of prey and the feed-
ing of animals in general.

Schaeffer, in 1910, published the results of some exceedingly
interesting feeding experiments on Stentor. Every test was most
carefully controlled in order that the exact kind and quantity
of food might be accurately recorded. Without going into the
details of the experiments I will simply quote his results which
are of interest and importance in the light of other experimental
work on the feeding habits of Protozoa. Schaeffer found that,
as food, Stentor preferred Euglena to Phacus; that it could
discriminate between Phacus triqueter and Phacus longicaudus,
also between Trachelomonas hispida and Trachelomonas volvo-
cina; that it manifested no choice between living organisms
and those killed in chemicals, as for example osmic acid, iodine

LIFE HISTORY OF TWO RARE CILIATES 395

or alcohol; that whole specimens were eaten while a jelly com-
posed of crushed Paramoecia and Euglena was rejected. As
a logical deduction from these results Schaeffer concluded that
Stentor does exercise a selection among particles which are
brought into the food pouch; it discriminates between digestible
and indigestible particles and between different kinds of organ-
isms. What then is the basis of this choice? Making a dis-
tinction between taste, a reaction to chemicals, and touch, a
reaction to form, he concludes that inasmuch as Stentor ate
alike living forms and those fixed in chemicals while it chose
entire organisms in preference to macerated ones, that the selec-
tion must be made on a tactual basis.

Didinium, a predatory ciliate, appears to the casual observer
to exercise a somewhat limited choice. Jennings, however,
explains the food habits of this form by the trial-error theory,
claiming that Didinium reacts not only to particles which may
serve as food but to all kinds of solid bodies, that it is constantly
coming into contact with various substances digestible and indi-
gestible, each one of which it ‘tries’ to pierce and swallow. If
successful the particle is chosen as food, if the trial be unsuccess-
ful it is rejected and classed with the errors. Jennings says
“there is no evidence that in some unknown way the infusoria
perceive their prey at a distance nor that they decide beforehand
to attack certain objects and leave others unattacked. They
simply prove all things and hold fast to that which is good.”
In this conclusion Jennings agrees with Maupas.

It is evident, however, that the trial-error explanation is
inadequate in dealing with the behavior of Actinobolus and
Spathidium. It is true that when the organisms are in a con-
dition of satiety, the small ciliates Halteria and Colpidium may
pass their enemies unharmed; if, on the other hand, Actinobolus
and Spathidium are hungry, no attempt is made to eat any other:
_ kind of food than the accustomed prey. We find here no ‘trials’
but a definite selection of food material. In experimentation
with unicellular forms under identical conditions not only has
a difference in reaction between different individuals been noted,
but also difference in reaction of the same individual at differ-

396 JULIA ELEANOR MOODY

ent times. These reactions may be accounted for by an ex-
planation on the basis of chemical or physical laws of attraction
or by assigning to the organism something akin to intelligence.
The question naturally suggests itself, how has the exercise of
choice become limited to one special form of food material in
the case of Actinobolus and Spathidium? If, as Jennings
claims, the avoiding action of the cilia in Paramoecium is in
itself an expression of choice, it is not difficult to conceive of
a still further development of this expression as illustrated in
the reversal of the ciliary current, the bending of the body
of the organism and the swimming away from foreign substances
or undesirable food material. Accepting the definition of instinct
as purposive action without consciousness of purpose, it seems
legitimate to apply the term in this sense to the phenomena
involved in the food taking behavior of protozoa. The word
thus used is simply a convenient term for expressing the outward
manifestation of a relation existing between the protoplasm
of an organism and an external stimulus or another form of
protoplasm possessed of different chemical or physical proper-
ties. It may be that the behavior of the early ancestors of the
hunter ciliates was based on the method of trial. In the course
of time, the protoplasm of the organism has become modified
chemically and physiologically to such an extent that a reaction
to one kind of protoplasm only is possible—in other words
forms like Actinobolus and Spathidium have become ‘educated’
through ‘error’ to the selection of one species of food, namely,
Halteria grandinella and Colpidium colpoda.

March 9, 1912

LITERATURE CITED

Bavsrant, E. G. 1889 Recherches expérimentales sur la merotomie des in-
fusoires ciliés. Recueil Zool., Suisse, tom. 5.

1891 Sur les régénérations successive du péristome chez les Stentors
et sur la rédle du noyau dans ces phénoméne. Zool. Anzeiger, Bd. 14.

Boveri, T. 1905 Zellenstudien V. Ueber die Abhingigkeit der Kerngrosse
und Zellenzahl der Ausgangszellen.

Branpt, H. 1877 Uber Actinosphaerium Eichhorni.

LIFE HISTORY OF TWO RARE CILIATES 397

Branpt, K. 1902 Beitrige zur Kenntnis der Colliden IT und II. Archiv fiir
Protistenkunde, vol. 1.

Birscuur, O. 1883 Protozoa. Bronn’s Klassen und Ordnungen des Thier-
Reichs.

Cauxins, G. N. 190la Some Protozoa of especial interest from Van Cortlandt
Park, New York. Am. Naturalist, vol. 35.

1901b The Protozoa. Columbia University Press.

1902a Studies on the life-history of Protozoa. II. The effect of stim-
uli on the life-cycle of Paramoecium caudatum. Archiv fiir Protist.

1902b III. The six hundred and twentieth generation of Paramoecium
caudatum. Biol. Bulletin, vol. 3.

1904 IV. Death of the series: Conclusions. Jour. Exp. Zool. 1.
1906 The protozoan life-cycle. Biol. Bulletin, vol. 10.

1907 The fertilization of Amoeba proteus. Biol. Bul., vol. 13.
1910 Protozodlogy.

191la Regeneration and cell-division in Uronychia. Jour. Exp. Zool.
vol. 10.

1911b Effects produced by cutting Paramoecium cells. 3iol.Bull.,
vol. 21.

1911le The scope of protozoology. Science, vol. 34.

Cup, C. M. 1911 A study of senescence and rejuvenescence based on experi-
ments with Planaria dorotocephala.. Arch. F. Entw.—Mech., Bd. 31.

Conxkuin, E. G. 1912 Cell size and nuclear size. Jour. Exp. Zool., vol. 12.

Dose, C. C. 1909 Chromidia and the binuclearity hypothesis. Q. J. Mic.
SComVOlemne:

DortEIn. 1907 Fortpflanzungserscheinungen bei Amoeben und verwandten
Organismen. SB. Ges. Morph. Physiol. Miinchen, Bd. 23.

Dusarpin, F. 1841 Historie naturelle des Infusoires.
EHRENBERG 1838 Die Infusionsthierchen als vollkommene Organismen.

Enriqures, P. 1908 Die Conjugation und sexuelle Differenzierung der In-
fusioren. Archiv f. Protistk, vol. 12

Entz, G. 1879 Uber einige Infusorien des Salzteiches zu Szamosfalva.

1883 Beitrage zur Kenntnis der Infusorien. Zeitsch. f. wiss. Zool.,
Bd. 38.

VoN ERLANGER’ 1889 Ziir Kenntnis einiger Infusorien. Zeitsch. f. wiss. Zool.,
Bd., 49.

ERDMANN, Ru. 1909 Experimentelle Untersuchung der Massverhiltnisse von
Plasma, Kern und Chromosomen in dem sich Sane eyes Seeigelei.
Arch. f. Zellforschung, Bd. 2.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

398 JULIA ELEANOR MOODY

Gates, R. R. 1909 Stature and chromosomes of Oenothera gigas. Arch. f.
Zellforschung, Bd. 3.

Grrassrmow 1902 Die Abhingigkeit der Grésse der Zelle von Menge ihrer
Kernmasse. Zeitsch. f. Allgem. Physiol., Bd. 1.

Goupscumipt, R. 1905 Die Chromidien der Protozoen. Arch. f. Prot, vol. 5.

Grecory, L. H. 1909 Observations on the life-history of Tillina magna. Jour.
Exp. Zool., vol. 6.

Gruser, A. 1885 Uber kiinstliche Teilung bei Infusorien. Biol. Centralblatt,
vol. 5.

Hertwic, R. 1887 Uber die Kernteilung der Infusorien. SB. Ges. Morph.
Physiol. Miinchen.

1899 Uber Encystierung und Kernvermehrung bei Arcella vulgaris.
Festschrift f. Kupffer.

1902 Die Protozoen und die Zelltheorie. Arch. f. Protistk, vol. 1.

1903 Uber Korrelation von Zell- und Kerngrésse und ihre Bedeutung
fiir die geschlechtliche Differenzierung und die Teilung der Zelle. Biol.
Centralb., Bd. 22.

1908 Uber neue Probleme der Zellenlehre. Arch. f. Zellforsch., Bd. 1.

Jennines, H. 8. 1898 The psychology of a protozoan. Am. Jour. of Psych.,
vol. 10.

1902 Studies on reactions to stimuli in unicellular animals. Am. Jour.
of Physiol., vol 8.

1906 Behavior of the lower organisms.

Linum, F. R. 1896 On the smallest parts of Stentor capable of regeneration.
Jour. Morph., vol. 12.

Lister, J. J. 1906 The life-history of the Foramenifera. Pres. Add. Sec. D.
Brit. Assoc.

Mater, H. N. 1903 Uber den feinen Bau der Wimperapparat der Infusorien.
Arch. f. Protistenk, vol. 2.

Mast, 8.0. 1909 Reactions of Didinium nasutum with specific reference to the
feeding habits and the functions of trichocysts. Biol. Bull., vol. 16.

Maupas, E. 1888 Recherches expérimentales sur la multiplication des infusoires
ciliés. Arch. de Zool. Expér. et Gen.

1883 Contribution a |’étude morphologique et anatomique des in-
fusoires ciliés.

Mesniu, F. 1905 Chromidies et questions connexes. Bull. Inst. Pasteur.
Minot, C. 8. 1908 Age, growth and death.

Mirropuanow 1905 Btude sur la structure, le développement et l’explosion
des trichocysts des Paramécies. Arch. f. Protistenk., Bd. 5.

LIFE HISTORY OF TWO RARE CILIATES 399

Miitter, O. F. 1786 Animalcula infusoria.

NERESHEIMER, E. 1908 Der Zeugungskreis des Ichthyophthirius. Ber. bayer.
Biol. Versuchungstation. Miinchen.

Nussspaum, M. 1884 Uber Spontane und Kiinstliche Theilung von Infusorien.
Vehr. d. natur. Ver. preus. Rheinland.

Prerty, J. A. 1852 Zur Kenntnis kleinster Lebensformen.

Pororr, M. 1907 Depression der Protozoenzelle und der Geschlechtszelle der
Metazoen. Arch. f. Prot., Bd. 1.

1908 Experimentelle cytologische Studien. Arch. f. Zellforschung,
Bd. v1:

1909 Experimentelle Zellstudien

II. Uber die Zellgrésse, ihre Fixierung und Vererbung. Arch. f. Zell-
forsch., Bd. 3.

ProwazeExk, 8. 1904 Untersuchungen iiber einige parasitische Flagellaten. Arb.
Kaiserl. Gesundheitsamte.

ScHaErrer, J. P. 1910 Selection of food in Stentor. Jour. Exp. Zool., vol. 8.

ScHAUDINN, F. 1896 Camptonema mutans. Sitz. Ber. k. preuss. Akad. Wiss.
Bd. 53.

1903 Untersuchungen iiber die Fortpflanzung einigen Rhizopoden.
Arb. kaiserl. Gesundheitsamte.

ScnewraKorr, W. 1889 Beitriige zur Kenntniss der holotrichen Ciliaten. Biblio-
theke zool., Heft. 5.

Stern, F. 1867 Der Organismus der Infusionsthiere, vol 2.
SrraspurGeR, E. 1893 Uber die Wirkungsphire der Kerne und die Zellgrésse.
Verworn, M. 1889 Protistenstudien.

1892 Die physiologische Bedeutung des Kernes. Arch. f. gesammt.
Physial:; Bd: 5.

Witson, E. B. 1906 The cell in development and inheritance.

Wooprurr, L. L. 1905 An experimental study on the life-history of hypo-
trichous Infusoria. Jour. Exp. Zool., vol. 2.

1909 Further studies on the life-cycle of Paramoecium. Biol. Bull.
Wola AZ.

1911 Two thousand generations of Paramoecium. Arch. f. Pro-
tistenk, vol. 2.

1911 Evidence of adaptations of Paramoecium to different environ-
ments. Biol. Bull., vol. 22.

Wooprurr, L. L., anp BairsELL, G. A. 1911 Rythms in reproductive activity
of Infusoria. Jour. Exp. Zool., vol.11.

DESCRIPTION OF PLATES

Unless otherwise stated, all figures are made from camera drawings or sketches
from the living cells.

The figures in plates 1 and 2 are of Spathidium spathula, those in plates 3 and
4 of Actinobolus radians.

PLATE 1

EXPLANATION OF FIGURES

1 Freehand sketch of Spathidium showing nucleus, contractile vacuole,
mouth, lines of insertion of the cilia and oral cilia.

2 Optical section of total preparation to show cuticle, cortex, endoplasm
and chromatin granules in the nucleus. X 400.

3 Sketch of total mount showing trichocysts in rim of the mouth. X 400.

4 and 5 Sketches of total preparations to show variations in form of the
nucleus. X 400.

6, 7 and 8 Freehand sketches of abnormal forms found a short time before
the death of the series.

9 Total preparation showing the nucleus divided into two distinct parts
each of which is tightly coiled. X 400.

10 Longitudinal section showing fragments of the macronucleus and an
elongated vesicular body containing deeply staining bodies in process of division.
* 415. :

11 A retort-shaped body found in section represented in fig. 14: X 1673.

12 Anterior region of body showing trichocysts, drawn from total prepara-
tions.

13. Sketch of mouth region drawn from living form.

14 Longitudinal section showing nucleus, food particles and retort-shaped
body shown in figure 11. X 415.

15 Sketch showing relation of the elongated and retort-shaped bodies found
in the serial sections pictured in figure 10 and 14. X 1673.

16 Total preparation illustrating another variation in the form of the nu-
cleus. 415.

17 Sketch of mouth region from living animal.

18 Section showing cortical foldings in the gullet. X 415.

19 and 20 Sketches of early division stage to show elongated chromatin
granules. X 415.

21 and 22 Sketches of a late division stage showing the fine chromatin gran-
ules. X 415.

400

1

PLATE

LIFE HISTORY OF TWO RARE CILIATES

JULIA E. MOODY

401

JOURNAL OF MORPHOLOGY, VOL. 22, NO. 3

PLATE 2
EXPLANATION OF FIGURES

23 to 27 Total preparations of early division stages showing greatly elon-
gated cell and nucleus. 400.

28 to 31 Late division stages illustrating great variation in the size and
form of nucleus in the new cells. 400.

32 Diagram to show the position of the incision made in experiments 1 and
2 in regeneration.

33 Diagram showing position of incision made in experiment 3.

34 Diagram to show place of cut made in experiment 4.

30, 36, 37, 38 Sketches made from the living form while emerging from the
cyst.

402

PLATE 2

LIFE HISTORY OF TWO RARE CILIATES

JULIA E. MOODY

JEM del

403

PLATE 3
EXPLANATION OF FIGURES

39 Sketch of living animal showing nucleus, contractile vacuole, tentacles
and cilia. Dorsal view.

40 Drawing of section through the mouth region to show the longitudinal
folds of the gullet; a partially digested Halteria in a flood vacuole and three
fragments of the nucleus. > 415.

41 Sketch from living animal. The periphery of the cell drawn under oil
immersion lens; to show peripheral vacuoles, cilia, tentacles, cuticle, cortex,
endoplasm and refractive bodies. > 787.

42 Sketch of section following the one represented in figure 40. x 415.

43 Sketch of living cell to show four greatly elongated tentacles.

44 Drawing of same individual swimming and dragging a long trail of de-
tached tentacles.

45 and 46 Degeneration stages of the same individual.

404

PLATE 3

LIFE HISTORY OF TWO RARE CILIATES

JULIA E. MOODY

PLATE 4

EXPLANATION OF FIGURES

47 Sketch of section to show two fragments of the nucleus, a retort-shaped
vesicular body containing masses of chromatin which appear to be in process

of division; and the peripheral vacuoles. 415.
48 The retort-shaped body shown in figure 47, greatly magnified. > 1673.
49 Sketch of total preparation showing an extended nucleus following the
circumference of the cell. X 415.
50. Total preparation to show the loosely coiled nucleus in the center of the
cell. 415.
51 Sketch of a section showing an almost straight nucleus.
Sketch of a total preparation in which the nucleus is seen tightly twisted.

52
x 415.
538 A portion of the nucleus pictured in figure 65, showing elongated chro-
matin granules. 787.
54 Early division stage drawn from a total preparation to show elongated
x 415.

nucleus and three vesicular bodies.
A series of division stages drawn from the living animal.

x 415.

NO, Oi, Bis as); (oil 4
60 Total preparation of a late division stage to show the elongated nuclei
and two vesicular bodies. X 415.

62, 63, 64, 65 Serial sections of a dividing cell, showing comma-shaped and
thread-like bodies scattered throughout the endoplasm. A section of the nucleus

x 415.

is shown in figures 64 and 65.

406

PLATE 4

LIFE HISTORY OF TWO RARE CILIATES

JULIA BE. MOODY

uy

55

cea a
as ESE se

JEM det

407

THE HEART AND ARTERIES OF POLYODON

C. H. DANFORTH
From the Department of Comparative Anatomy of Harvard Medical School

NINETEEN FIGURES

The following account of the arterial system of the ganoid
Polyodon was begun in the Anatomical Department of Washing-
ton University and completed in the Department of Compara-
tive Anatomy at Harvard Medica School. The writer grate-
fully acknowledges his indebtedness to the staffs in both places.
The work is approached entirely from a morphological viewpoint
and only the gross anatomical relationships are considered. The
material used has been chiefly fish of about a meter in length
which were secured in the vicinity of St. Louis. These were
injected with gelatin and starch masses, the latter proving quite
satisfactory for present purposes. ‘The results have been checked
by a study of serial sections of a 74 mm. specimen which has
already been briefly described by the writer (Danforth, ’11).

Allis’s (11) paper on the pseudobranchial and carotid arteries
of Polyodon did not appear until the present work had been
finished except for the drawings. Since that paper only par-
tially covers the field attempted here and moreover some com-
parison of our results seems desirable the paragraphs on the
pseudobranchial and carotid vessels are retained without very
much condensation.

PERICARDIUM AND HEART

The pericardium has the usual conical or rounded form with
the base directed caudally against the septum transversum.
A dorsal and two lateral faces are vaguely indicated. The lin-
ing is a uniform serous membrane with no macroscopic openings
except that of the pericardio-peritoneal canal. This is a rather

409

410 Cc. H. DANFORTH

large passage which opens from the pericardium into the coelom,
the posterior mouth being in the usual position ventral to the
oesophagus and dorsal to the liver. The pericardial mouth opens
dorsally between the entrance of the right and left ducts of
Cuvier. In a specimen where the greatest lateral diameter of the
pericardial cavity was 36 mm., the smallest diameter of the
pericardio-peritoneal canal was 5 mm. The cana itself was
about 15 mm. long.

The heart is attached to the pericardial wall anteriorly by
the truncus arteriosus and posteriorly by the union of the great
veins with the sinus venosus, by two cords consisting each of
a coronary vein and a posterior coronary artery, and by fine
strands running from the septum transversum to lymphoid masses
on the ventricle. The attachment of the sinus to the pericar-
dial wall is somewhat in the form of an upright H, of which
the great veins form the limbs. In the dorsal notch of the H
is the pericardial opening of the pericardio-peritoneal canal and
through the ventral notch pass the coronary vessels. These
are in two free strands each consisting of an artery and a vein.
Usually the left vein and right artery are large and the right
vein and left artery small. The veins were not studied in detail;
the arteries will be discussed further on.

The sinus venosus is very asymmetrical. On the left its
anterio-posterior diameter is short so that the ve ns almost enter
the auricle, but on the right, lateral to the mouth of the veins,
there is a saccular dilation equal in capacity to about one-third
of the auricle. Posteriorly there are a few internal trabeculae
or reinforcing strands, but these are not much developed. ‘The
arterial blood supply, is in part at least, by small twigs of cor-
onary orgin coming in from the septum transversum.

Externally the auricle appears nearly symmetrical but slightly
inclined to the left. It has a smooth surface and is somewhat
crenulated around the margin. In sagittal section it is in the
form of a right triangle, high behind and low and thin in front.
Within there is a strong development of trabeculae, mostly
flattened and forked, all around the edge except for a short
distance on the left, lateral to the auriculo-ventricular opening.

THE HEART AND ARTERIES OF POLYODON 411

Thus there are no trabeculae crossing the region opposite the
opening, but a crescent of them radiates around it, their con-
traction doubtless focusing the blood on this point. The weak-
est place in the wall, therefore, 1s opposite the mouth. The
sino-auricular valve is a pair of folds on the left placed oblique
to the sagittal plane with the ventral end of the slit between
them more medial than the dorsal. The slit makes an angle
of about 45 degrees with the perpendicular. The nearly round
auriculo-ventricular opening, is placed lateral, ventral, and an-
terior to the opening from the sinus. It is guarded by the auri-
culo-ventricular valve shown in fig. 14 B.t

The ventricular part of the heart is almost completely sur-
rounded and concealed by large lobes of lymphoid tissue appended
to its outer wall and richly supplied from the coronary artery.
Otherwise it presents no features calling for special mention.

The conus arteriosus of Polyodon is well developed. The
number of valves, however, is relatively few as compared with
some other ganoids. Figs. 1 and 2 each represent a conus that
has been cut along the mid-ventral line and opened to show the
cusps of the valves. In fig. 11 a portion of: the lateral wall has
been removed exposing the valves within. It will be seen ‘from
these figures that in this species the conus is a variable structure,
at least in regard to the number of valves. Perhaps the most
common form has three valves of four cusps each, the cusps
being arranged in longitudinal rows as is usual among elasmo-
branchs and ganoids. Between the first and second valve there
is a considerable space. In some individuals this space is occu-
pied by another valve which may be either rudimentary (fig.11),
or well developed (fig. 2). Although there are typically four
more or less equal cusps to each valve their number and form
vary and all stages from the merest rudiment to well developed
cusps can be found. In some cases, as in the second tier shown
in figure 1, multiplicity seems to result from a division of one of
the cusps. In agreement with most other forms the first (most
anterior) valve, supposed to correspond with the single valve

1The figures in this paper were made by Mr. Wm. T. Oliver from the original
drawings by the writer.

412 C. H.

DANFORTH

ABBREVIATIONS

a.bi., terminal bifurcation of afferent
branchial arteries

a.br.a. (1, 2, 3, 4), afferent branchial
artery

a.br.e. (1, 2, 3, 4), efferent branchial
artery

a.cc., common carotid

a.ce., external carotid

a.ci., internal carotid

a.co-ca., arteria coraco-cardiaca

a.coe., coeliac artery

a.com., branch of external carotid to
region of pseudobranch

a.co-me, coeliaco-mesenteric artery

a.cn., artery of heart

a.cor., coronary artery

a.dor., dorsal aorta

a. eps., efferent pseudobranchial artery

a. fa., facial artery

a. fg., artery to the F-shaped groove
of Bridge

a.fil.a., afferent filamentar artery

a.fil.e., efferent fiamentar artery

a.fil.s., basal branch of afferent fila-
meftar artery

a.fil.tr., transverse filamentar artery

a.fil.x., network of filamentar branches

a.hb. (2, 8, 4), hypobranchial artery

a.hb.y., a terminal branch of median
hypobranchial artery

a.he.a., anterior hepatic artery

a.he.p., posterior hepatic artery

a.hy., afferent hyoidean artery

a.hyo., hyodpercular artery

a.an., innominate artery

a.lhb., lhb.’, thb.’’, thb.’"", lhb.'’"" possi-
ple rudiments of lateral hypobran-
chial arteries

a.md., afferent mandibular artery

a.mhb., median hypobranchial from
second recurrent arteries

a.mhb.’, median hypobranchial from
fourth recurrent arteries

a.nu.; nutrient artery of gill

a.om., arteria ophthalmica magna

a.on., orbitonasal artery

a.op., ophthalmic branch of orbitonasal
artery

a.pa., parietal arteries

a.ph., pharyngeal branch of second
efferent artery

a.rc., recurrent branch of afferent
branchial artery

a.re., artery to kidney

a.rec., rectal artery

a.ret., arteria retinalis

a.sc., Subclavian artery

a.sp., subpericardial branch of cor-
onary artery

a.spl., splanchnic artery

a.sthy., artery to the m. sternohyoideus

a.thd., arteria thoracico-dorsalis (?)

a.thy., arteria thoracico-ventralis (?)

au., auricle

bd., bile duct

br., branchiostegal ray

c.br., branchial cartilage

c.cer. (1, 2, 3, 4), ceratobranchial carti-
lage

c.ch., ceratohya cartilage

c.co. (1,2,3), copulae

c.ep. (1,2,3,4), epibranchial cartilage

c.fi., filamentar cartilage

c.hyb. (1,2,3,4), hypobranchial cartilage

c.hyh., hypohyal cartilage

c.hyo., hyomandibular cartilage

c.th., interhyal cartilage

c.mc., Meckel’s cartilage

c.pbr. (pbr.’ pbr.”), pharyngobranchial
cartilages

c.sym., symplectic cartilage

ca., caeca

co., coelom

con., conus arteriosus

div., spiracular diverticulum

gb., gall bladder

gr., gill rakers

hb.a., anterior hemibranch

hb.p., posterior hemibranch

l., liver

lig., longitudinal ligament of dorsal
aorta

THE HEART AND ARTERIES OF POLYODON

lym., lymphatic vessel

m.adm., m. adductor mandibularis

m.adh., m. adductor hyomandibularis

m.bmd.m.bmd.,’ origin and _ insertion
of m. branchiomandibularis

m.lev. (1,2,3,4), m. levator arcuus
branchialis

m.obv. (1,2,3,4), m. obliquus ventralis

m.pr., m. protractor mandibularis

m.sthy., tendon of sternohyoideus mus-
cle

no., notochord

oc., oral cavity

oper., operculum

os.den., dentary bone

os. mz., maxillary bone

0s.pa., parasphenoid bone

0s.ps., post scapula

Fig. 1 A conus arteriosus cut along the midventral line and spread open.
this specimen there are but three valves, one of which has an extra cusp.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

pe., pericardial cavity
rec., rectum

rsp., respiratory folds
sb., swim bladder

si., small intestine
spi., spiracle

spl., spleen

st., stomach

sv., Sinus venosus

thy., thyreoid gland
v.de., duct of Cuvier
v.hp., hepatic vein
v.j.. jugular vein.
v.ji., inferior jugular
v.p., portal vein

c.pc., post cardinal vein
ven., ventricle

413

In

414 A. H. DANFORTH

of teleosts, is usually the one best developed. The cusps them-
selves are of the characteristic type described by Stohr (’76),
having a thick middle portion and thin lateral attachments
that are more or less fenestrated. There are also numerous
strands on the inner side of the flaps. Stdhr (’76) and Boas
(80) have published accounts of conus forms in a number of
the ganoids, but neither of them includes Polyodon in his de-
scriptions. These authors point out that a ‘beautiful transi-

Fig. 2. A conus with four well developed valves

tion’ from the ganoid to the teleostean type of conus is found
in the hearts of Amia on the side of the ganoids and of Butyrinus
among the teleosts. The former has three valves of four cusps
each. Of these cusps two are reduced in size in each valve.
Butyrinus has two valves, the anterior with two cusps, the pos-
terior with four as in Amia. Senior (’07) has recently described
the conus of Tarpon which contains two valves of but two cusps
each. With these exceptions almost all teleosts have a single

THE HEART AND ARTERIES OF POLYODON 415

valve of two cusps. If it were desirable to show a still closer
- series of gradations, the variable conus of Polyodon might be
placed before that of Amia in a series leading back through such
forms as Lepidosteus and Polypterus to the complicated struc-
ture found among the Dipnoi. Such a finely graded series,
however suggestive it might be, would probably not correspond
to any line of descent. The morphological value of the exact
number and arrangement of these valves cannot be very great,
especially where there is a tendency to multiplicity.

Fig. 3 Dorsal aspect of the ventral aorta

VENTRAL AORTA AND AFFERENT BRANCHIAL ARTERIES

The ventral aorta (fig. 3) is greatly elongated. In a fish of
10 decimeters it is fully one-tenth of the entire length of the
specimen. All of its branches arise near together at the anterior

416 Cc. H. DANFORTH

end. They are disposed in three pairs, morphologically com-
parable to similar vessels in elasmobranchs. The most anterior,
resulting from a terminal bifurcation of the aorta, is a short
trunk on either side, which presently divides to form afferent
hyoidean (ahy.) and first branchial (a.br.a. 1) arteries. Next
behind this comes a paired vessel which arises from the dorsal
aspect of the aorta and supplies arteries (a.br.a. 3, a.br.a. 4) to the
third and fourth gills. Finally the afferent artery (a.br.a. 2) to
the second gill, which is the only one to come directly from the
aorta, is in point of origin the most posterior of all. This is
due, of course, to the displacement headwards of the common
trunk of supply to the third and fourth gills.

The afferent hyoidean (fig. 4,a.hy.) and first branchial (a.br.a.1)
arteries immediately pass ventral to the tendon of the M. sterno-
hyoideus (m.sthy.), while all the other afferent arteries are dorsal
to it. The former, running close beneath the skin, follows the
hyoid arch for a considerable distance. Anteriorly it.lies just
medial to and parallel with the A. hyoidea (a.md.), the vessels
and the cartilage suggesting the arrangement of parts in a gill.
Further back it does not follow the arch so c’osely and runs
more nearly parallel with the median line. This vessel seems
clearly to correspond to the afferent hyoidean artery of other
ganoids and selachians. With teleosts, although sometimes
present in young (e.g., Ameiurus, Salmo), it appears typically
to be absent in the adult. In Polyodon I have frequently been
unable to find it, so its absence may here be a more or less com-
mon anomaly—an anomaly which might perhaps be expected
on physiological grounds inasmuch as this artery can distribute
only venous blood to the tissues.

The first afferent branchial artery, after having passed under
the tendon of the sternohyoideus, turns obliquely outward and
backward to enter its gill along the posterio-ventral border of
the m. obliquus ventralis I (m.obv.1). The second enters its
gill in exactly the same way except that it goes dorsal instead of
ventral to the tendon (fig. 4). .

The trunk (a.an.) on either side which supplies the third and
fourth gills runs back near the median line and dorsal to the

THE HEART AND ARTERIES OF POLYODON 417

ie Snomd:

c.mc
a. hb. y.
c. hyh.
a. md.
a. hy. thy:
a. Ihb.
a.br.a.1
a. mhb m. obv. 1
a. hb. 2
c. hyb. 2
a. sthy
a. br. a. 2
c. cer, 2
a. Ihb. m. obv. 2
c.co.1
a.an
m. bmd.
c, hyb. 3
a. Ihb.” :
© co. 2
a. bra. 3
c.cer. 3
m. :
a. Ihb”” Obv. 3
m. sthy.
a.hb. 4
c. cer, 4
a Inb/”
a. bra. 4
m. obv, 4
a. Ihb.” ©. cer. 5

Fig. 4 The principal structures of the hypobranchial region seen from below.
The anterior cartilage elements are displaced somewhat laterally and portions of

the copulae omitted.

418 Cc. H. DANFORTH

aorta. At the level of the third hypo-branchial cartilage it
divides into the two afferent arteries. One (a.br.a. 3) enters the
third gill in the same way as those described above, while the
other (a.br.a. 4) reaches the fourth in what at first sight appears
a very unusual manner. It ascends in an oblique groove on
the lateral face of the second copula (fig. 5) to gain the floor of
the mouth where it is separated from the oral cavity only by
the mucosa and the slightest amount of subjacent tissue (fig.
6, A). It crosses dorsally the anterior end of the ventral carti-
lage of the fourth branchial arch and then turns down in a groove
on the medial and posterior aspect of that cartilage to enter the
gill along the m. obliquus ventralis IV, thus coming into corre-

c.cer.5 c.cer4 c.co.3

c, co, 2

a. bra. 4 abna3 aan.

Fig. 5 The origin and proximal relations of the fourth afferent branchial
artery.

spondence with all the more anterior afferent arteries. Caudad
to the groove fortheartery there is a ligament binding the branch-
chial cartilage to the second copula, and this is interpreted by
Bridge (’79) and Van Wijhe (’82) as a second articulation. Such
an interpretation seems justifiable, and, if it be correct, brings
Polyodon into accord with Amia and Acipenser, in so far as
this point is concerned. In the third and fourth arches of Amia
the hypobranchial articulates with the copula near the floor
of the mouth and with its ventral keel at a deeper level (fig. 6, B).
The artery passes through the enclosure thus formed. With
Polyodon the same condition exists in the third arch, but in
the fourth, where the hypobranchial has disappeared or lost

THE HEART AND ARTERIES OF POLYODON 419

its individuality, the ventral articulation is displaced forward
and upward to the level of a dorsal articulation, and the original
dorsal (if it be such) has become posterior in position and is
nearly obliterated. In this readjustment the artery (a.br.a. 4)
has retained its original relative position and is in consequence
brought up to the floor of the mouth (fig. 6, A).

Fig. 6 A, section through a portion of the hypobranchial region of Polyodon
(74 mm. specimen); B, a similar section of Amia at the same level (Harvard
Embryological Collection, No. 273, sections 270-285).

THE BRANCHIAL CIRCULATION

On entering their respective gills the afferent vessels (a.br.a) run
a considerable distance without giving rise to any filamentar
arteries. At a point well towards the middle of the cerato-
branchial region each gives off a recurrent vessel (a.rc., fig.7)
which bears all the filamentar arteries medial to this point.
Dorsally all the afferent vessels except the fourth bifurcate near
the end of the gill, sending a division to the two hemibranchs as
they diverge around the insertion of the m. levator arcus bran-
chialis (a.b2.). Allis (loc. cit., p. 259) discusses these vessels
briefly, but does not mention their terminal bifurcation. He
finds the recurrent branches in larvae (130 mm. to 170 mm.),
but in view of drawings in his possession thinks they may be

420 Cc. H. DANFORTH

absent from the first arch in the adult. The present writer finds
them occuring invariably in all the arches.

The efferent arteries, with the partial exception of the fourth, to
be described presently, parallel very nearly the afferent vessels as
indicated in figure 7. They arise ventrally by two long branches,
one from each hemibranch, which unite at about the same level
as that at which the recurrent afferent artery is given off. This
corresponds very well with what Silvester (’04) found to be the

m. lev. 1 a. bi.

Fig. 7 Anterior aspect of the first gill after the removal of some of the super-
ficial tissues. :

condition in a large number of teleosts. Dorsally the efferent
vessels all send off small branches to accompany the terminal
bifurcations of the afferent arteries. Throughout the whole
gill the efferent fiiamentar arteries tend to unite in tree-like
groups (cf. fig. 9), a single stem often draining several filaments.
Morphologically the dorsal and ventral branches may be simply
enlarged stems of this sort that have developed with the increased
_ size of the gill. If such be the case they probably do not indicate
an incomplete fusion of a pair of efferent vessels such as occurs

THE HEART AND ARTERIES OF POLYODON Ait

in each holobranch of sharks, as Allis seems to suppose. As
shown in the figure, numerous nutrient arteries (a.nwu.) are given
off all along the gill. Largest and most important of these are
the hypobranchials discussed below, which arise from the ante-
rior ventral divisions of the efferent arteries.

a fil. e. a. fil. tr.

Fig. 8 Stereogram of a region near the middle of a gill. The basal parts of
two and one-half gill-filaments are shown on each side.

The arrangement of vessels near the middle of a gill is shown
semidiagramatically in figure 8. This arrangement is described
by Allen (07, p. 106). The efferent artery (a.br.e) lies deep-
est and next above it is the afferent artery (a.br.a). The third
vessel (lym.) still higher is the branchial vein or lymphatic of
Allen’s account. The afferent filamentar artery at its origin
immediately gives off a short spur (a,fil.s.) to supply the region
lateral to the main vessel and below its own point of origin.

422 C. H. DANFORTH

The main stem (a.fil.a.) ascends to the top of the filament, giv-
ing off throughout its whole length a series of short cross pieces
which reunite in an irregular network (a.fil.x.) from which the
‘afferent transverse filamentar arteries’ (a.fil.tr.) take origin.
Each of these supplies several of the flap-like folds of respiratory
epithelium (rsp.) on either side of the filamentar cartilage (c.fil.).
From the ultimate capillaries the blood is returned directly
to the efferent filamentar artery (a.fil.e.) which descends on the
other side of the cartilage to pass through a notch in its base.

It is stated above that the efferent branchial artery of the
fourth gill varies from the corresponding vessels anterior to it.
This divergence is correlated with other modifications in the
region. The fourth gill, unlike the others, is not a complete
holobranch. Its anterior hemibranch (fig. 9, hb.a.) is entire but
the posterior, (hb.p.)} although constantly present, extends only
slightly into the epibranchial region where a fusion has taken
place, obliterating the dorsal half of the cleft. The ventral
half of the cleft, however, remains open. This simple but un-
usual condition has proved misleading to several writers. Van
Wijhe (op. cit.) says that each of the four first gill-arches bears
on its outer side a whole gill. Jordan (’99) diagnosing the family
Polyodontidae, says ‘gills 43’. Finally Allen twice states (1.c.,
p. 106, p. 108) that.‘‘the fourth or last branchial arch has but
one row of filaments, and is therefore a hemibranch.” Strictly
speaking all of these descriptions are incorrect. Apparently
Van Wijhe failed to notice the absence of the dorsal half of the
fourth hemibranch, while Allen overlooked the presence of the
ventral half. Jordan’s statement is probably based on the car-
tilaginous arches and not on the gills themselves. A consider-
able number of specimens examined by the present writer proved
very uniform in this particular except that in a young one, 74
mm. long, the filaments were found to be rudimentary in the
epibranchial region of all the gills. In its lower half the fourth
gill is like those in front of it and its efferent artery gives off the
characteristic hypobranchial vessel in the usual manner.

Near the point where it rounds the articulation between the
epi- and cerato-branchial cartilages, the fourth efferent artery

THE HEART AND ARTERIES OF POLYODON 423

gives off a large posteriorly directed branch (figs. 9, 11 a.cor.),
which for the present may be termed the coronary artery. It
has a wider range of origin (cf. figs. 10 and 11) than is implied
by Allis’s account, but always leaves the gill in the epibranchial
region, and this, in connection with the fact that there is already
a full complement of recurrent arteries below, precluded the

hb. a.

3 Mle i

S =
eee

a. hb, 4 a. re. a. nu. * a. bra. 4

Fig. 9 Posterior view of dorsal part of fourth gill
Fig. 10 The arteries of the fourth gill

possibility of its belonging to the hypobranchial system. Passing
over the dorsal end of the fifth gill cleft it comes into a position
lateral to the oesophagus from which point its course will be
traced in a subsequent paragraph.

Near the origin of the ‘coronary artery’ and in the region
where the posterior hemibranch is lacking (fig. 9), the efferent ves-
sel separates into two parts. The main division (a.br.e. 4), which

424 Cc. H. DANFORTH

gives rise only to nutrient branches, passes over the posterior
face of the muscle connecting the cerato- and epi-branchial
cartilages and across the epibranchial directly to the pharynx.
The other part (figs. 9, 10), which gives rise to both filamentar
and nutrient branches, follows the dorsal edge of the cartilage
and supplies the filaments of the anterior hemibranch. It goes
anterior to the attachment of the m. levator arcus branchialis
and reunites with the vessel from which it arose. Allis infers
from Allen’s drawings that the loop thus formed is interrupted
medially in the adult, but such is not always the case, for in
all the specimens which were examined by the writer the loop
was found to be complete as shown in the figures.

Although the filaments of each of the first three posterior
hemibranchs are practically in a continuous line above and below’
with those of the next succeeding anterior hemibranchs, the eby
completely surrounding the clefts with filaments, I have been
unable by gross methods to find any connection at this point
between the arteries of the successive gills. Allis, however,
finds an indirect connection between the arteries of the third
and fourth gills.

THE HYPOBRANCHIAL ARTERIES

In all groups of fishes a system of hypobranchial arteries is
of constant occurrence, but the details of arrangement are rather
variable. These arteries are derived from anastomoses between
recurrent vessels which arise from the efferent arteries within
the gills. Frequently secondary longitudinal trunks are formed
which from their position T. J. Parker (’84, ’86) has designated
the lateral and median hypobranchials. ‘Their transverse con-
nections he calls commissural arteries. That author, it may
be observed, regarded the whole® system as derived from the
subclavian artery because in the sharks, with which he worked,
there is a connection between the two. The connection with
the gill arteries he thought to be secondary. This view, how-
ever, has received little subsequent support and is not now
generally maintained. G. H. Parker and F. K. Davis (’99)

THE HEART AND ARTERIES OF POLYODON 425

corrected and elaborated the original account and extended their
observations to one of the ganoids, Amia. Parker (’00), Sil-
-vester (’04) and others have studied these vessels in teleosts.
In Polyodon I find a considerable latitude of variation in minor
points even on opposite sides of the same fish, but no asymmetry
of constant occurrence, such as characterizes some of the tele-
osts, has been noticed. In figure 4 there is a slightly schematic
representation of the whole system in a specimen where the
parts attained a very full development.

The recurrent artery from the first gill (a.md.), generally
designated as A. hyoidea or afferent pseudobranchial artery,
is apparently constant in occurrence and position. It does not
anastomose with the other hypobranchial arteries and in this
region gives off only very small branches or none whatever.
From its origin (fig.7) it crosses the m. obliquus ventralis I diagon-
ally and loops around its tendon, passing first dorsal, then medial
and finally ventral to it as shown in figure 4. On rounding the
tendon it gains the posterio-medial aspect on the hypohyal
where, however, there is neither foramen nor groove. Allis (11)
finds branches here to the local musculature. It crosses the
lateral aspect of the ceratohyal (c.ch.) in a diagonal furrow (fig.
16) which traverses nearly the whole length of the proximal
cartilaginous portion of that element. Turning dorsally it
passes in front of the interhyal to the lower edge of the symplec-
tic where it again enters a shallow groove in the cartilage. Here
it frequently gives rise to a vessel (fig. 16) that runs back under
the end of the hyomandibular. Allis (l.c.,p. 260) is convinced
that this branch ts homologous with similar vessels in Amia
and Salmo and that it represents in all these the ‘‘remnant of a
commissure that in younger larvae undoubtedly connected the
hyoidean and mandibular aortic arches.” It is absent in many
of the adult Polyodons studied. Leaving the dorsal aspect of
the symplectic the afferent pseudobranchial artery runs forward
in a course ventral to the protractor hyomandibularis muscle,
but separated from the oral mucosa by a thick layer of fatty
tissue. Anteriorly it makes an S-shaped bend (fig. 15) and
enters the pseudobranch, where it usually breaks up into several

426 C. H. DANFORTH

divisions, from which the ultimate filamentar arteries (about
twenty) arise.

This vessel and the afferent hyoidean artery referred to above.
apparently correspond to the similar arteries in Amia and the
young trout. These vessels Allis (00) and Maurer (’88) believe
on embryological grounds to be the true afferent arteries of the
first and second arches, and this would seem to be the impli-
cation of Greil (’03) in reference to elasmobranchs. But Wright
(85) working with Lepidosteus, and Silvester (’04) with teleosts
are inclined to other interpretations. It can hardly be profit-
able to discuss this question in connection with Polyodon until
something of its embryology is known.

The hypobranchial arteries posterior to the hyoidean are more
variable and tend to become asymmetrical. The recurrent
vessel (fig. 4, a.hb.2) from the second gill, when fully developed,
on entering the hypobranchial region passes behind and dorsal
to a small accessory tendon from the m. sternohyoideus to the
second branchial cartilage. Here it gives off an anterior branch
(a. lhb.) which passes dorsal to all other arteries and tendons
to supply the lateral lobe of the thyreoid and the articular sur-
faces of the anterior basibranchials. At about the same place
it also gives rise to a posterior branch (a.lhb’.) which runs back
dorsal to the second afferent artery to reach the ventral aspect
of the third basibranchial cartilage. The recurrent vessel itself,
turning slightly backward and inward, presently gives off a third
branch (a.sthy.). This is to the sternohyoideus (m.sthy.), the
tendon of which it follows along the dorso-medial side. It
generally passes laterad of the m. branchiomandibularis (m.
bmd.), but may go mesad of it. In one specimen, unfortunately
injured, the vessel seemed to encircle the muscle. The vessels
of the two sides meet behind on the belly of the sternohyoideus
but remain more or less distinct. After giving off these three
branches the main trunk of the recurrent artery reaches the
median line ventral to the aorta, where it meets a corresponding
artery (if present) from the other side. Frequently the two
vessels (a.mhb.), usually of unequal size, run forward side by
side without uniting. ‘They supply the median lobe, and possi-

THE HEART AND ARTERIES OF POLYODON 427

bly more, of the thyreoid and terminate as the nutrient arteries
of the branchiomandibularis (m. bmd.) which they reach by pass-
ing between its two tendons of insertion to gain the ventral side
of the muscle. But while still dorsal to the branchiomandibu-
laris several minor branches are given off. One of these (a.hb.y.)
is very constant in its occurrence. It arises near the median
line and runs laterad beneath and close to the insertion of both
the sternohyoideus and transversalis I. It here comes into
contact with the hyoidean artery, twice crossing its dorsal sur-
face, but apparently never anastomoses with it. Its final dis-
tribution is over the hypobranchial cartilage and the surrounding
connective tissue.

In these vessels what may represent lateral, medial and com-
missural elements can, perhaps, all be recognized, but this is
a point which should not be pressed too strongly, for the vessels
are all variable and it is doubtful if they have much significance
for comparative purposes. The two lateral branches could to-
gether be interpreted as constituting a lateral hypobranchial
artery, and the fact that the anterior division sends a twig to
the thyreoid is in agreement with that vessel in the elasmo-
branchs as described by Ferguson (11). Whether or not the
posterior branch ever effects a communication with the arteries
from the third or fourth gills I cannot say. Apparently it usually
does not, although in some cases the terminal twigs come very
close to one another. The portion of the recurrent vessel be-
tween the ‘lateral hypobranchial’ and the middle line is in the
correct position for a commissural artery and the rest of the
vessel (a.mhb.) is the median hypobranchial. It is in accordance
with teleostean conditions, according to Silvester (04) and
Gudernatsch (711), for the median hypobranchial to supply the
thyreoid. The supply of the median part of the thyreoid, when
this vessel is lacking, has not been fully determined, but appar-
ently the gland then draws exclusively on the lateral hypo-
branchials. The median hypobranchial is not produced back-
ward beneath the aorta as it is in almost every fish thus far
described, but all the derivatives of the second afferent artery
are confined to the region anterior to the third gill.

428 Cc. H. DANFORTH

From the third gill there may or may not be a recurrent
vessel. When present it is comparable to the others in its gen-
eral relations but has a rather limited distribution.

The fourth efferent artery, so far as my experience goes, invari-
ably gives off a recurrent vessel which has median and lateral
branches. From the former are derived the anterior vasa of
the ventral aorta (fig. 8 a.mhb.’) and other small twigs. The
lateral branches, possibly to be considered collectively as pos-
terior remnants of the.lateral hypobranchial, are anterior and
posterior as regards their direction. The former (lhb.’’’) is small
and may anastomose anteriorly with a similar vessel from the
third gill when there is one present. The main trunk of the
posterior branch (lhb."’"”), which is larger and more diffuse in
its distribution, gains the median aspect of the fifth cerato-
branchial and runs back for a considerable distance. Its ter-
minal twigs reach those of the vessels from behind the heart
(fig.12) and may even anastomose with them, but any connection
here is probably only secondary. In distribution, but not in
origin, it suggests the dorsal median hypobranchial artery of
Silvester’s account.

The restricted distribution of the hypobranchial arteries of
Polyodon is probably to be correlated with the fact that the
hypobranchial region as a whole is greatly reduced and pushed
far forward so that the direct connection between this and the
pectoral region is unusually slender, consisting for a considerable
distance of little but cartilage and tendons. The great varia-
bility of the minor vessels on the other hand is a characteristic
shared by fishes in general. But that there should be no coron-
ary artery in connection with this system is interesting and
apparently unique. Allen (op. cit., pl. 3, fig. 5) shows a vessel
arising in the second gill and indicated as ‘coronary artery.’
This figure was made to show lymphatics, however, and is almost
certainly incorrect so far as the artery is concerned. The vessel
in question is either the commissural or the artery of supply to
the sternohyoideus; and Allen himself states in the text that
the coronary artery ‘“‘comes from the fourth right efferent bran-
chial artery and approaches the heart from the rear.” This

THE HEART AND ARTERIES OF POLYODON 429

is true, but the statement fails to make clear the important
fact that the coronary does not come from the hypobranchial
end of the efferent artery as in other fishes. Allis does not dis-
cuss the hypobranchial arteries.

THE CORONARY ARTERY

The coronary arteries as usually described for fishes are desig-
nated as anterior and posterior. The former arise from hypo-_
‘ branchial derivatives of the efferent branchial vessels and reach
the heart by passing along the conus arteriosus. The latter,
found in skates, arise from the coracoid branch of the subclavian
artery and reach the heart from behind. Depending on their
relation to the conus, anterior coronary arteries are recognized
as dorsal or ventral. In Polyodon anterior coronary arteries
are all entirely lacking and, further, the arteries of supply which
reach the heart by way of the septum transversum do not cor-
respond in other respects to posterior coronaries.

The vessel which for convenience may be called the coronary
artery (figs. 11 and 12 a. cor.), although it represents much more
than is usually covered by that term, arises as above described
from the fourth efferent branchial artery which it leaves in the
epibranchial region. It is somewhat variable in its point of
origin (figs. 9, 10 and 11, a. cor.) and is not symmetrical with
the corresponding vessel of the opposite side, one or the other
always being much the larger. Usually it is the right vessel
that is best developed. In the region of the oesophagus it gives
off a few small branches and then descends in a somewhat spiral
course to the base of the pericardium. Its relation in this region
to the anterior cardinal and ductus Cuvierii (fig. 11, v. dc.) is:
variable. It lies for the most part anterior and somewhat medial
to, the subclavian artery (fig. 12, a.sc.). Dorsal to the pericar-
dium there arises a long branch (a.co.-ca.) that runs forward
between the fifth ceratobranchial and median.hypopharyngeal
cartilages to supply the dorsal parts of the mm. pharyngoclavi-
culares (fig. 12). This vessel occupies the position of a part
of the a. coraco-cardiaca of Seyllium (Hyrtl ’58, Carazzi ’04),

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

430 C. H. DANFORTH

c. ep. 4 a.brna.4 v.j. : ve pc.

Vv. jl. a, co-ca. c. cer, 5con. pe. ven. au.

Fig. 11 Semi-schematic representation of the coronary artery and neighbor-
ing structures. In this specimen the left coronary artery is the one best devel-
oped.

a point which is emphasized by the fact that it sometimes anasto-
moses anteriorly with the fourth hypobranchial artery. Presum-
ably there is such an anastomosis in the larvae studied by Allis
(11, Le., p. 283). The main vessel enters the septum trans-
versum where branches are given off to the region below the
pericardium (a.sp.), to the liver (a.he.a.), and to the heart (a.cn.).

The artery which passes into the region ventral to the peri-
cardium (fig. 12, a.sp.) sends forward several anterior branches
which are chiefly distributed to the mm. pharyngo-claviculares
and the posterior part of the sternohyoideus. There is also
an anastomosing branch (fig. 12) connecting it with branches
of the subclavian artery. The anastomosis is in the position
of, and in a measure suggests, the coraco-hypobranchial anasto-
mosis of the elasmobranchs as described especially by Pitzorno

THE HEART AND ARTERIES OF POLYODON 431

Fig. 12 Dissection to show the coronary and subclavian arteries

(05), but although the relationships of the subclavians are the
same the other arteries are not entirely comparable. In several
specimens the anastomosing vessel was apparently lacking.
The branches from the coronary artery to the liver may arise
from both the right and left coronaries or from the larger one
alone (figs. 11, 12, a.he.a.). These vessels, which may be desig-
nated as anterior hepatic arteries, are distributed to the whole
anterior part of the liver. Their arrangement on entering the
liver is fairly constant, there being two arteries in relation to
each hepatic vein (fig. 13, a.he.a.). Within the liver they follow
the veins and ramify in the tissues till lost to macroscopic meth-
ods. In some, possibly all, cases they supply the gall bladder
(figs. 11, 13, gb.). In well injected specimens anastomoses are
easily traced between the anterior and posterior (a.he.p.) hepatic
arteries (fig. 13) but there is nothing to suggest that these are
more than secondary connections. The writer does not know

432 Cc. H. DANFORTH

Fig. 13 Outline of liver showing main branches of anterior and posterior
hepatic arteries.

of any other forms with vessels of this type and has no sugges-
tion as to their possible homologies. They are not mentioned
by either Allen or Allis in the papers cited.

The terminal branch of the ‘coronary artery’ (fig. 11, a.cn.)
or the part which actually supplies the heart and seems to corre-
spond to the posterior coronary of the skate, is generally, but not
always, best developed on the right side. The smaller artery
is, in all but exceptional cases, an insignificant vessel that runs
toward the heart along the large coronary vein. The larger
artery crosses the posterior part of the pericardial cavity in a
free strand which also contains the small coronary vein. It
passes under the auricle to the dorsal aspect of the ventricle
(fig. 11) where, at a point slightly behind and to the right of the
auriculoventricular junction, it breaks up into three or more —
branches. Like other arteries approaching their final distri-

THE HEART AND ARTERIES OF POLYODON 433

Fig. 14 Ventral (A) and dorsal (B) views of the ventricle showing distribu-
tion of the coronary artery. Branches with cut ends supply the investing lym-
phoid tissue.

bution it is extremely variable. Figure 14 is from sketches of the
dorsal and ventral aspects of a typical ventricle. To bring the
arteries into view the surrounding lymphoid masses were removed.
Vessels represented with cut ends went to supply these structures.

The question as to the homology of the ‘coronary artery’
in Polyodon is not easily disposed of. The morphological impor-
tance of it depends on the value that may be attached to such
structures as units. ‘The vessels of fishes are variable to a marked
degree and it does not seem difficult for new channels to be
established nor is it surprising to find related forms more or less
divergent in the matter of the distribution of minor arteries.
Carazzi (op cit.) states that in working with a single form (Scyli-
lium) he finds in various individuals the arrangements described
for different species and even different genera. He seems to
be inclined to emphasize the functional rather than the mor-
phological significance of the different plans. How far the
divergent arteries of Polyodon are new developments or to what
extent they are direct modifications of a more primitive condi-
tion cannot be judged with any assurance till their embryology
is known.

434 Cc. H. DANFORTH
ARTERIES OF THE PHARYNX

Each efferent branchial artery on approaching the roof of
the pharynx leaves its groove in the epibranchial cartilage and
crosses the anterior aspect of the m. levator arcus branchialis
(m. lev. 1, fig. 7). Medially it turns ventrally and somewhat
posteriorly around the muscle a little proximad of its insertion.
The relations with the cartilages vary somewhat in the different
arches (fig. 15). The first passes through a triangular space
bounded dorso-medially by the parasphenoid bone, antero-lat-
erally by the first pharyngobranchial, and postero-laterally by
the anterior end of the first epibranchial and second pharyngo-
branchial cartilages. Within this triangular space it gives rise
to the common carotid artery (a.cc.). Leaving the triangle it
crosses the medial part of the second pharyngobranchial, making
a groove in its ventral surface to gain the deep median exca-
vation between the wings of the parasphenoid behind. It unites
with the corresponding vessel of the other side in a plane passing
near the middle of the third pharyngobranchials. The third
pharyngobranchial does not reach the parasphenoid above and
so a triangle corresponding to the one in front is not completely
closed in behind. This allows the second efferent artery to
slip back to a more posterior position as shown in the figure.
It reaches the median vessel in the plane of the end of the third
epibranchial cartilage. The third and fourth vessels likewise
have migrated somewhat backwards. All of these arteries are
forced to make a ventral dip around the knife-like edge of the
parasphenoid. On reaching the mesial side of the wing of that
bone they turn abruptly upward and backward to enter the
median vessel on its ventral side. The arrangement of these
vessels, each separately emptying directly into a single median
dorsal aorta, is in accord with some of the other ganoids and
simpler sharks and is seemingly more primitive than the radices
aortae of higher sharks and teleosts.

Pharyngeal branches of the main trunks are not well devel-
oped. There are, however numerous small twigs (fig. 15) that
supply the roof of the mouth and one prominent artery (a.ph.)

THE HEART AND ARTERIES OF POLYODON 435

from the second efferent vessel which runs up over the wing
of the parasphenoid to the cartilage of the occipital region, where
it is distributed around the origin of the levator muscles and may
possibly send twigs through into the cranial cavity or anasto-
mose with others coming out. Allis describes this vessel as
coming from the first efferent artery; presumably it may arise
from either the first or second. Allis also describes a commissure
between the third and fourth efferent arteries. No large vessels
were found going back to the oesophagus.

THE COMMON CAROTID ARTERY

The common carotid (fig. 15, a.cc.) arises from the first efferent
branchial artery just in front of the anterior epibranchial carti-
lage. It is a short vessel which runs diagonally forward and
inward along the mesial aspect of the first pharyngobranchial,
near the end of which it may be said to terminate by dividing
into internal and external carotids. In the adult this division
is a very unequal one, the internal carotid, which arises from
the posterior and ventral side of the vessel, being so small that
it was at first entirely overlooked or mistaken for one of the
pharyngeal twigs that are given off in this region. The minute
internal carotid will be described further on in connection with
the artery from the pseudobranch; the distal continuation of
the main trunk from the end of the common carotid is here
described as the external carotid, even though the hypoépercu-
lar and orbito-nasal arteries are also involved.

THE EXTERNAL CAROTID ARTERY

The external carotid (a.ce. figs. 15, 16), appearing as a pro-
longation of the common carotid, continues around the anterior
end of the first pharyngobranchial cartilage in which it makes
a shallow groove. Laterally it passes through a rather large
foramen into a longitudinal channel, the facial canal of the basis
cranii. Here it gives off a large branch (a.hyo.) which follows
the hyomandibular nerve out posteriorly through the facial
foramen. The canal is produced forward beyond the point

C. H. DANFORTH

436

Ay Ge

a. fa.

4 ret.

a. hyo.

c, pbr.”

Yy
AL V4 o Uy
Z, 7
LULL

c. ep. 2
a. md.

SSS SSS SSS sss sss SSS SSS SSS SESS

~
m7

= 2
SS
~

oN

i
5)
Uy

yy
y

Zo

Fig. 15 Dissection of the roof of mouth and pharynx

SSSSSSSSSSSSSSSSSSSSSSESS SSS SEES

a. ph
a. dor.
a
a. co-me.

THE HEART AND ARTERIES OF POLYODON 437

where the hyomandibular nerve enters it and through this ante-
rior part of the channel the artery again escapes from the cranial
wall, emerging dorsal to the m. protractor hyomandibularis be-
tween its two slips of origin. As it leaves the canal it gives off
a second branch. This, together with the first, probably repre-
sents the equivalent of the hypodpercular artery of Silvester’s
descriptions (for teleosts) or the hypodpercular and external
carotid of Amia (Allis).

The posterior branch is here called the hyodpercular (a.hyo.)
and is likewise designated by Allis in his recent paper. Leaving
the facial foramen along with the nerve, it ascends abruptly
the posterior aspect of the hyomandibular bone, gives off a
large muscular branch to the adductor hyomandibularis (fig.,
16 m.adh.) and assumes a superficial position along the insertion
of this muscle. Several small branches cross over to the ante-
rior side of the hyomandibular bone and a very long one runs
back under the operculum (oper.) to the inner side of the oper-
cular flap. The main vessel, still following the hyomandibular
nerve, traverses a groove in the distal end of the cartilage from
which it sends a second branch to the opercular flap and then
passes down under the branchiostegal ray (br.) to be distributed
posterior and medial to the area supplied by the facial artery.
For a further discussion of this artery the reader is referred to
Allis’s paper (1. ¢., p. 286).

The second branch of. the external carotid immediately sepa-
rates into two divisions. In the specimen described by Allis (L.e.,
p. 285) they arose separately from the main trunk. One runs
laterally beneath the ep thelium in front of the spiracular cleft
and along the anterior face of the hyomandibular bone. The
other divides into external and internal branches. The former
(a.com.) ramifies over the cartilage around the spiracular cleft
and diverticulum and supplies a nutrient branch to the anterior
division of the protractor muscle. The latter (a.fg.) passes
through the cartilage into a large ‘fat-space’ (the F-shaped groove
of Bridge) beneath the frontal bone and medial to the hyoman-
dibular articulation where it breaks up into small twigs.

438 Cc. H. DANFORTH

Neither of the above arteries seems to fill completely the
place of the external carotid of Amia although the second comes
very near to it and, with the first, also covers the hyodpercular
of that form. The teleostean vessel which Silvester, endeavor-
ing to follow Allis, called the hyodpercular artery, but which,
as Allis subsequently (08 c) pointed out, is presumably com-
parable to the external carotid of Amia, finds a more or less
complete homologue in the two vessels above described. The
fact that they arise separately from a longitudinal trunk and
not by a single stem from one of the several points where the
‘hyodpercular’ (‘external carotid’) may appear in other forms,
is probably of little morphological import. The branch (a. com.)
from the anterior of these, which is described above as crossing
beneath the epithelium of the spiracular cleft, was not found
to connect directly with the spiracular vessel although it is
strongly suggestive of the commissure that occurs here in Amia,
the Loricati, and the teleosts described by Silvester.

The further continuation of the external carotid, apart from
its proximal connection, is very much like the artery in teleosts
which Silvester calls the external carotid and Allis (’08 c) the
orbito-nasal. It is partly encircled by the a. opthalmica magna
(fig. 15, a.om.) just as is the orbitonasal (Allis) in teleosts and,
like that artery, it forms a large longitudinal trunk supplying
branches to the orbit, eye muscles and nasal region. After
leaving the facial canal and giving off its second bran: h as above
described, it runs forward medial to the anterior portion of the
protractor hyomandibularis already referred to. Along this
part of the course there arise several small twigs probably of
no morphological importance. As it approaches the mandibu-
lar nerve it turns somewhat laterad under the protractor man-
dibularis muscle where it gives off the large facial artery and
then continues forward beneath the orbit and recti muscles.
Allis (1. ¢., p. 285) deseribes the external carotid as terminating
here by dividing into the ‘orbito-nasal and maxillo-mandibularis
arteries.’

The facial artery (figs. 15, 16, a.fa.) passes over the nerve to
its anterior side and then follows it out, lying at first on the

THE HEART AND ARTERIES OF POLYODON 439

ventral side of the protractor and then between that muscle
above and the m. adductor mandibulae (m.adm.) below. In
this position it gives off a large branch which immediately sepa-
rates into posterior and anterior divisions. The former is a
nutrient artery to the m. adductor mandibulae, but the latter,
although supplying some muscular twigs, is distributed mostly
to the abundant fatty tissues and follows the palatoquadrate
forward to the median line. This recalls the vessel in Polypterus
which arises from the internal carotid and has recently been
considered by Allis (08 b) as a possible remnant of the efferent
mandibular artery of that form. Here it is not likely that it
has any such significance. Further back the facial artery passes
over the lateral aspect of the m. adductor mandibulae and in
between the palatoquadrate cartilage and maxillary bone. Its
rather numerous branches, some of them muscular, are indicated
in figure 16. Below the angle of the mouth it still adheres to
the lateral surface of the muscle as the latter gains its insertion
on Meckel’s cartilage. The main vessel finally becomes super-
ficial ventrally by emerging from between the cartilage and
overlying bone. From this point it may be traced forward
along the cartilage nearly to the median line.

The next prominent branch (a.op.) from the external carotid
enters the region back of the orbit, and is apparently the one
designated by Allis as the ophthalmic branch. It gives a branch
to the rectus muscles of the eye, just as does a similar vessel
from the external carotid (orbito-nasal) of Lopholatilus and other
teleosts (Silvester), and divides into two main branches. One
of these supplies the fatty tissue behind and above the eye while
the other goes to the similar tissue medial and anterior to the
eye. Some of the terminal branches of the latter supply the
m. obliquus superior and others reach the nasal region, being
distributed apparently to a part of the sensory epithelium.

Two other branches (fig. 16) arise from the external carotid
before it enters the rostrum. One of these runs ventral to the
orbit, supplying the m. obliquus inferior and the superficial tis-
sues below and in front of the orbit. The other has a somewhat
similar but more anterior distribution except that its largest

440 Cc. H. DANFORTH

m. adh. oper.

Fig. 16 Sketch of the external carotid and its branches. The deeper parts
are indicated by lighter lines.

branch crosses the posterior face of the antorbital cartilage and
then goes through a foramen into the nasal capsule. How far
this perforating branch is homologous with the anterior end
of the orbitonasal artery which perforates the antorbital process
of teleosts it is difficult to say.

The artery of the rostrum (fig. 17), the terminal extremity
of the external carotid, enters the snout by passing across a
broad groove beneath the antorbital process. It then bends
dorsally and, lying under the overhanging dorsal expansion of
the central cartilaginous core, retains its individuality nearly
to the anterior end. In its course through the rostrum it gives
off a series of long medial vessels, which lie close to the carti-
lage except near their terminations where they may turn laterad,
and another series of short lateral vessels, some of which actually
arise on the medial side and then pass around the main stem
dorsally. Through the greater part of its length the rostral
cartilage is hollow and filled with fat behind and a kind of mucous
tissue in front. The nutrient artery to this tissue arises from
the main vessel shortly after the latter enters the rostrum. At
its origin it bends somewhat caudad and then goes through a

THE HEART AND ARTERIES OF POLYODON 441

Fig. 17 Arterial supply of the rostrum. The cartilage is cross hatched.

foramen in the cartilage. The two nutrient arteries, one from
either side, run obliquely forward and inward giving off anterior
and posterior branches. The latter is deep and reaches the
dorsal wall of the cartilage where possibly it sends through per-
forating branches. There are other perforating branches further
forward. The main stems of these nutrient arteries finally
fuse and the resulting median vessel is continued anteriorly
through the mucous tissue. It ends finally in one or more
perforating vessels. .

AAD. Cc. H. DANFORTH

Although the foregoing account shows quite clearly that the
extensive system of vessels here described as the external caro-
tid (fig. 16) and its branches is really much more than is usually
embraced by that term, this designation is employed for conven-
ience in the description of a natural system, the parts of which
cannot be satisfactorily homologized by macroscopic methods
alone. The several interesting questions that naturally arise
in this connection must undoubtedly await embryological studies
for their final solution.

INTERNAL CAROTID AND EFFERENT PSEUDOBRANCHIAL
ARTERIES

The internal carotid artery (fig. 15, a.ci.) is intimately con-
nected with the efferent pseudobranchial (a. eps.). It at first
runs anteriorly in the roof of the mouth, lying on the ventral
surface of the parasphenoid bone. Anteriorly it turns laterally
and, dorsal to an expansion of the bone, enters a canal in the
basis cranii, where it is soon joined by the efferent pseudobranch-
chial (a. eps.) which is the larger of the two. This vessel arises
in the pseudobranch from about twenty filamentar arteries and,
gaining the ventral surface of the protractor hyomandibularis,
passes forward across that muscle to a canal which it enters
at a point dorsal and anterior to the opening of the internal
carotid canal. The canals unite in front and the arteries within
anastomose. This anastomosis may apparently be effected
through a large connecting branch or more commonly the two
arteries completely fuse for a short distance. Anterior to the anas-
tomosis the internal carotid becomes the larger vessel and the
efferent pseudobranchial, now greatly reduced, proceeds as the
a. ophthalmica magna (fig. 15, a.om.). This artery, while still
in the cartilage, gives rise to several small twigs which could not
be traced far, and then escapes from the cranial wall above the
trigeminal nerve. It is at first posterior to the rectus muscles
and then comes to lie in between them, following an independent
course to the back of the eyeball which it enters just posterior
to the entrance of the optic nerve. A few small twigs may be

THE HEART AND ARTERIES OF POLYODON 443

given off on the outside of the sclerotic, some of them possibly
anastomosing with the a. retinalis.

Returning to the internal carotid proper, we find it proceed-
ing from its anastomosis with the efferent pseudobranchial
artery through the basal cartilage into the cranial cavity. This
region is carefully described by Allis (l.c., pp. 289-290). In
this part of its course it gives off a long slender vessel, apparently
confined entirely to the cartilage, which arches forward, ultimately
reaching the median line. Just inside the cranium the a. retin-
alis (a.ret.) separates and accompanies the optic nerve to the
eye. This is a very minute vessel which it is difficult to trace.
Lateral to the diencephalon the internal carotid divides pal-
mately into three branches, of which the posterior is largest,
the anterior smallest. Allis apparently found only two of these,
but he did not attempt to trace the encephalic arteries. The
anterior division, which is sometimes greatly reduced, runs
along the ventro-lateral side of the corpus striatum, dividing
in no very constant manner into the twigs that supply this
region. Its terminal rami run in among the coarse bundles
of the olfactory nerve and one of them gains the dorsal side of
the telencephalon. If the anterior division be reduced in size,
as is often the case, its place is taken by a branch from one of
the other two, either of which may send out a branch to cover
practically the same region. The middle division of the en-
cephalic artery, which, as just stated, often shows an anterior
branch, ascends in the angle between the telencephalon and
diencephalon. As it approaches the epiphyseal region it turns
abruptly backward over the dome-shaped midbrain where it
is joined by its fellow of the opposite side. “ There may be merely
cross anastomoses between the two or they may fuse completely,
in which case a single median vessel supplies the posterior aspect
of the midbrain and gives off on either side_a large branch which
turns laterally and ventrally between the optic lobes and cere-
bellum. These arteries are subject to considerable variations
and these variations are correlated for the most part with inverse
variations in the lateral rami of the third encephalic artery
with which they anastomose. The posterior and largest of the

444 Cc. H. DANFORTH

encephalic arteries like the second may or may not give rise
to a vessel to the corpus striatum. In one specimen, where
such a branch was well developed, no anastomosis could be
detected between it and the anterior artery by the gross methods
employed. The main trunk turns medially between the in-
ferior lobe and the anterior end of the medulla. As it arches
backward it gives off one or two lateral branches which run
up between the optic lobe and the cerebellum to anastomose
as above described with the dorsal artery ‘of the brain. The
posterior arteries of the two sides join each other in the angle
between the inferior lobes and the medulla, but from this point
of union two vessels instead of one run caudally as far back as
the level of the vagus nerve and may even extend further back
before finally uniting. From these two parallel basilar arteries
in front and their single posterior prolongation behind, a number
of small branches are given off to the medulla, to the roots of
the nerves, (nearly all of which are accompanied laterally by
small arteries) and to the floor of the cranial cavity. Several
small vessels of the latter group go to the optic capsule. In
addition to these smaller branches there are on either side two
large branches which seem to be constant in occurrence, but
not in position. The more anterior of these runs laterally, either
in front of the auditory nerve and between the roots of the V-
VII complex, or posterior to all of these, to gain the lateral side
of the medulla along which it runs caudally. The other and
more posterior gives off a small branch and then follows the
ninth nerve laterally as far as the auditory region where it turns
aside to be distributed about the sacculus. No end to end
connection between the basilar artery and the vertebral branch
of the subclavian, which enters the spinal canal further caudad,
could be detected, although it is not improbable that small anas-
tomoses between the two arteries do occur. It is evident, at
any rate, that the type of circulation, common to many of the
vertebrates, including some of the fishes, in which the vertebral
arteries are an important factor in the cerebral circulation, has
not been attained by Polyodon.

THE HEART AND ARTERIES OF POLYODON _ 445

It will be seen from the foregoing account that the relations
of the internal carotid are rather more in accord with conditions
in such other ganoids as Amia and Acipenser than is the case
with the external carotid. The connection of the internal caro-
tid with the efferent pseudobranchial and ophthalmica magna
arteries is clearly explained in Allis’s paper, and calls for no
farther discussion here. It remains only to be said that in my
larva, which is younger than those studied by Allis, the relative
size of the internal carotid and the efferent pseudobranchial
artery running beside it is very different from what it is in the
adult. In a 74 mm. specimen the internal carotid, instead of
being much the smaller vessel, is five or six times larger than
the other. This would seem to indicate that in Polyodon the
presence of an anterior carotid (i.e., one deriving its supply
chiefly from the pseudobranch) is secondary and not the primary
condition as it is said to be in sharks.

THE DORSAL AORTA

The origin of the dorsal aorta from the efferent branchial
arteries has already been described. A few millimeters behind
the entrance of the last pair of arteries the aortic wall becomes
chondrified and the vessel from this point on is invested by a
thick wall of cartilage which is interrupted only in the mid-
dorsal line. It is decidedly flattened dorso-ventrally, so that
in cross section it is in the form of a small sector of a large circle
or even crescentic in outline. In surface view, when dissected
out, the dorsal aspect shows regular metameric markings in
the cartilage, but the ventral aspect is ridged and furrowed in
a very irregular manner. <A most striking feature in connection
with the vessel is a longitudinal ligament (fig. 18, lig.), suspended
in the lumen throughout its entire length. This ligament is
attached in front and behind and supported throughout by a
continuous thin membrane of fibrous tissue which binds it to
the notochordal sheath along the non-chondrified dorsal line of
the aortic wall. The flat suspensory portion is so deep that the

JOURNAL OF MORPHOLOGY. VOL. 23, NO. 3

446 Cc. H. DANFORTH

rounded part below rests on the ventral inside wall of the aorta,
thereby dividing it into essentially distinct longitudinal com-
partments. In specimens fixed in a curved position the igament
stretched diagonally from side to side in accordance with the
bends in the body. Perhaps, in addition to the function sus-
gested by this condition, it may also act in opposition to the
interspinous ligament. This peculiar structure was also observed
by Bridge (’79), who suggested tentatively that it might repre-
sent the subchordal rod of elasmobranchs. I find a similar
structure occuring in Scaphirhynchus and a rudiment of it in
the trout and Ameiurus.

The branches of the dorsal aorta are an irregular series of
segmental arteries, beginning with the subclavian in front and
given off throughout the whole course; a series of short ventral
branches to the air-bladder, gonad, and kidney; and the large
coeliaco-mesenteric artery which arises at the anterior extremity
of the coelom. The aorta itself terminates as it approaches
the end of the tail by dividing into short lateral branches which
pass out of the axial cartilage and again divide, this time into
dorsal and ventral branches for the two lobes of the tail.

THE SEGMENTAL ARTERIES

There are two main divisions of the segmental arteries, the
dorsal, or parietal, the ventral, which in the coelomic region
are the splanchnic arteries. They may arise along the lateral
triangle of the aorta by short common stems or, more frequently
the dorsal and ventral divisions are quite distinct from each
other. These vessels are somewhat irregular in their appear-
ance and vary greatly in size. For the most part, however,
there is one pair of good sized arteries for every two or three pair
of nerves. Sometimes the two members of a pair will be very
different in size and sometimes two successive arteries on the
same side will both be large, especially in the region of .a fin.
Each dorsal parietal artery (fig. 18, a. pa.), on emerging from
the aortic wall, sends a large horizontal branch diagonally out-

THE HEART AND ARTERIES OF POLYODON 447

ward and backward in an intermuscular septum. This branch
ramifies somewhat in the septum but maintains its individuality
until it reaches the lateral line where it divides into dorsal and
ventral branches and shows some inclination to anastomose
with similar vessels in front and behind it. The main stem next
gives off a spray of small vessels to the notochordal sheath and
higher up a branch that enters the spinal canal. Beside the
neural spine there arises a second lateral branch very much

Fig. 18 Diagramatic cross section in the region of the kidney

like the first. Dorsally the parietal artery itself ends in a ter-
minal bifurcation, or, in the region of the dorsal fin, becomes one
of the arteries of supply to that structure. In a fish about a
meter long there were five pairs of dorsal arteries to the fin.
The splanchnic arteries (fig. 18, a. spl.) may give off small
ventral branches to the air-bladder, kidney or gonad, lateral
to which they continue ventrally close to the peritoneum. They
show a slight tendency to anastomose and form longitudinal
connecting vessels below. Several of them supply each ventral

448 C. H. DANFORTH

fin and in some cases at least, terminal branches around the
anus run forward on the rectum to become continuous or to
anastomose with the mesenteric artery. Posterior to the coelom
the vessels comparable to these partake more of the character
of the parietal arteries. About eight or nine pairs of them supply
the anal fin.

The subclavian arteries (figs. 12, 15, a. sc.) have the general
relations of the other segmental arteries. They arise at a con-
siderable distance behind and entirely independent of the fourth
efferent branchial arteries. Each at once gives off the charac-
teristic dorsal branch which ramifies in the ‘occipital region and
supplies the neural canal but was not found to. anastomose
directly with the basilar artery, The lateral division runs ventrad
and laterad behind the jugular vein and in front of the post scapu-
lar. Aside from the anterior branches already described in con-
nection with the coronary and hypobranchial arteries, it gives
off posteriorly above the scapular a long superficial branch (a. thd.)
which is probably to be identified with the a. thoracico-dorsalis
of the selachians (Pitzorno, loc. cit.), and a ventral branch
(a. thv.) which is probably the a. thoracico-ventralis of those -
forms. Its principal remaining branches supply the muscula-
ture of the fin.

ARTERIES TO THE KIDNEY

The arteries that directly supply the dorsal side of the kidney
(fig. 18, a.re.) and swim-bladder are numerous small vessels
that arise irregularly along the ventral side of the aorta. Within
the kidney there is also sometimes developed a longitudinal
stem of greater size. The finer anatomy of these vessels in
relation to the renal tissues was not studied.

THE COELIACO-MESENTERIC ARTERY

The coeliaco-mesenteric artery (figs. 15, 19, a. co me.) arises
from the dorsal aorta between the levels of the anterior end of
the coelom and the posterior margin of the pronephros. Enclosed

THE HEART AND ARTERIES OF POLYODON 449

in a free strand of tissue, it descends along the anterior face of
the air-bladder nearly to the cardiac end of the oesophagus, a
distance of about a decimeter in a fish a meter long. Throughout
the greater part of this portion of its course it is closely envel-
oped between the air-bladder and the dorsal aspect of the right
lobe of the liver, making a deep furrow in the later. On the
right side near the junction of the oesophagus and stomach
it gives off its first large branch, the coeliac division (igsw 13 19
a. coe.), and comes into relation with the portal vein. Asso-
ciated with the vein in a free strand, it passes under the oeso-
phagus to the right side of the intestine to which it gives a few
branches and then enters the spleen which it supplies copiously.
Emerging behind, dorsally and somewhat to the right it gives
off a few small branches which go to the rectum, and a large
one which enters the mesentery and bifurcates, one division
going to the fundus of the stomach and the other to the pos-
terior ventral aspect of the air-bladder where it anastomoses
with the dorsal arteries to that organ. This casual anastomosis
may become important for, in one specimen, the main blood
supply to the whole posterior part of the alimentary canal came
down through the channel thus opened up, making a kind of
anomalous posterior mesenteric artery. Behind the spleen the
main vessel (a.rec.) is continued in the median line down the
dorsal side of the rectum to which it gives off several branches,
and, at least in some cases, becomes directly continuous and
also anastomoses indirectly with some of the posterior ventral
segmental arteries.

The coeliac division of the artery gives rise first to oesophageal
branches, some of the anterior of which may send twigs to the
swim-bladder above and the liver below. At about the same
level there also arises a slender artery (a.he.p.) to the liver which
- often reaches as far as the gall-bladder. This is the principal
right posterior hepatic artery. The main vessel, subject to
considerable variation, gives rise to branches to all the neigh-
boring organs and then turns dorsally in the angle on the right
side of the gastro-intestinal junction. As it approaches the

Cuno. DANE ORE:

450

‘991

ds

Fig. 19 Schema of the coeliaco-mesenteric system

THE HEART AND ARTERIES OF POLYODON 451

mouth of the pyloric gland it bifureates. The anterior division
branches palmately into a number of arteries to the stomach,
a left posterior hepatic artery, which, as previously stated, anas-
tomoses with the anterior hepatics, and a large artery to the
anterior part of the pyloric gland. The posterior division breaks
up into three branches, one of which supplies the posterior part
of the pyloric caeca, one to the small intestine and one to the
side of the stomach.

As is the case with many other forms the details of distri-
bution vary greatly with all of these arteries, and a wide latitude
must be allowed for individual differences.

CONCLUSION

The foregoing account of the general anatomy of the arterial
system of Polyodon shows that, while in many respects the
ganoid type, especially as exemplified by Acipenser, is quite
clearly indicated, there are, nevertheless, some features which
suggest elasmobranch and teleostean conditions. Among the
former may be mentioned the posterior coronary arteries which
to some extent resemble those of the skates, and among the
latter the orbitonasal artery of teleosts, although it must be
confessed that the homologies of this vessel are by no means
clearly established. Marked peculiarities of Polyodon, appar-
ently not common to other fish so far as known, are the absence
of anterior coronary arteries, the origin of the posterior coronary
in the epibranchial region, and the existence of anterior hepatic
arteries. The evidence of these characters, so far as it goes,
indicates that Polyodon is not very closely related to any of
the other forms. Whether its resemblance to the skates, both
here and in some of its skeletal structures, is anything more
than a chance parallelism may well be doubted.

There is a close agreement between the account given here
and that of Allis’s paper wherever the two overlap. Small
differences in regard to number and origin of vessels fall easily
within the limits of individual variation which, in Polyodon, is
very great.

452 Cc. H. DANFORTH

The many differences, however, which Allis records as exist-
ing between his larvae and drawings which he has of adults
ean hardly be due, as he supposes, to immaturity of the former
since all of the possible larval characters which he mentions
are common to fish a meter or more in length.

THE HEART AND ARTERIES OF POLYODON 453

LITERATURE

ALLEN, Wm. F. 1907 Distribution of the subcutaneous vessels in the head
region of the ganoids, Polyodon and Lepisosteus. Proc. Wash. Acad.
Sci., vol. 9, pp. 79-125.

Auuis, E. P. 1897 The cranial muscles and cranial and first spinal nerves in
Amia.calva. Jour. Morph., vol. 12, pp. 487-808.

1900 The pseudobranchial circulationin Amiacalva. Zool.Jahrb. Abt.
Anat., Bd. 14, pp. 107-134.

1908 a The pseudobranchial and carotid arteries in Ameiurus. Anat.
Anz., Bd. 33, pp. 256-270.

1908 b The pseudobranchial and carotid arteries in Polypterus. Anat.
Anz., Bd. 33, pp. 217-227.

1908 c The pseudobranchial and carotid arteries in the gnathostome
fishes. Zool. Jahrb., Abt. Anat. Bd. 27, pp. 103-104.

1911 The pseudobranchial and carotid arteries in Polyodon spathula.
Anat. Anz., Bd. 39, pp. 257-262, 282-293.
Ayers, H. 1889 The morphology of the carotids, based on a study of the blood.

vessels of Chlamydoselachus anguineus, Garman. Bull. Mus. Comp-
Zool. Harvard College, vol. 17, pp, 191-223.

Boas, J. E. V. 1880a Uber Herz und Arterienbogen bei Ceratodus und Pro-
topterus. Morph. Jahrb., Bd. 6, pp. 321-353.

1880b Uber den Conus arteriosus bei Buterinus und bei anderen Knoch-
enfischen. Ibid, pp. 527-533.

Brivce, T. W. 1879 On the osteology of Polyodon folium. Phil. Trans. Roy.
Soc. London, vol. 179, pp. 683-733.

Carazzi, D. 1904 Sulla circolazione arteriosa cardiaca ed esofagea della Scyl-
lium catulus. Intern. Monatschr. Anat., Physiol.. Bd. 21, pp. 1-20.

Cavauib, M. 1904 La vésicule bilaire et sa circulation artérielle, chez quelques
poissons de mer (Torpedo galvani, Scyllium catulus, Galeus canis.)
C. R. Soe. Biol. Paris, T. 55, pp. 1386-1388.

DanrortH, C. H. 1911 A 74 mm. Polyodon. Biol. Bull., vol. 20, pp. 201-204.

Fereuson, J. 8. 1911 The anatomy of the thyroid gland of Elasmobranchs,
with remarks upon the hypobranchial circulation of these fishes. Amer.
Jour. Anat., vol.11, pp. 151-208.

GreiL, A. 1903 Ueber die Entwickelung des Truncus arteriosus der Anamnia.
Verh. Anat. Ges., Bd. 17, pp. 91-105.

Gupernatscu, J. F. 1911 The thyreoid gland of the Teleosts. Jour. Morph.,
vol. 21, no. 4, Supplement, pp. 709-782.

Hyrtt, J. 1858 Das Arterielle Gefisssystem der Rochen. Denkschr. d. Kais.
Akad. d. Wissens. Wien, Bd. 15.

454 Cc. H. DANFORTH

Jorpan, D.S. 1899 A manual of the vertebrate animals of the northern United
States. Chicago. Eighth ed. p. 33.

Maurer, F. 1888 Die Kiemen und ihre Gefisse bei anuren und urodelen Am-
phibien, und die Umbildungen der beiden ersten Arterienbogen bei
Teleostiern. Morph. Jahrb., Bd. 14, pp. 175-222.

Parker, G.H. 1900 Note on the blood vessels of the heart in the sunfish(Ortha-
goriscus mola Linn.). Anat. Anz., Bd. 17, pp. 313-316.

Parker, G. H., and Davis, F. K. 1899 The blood vessels of the heart in Chara-
charias, Raja, and Amia. Proc. Boston. Soc. Nat. Hist., vol. 29, pp.
163-178.

Parker, T. J. 1884 A course in zootomy. London.

1886 On the blood-vessels of Mustelus antarticus. Philos. Trans.
Roy. Soe., London, vol. 177. ;

Pirzorno, M. 1905 Ricerche di morphologia comparata sopra le arterie suc-
clavia ed ascellare Selachii. Monit. Zool. Ital., vol. 16, pp. 94-103.

Senior, H. D. 1907a The conus arteriosus in Tarpon atlanticus (Cuvier and
Valenciennes). Biol. Bull., vol. 12, pp. 146-151.

1907, b Note on the conus arteriosus of Megalops cyprinoides (Brous-
sonet). Biol. Bull., vol. 12, pp. 378-379.

Sinvester, C. F. 1904 The blood-vescular system of the tile-fish, Lopholatilus
chamaeleonticeps. Bull. Bureau Fisheries, vol 24, pp. 89-144.

Stour, P. 1876 Ueber den Klappenapparat im Conus arteriosus der Selachier
und Ganoiden. Morph. Jahrb., Bd. 2, pp. 197-228.

WisHe, J. W. van. 1882 Ueber das Visceralskelett und die Nerven des Kopfes
der Ganoiden und von Ceratodus. Niederl. Archiv Zool., Bd. 5.

Wriaeut, R. R. 1885 On the hyomandibular clefts and pseudobranchs of Lepi-
dosteus and Amia. Journ. Anat. and Physiol., vol 19.

THE EMBRYOLOGY OF CRYPTOBRANCHUS ALLEGHE-

II. GENERAL EMBRYONIC AND LARVAL DEVELOPMENT,
SPECIAL REFERENCE TO EXTERNAL FEATURES!

Vi
WALES

VIII.

I:

X.

XT.
XII.

NIENSIS, INCLUDING COMPARISONS WITH
SOME OTHER VERTEBRATES

BERTRAM G. SMITH

From the Zoological Laboratory of Columbia University

TWO HUNDRED AND TWENTY-THREE FIGURES (EIGHT PLATES)

CONTENTS

TMC ROLLE EMORAs Sse atc Seas tect o- 0's tees Seis GP gat sid an eee ree
[Olea 30 to ea Aer eve Ot he ee ena Ae bee Ht eee ee Raa
AP eDESeripLiONy DY} StACES wages es. ume esissG Gye ae lain eas
Bee SUMMA y: Weert See mtney. sean ee ai cue seis wre, seat g ladeen ee
Gastrulation and early formation of the embryo......................
en MescertptioneOy, SUACESR a oad s dette heel eo keel A ene eee
JE, WSRUTCTRE ETA CeNe ean ihe Cha Ciecha ei eo ane pi Pan Ee RIE caearhct Aas 3
Development after the closure of the neural folds............
A. Description by stages to the time of hatching...................
B. Larval development, and the metamorphosis....................
@reRost=lanvalistacesessue ren me eed elet enna cma enone
AS tI Pees cae yay eaetns pas eh i ecy Saevey > agers Gap eae ean
{i Mriea ee teCONG Lay Me Mente ae oko oie ees einen ean eR SRAIE ohe Pig a cic
JM onaKera CC INNIT SS dake clan Magen ees Sa PERE orcs! SO A.e
Phy lemeniyecc ye ae rtenter pe dao! et hv aeal ote afl # nnn eo alee
Bibliography

VI. INTRODUCTION

WITH

458

imepareeiste 458

485
486
486
518
519
519
528

S Bicha Ee d31

532

The present contribution is one of a series of papers on the |
embryology of Cryptobranchus, of which Part I, dealing with
the breeding habits, ovogenesis, maturation and fertilization,

has already been published (Smith 712).

In the preparation of

1 Part I was published in the Journal of Morphology, vol. 23, no. 1, March, 1912,

455

456 BERTRAM G. SMITH

this paper I am greatly indebted to Prof. Bashford Dean, under
whose guidance it has been brought to completion.

During the past seven years the collection of an abundance of
material has enabled me to preserve an ample supply in every
stage of development. At least fifteen thousand embryos have
been secured from nests, and nearly as many more have been
obtained by artificial fertilization.

For convenience in description, the embryonic and larval his-
tory has been divided into stages, based chiefly on external char-
acters. In making the division, the usual difficulty has been
encountered, that the rate of development of each structure
varies more or less in different embryos. In detetmining what
shall constitute the interval between stages, the guiding principle
has been to establish stages only so far apart that individual
variations in the rate of development of the most important
characteristics selected as criteria for classification shall not
overlap. For purposes of more intensive study, each stage may
be divided into tenths; this device is useful in following any single
character or set of characters. Since development is a continu-
ous process, the importance of studying a close series cannot be
too strongly emphasized.

As an aid to obtaining the exact sequence of events, stress
has been laid on the study of a series of stages preserved at short
intervals, from a single lot of eggs fertilized at the same time.
Every period of the embryonic development has been covered
repeatedly in this way, an entire spawning of eggs being some-
times used in the study of three or four stages as distinguished
in this account. Thus not only a close series, but a large number
of embryos in each stage, representing several different spawnings,
have been studied, so that the typical course of development
could readily be distinguished from variations or abnormalities.

Moreover, the entire embryonic and larval history has been .
carefully followed in living material, repeatedly and for the most
part with embryos freshly collected. The study of living mate-
rial is of especial importance in the late cleavage and gastrula
stages of Cryptobranchus; for in these stages the translucent
condition of parts of the unpigmented embryo enables one to

EMBRYOLOGY OF CRYPTOBRANCHUS 457

gain a fair idea of what is going on inside. The ciliation of the
ectoderm of the late embryo, and some features of the circulatory
system, are best studied in living material.

An accurate and complete time record (Section X) of the
course of development has been obtained by comparison of many
different records of material kept alive during long periods of
time; these results were checked by observing the rate of devel-
opment of material freshly collected. One lot of embryos, col-
lected in the fall of 1906 at the time of the closure of the neural
folds, were kept alive in the Zoological Laboratory of the Univer-
sity of Michigan throughout the entire larval period, and their
metamorphosis was observed at the end of the second year after
fertilization. Specimens were preserved at intervals; shortly
after metamorphosis the half-dozen remaining individuals died
from causes unknown. Another lot of embryos collected in
an advanced gastrula stage in the fall of 1910, were kept alive
and in good condition in the Zoological Laboratory of the Univer-
sity of Wisconsin until May, 1911, when the last ones were pre-
served. .

The study of external and internal structure has gone hand
in hand, except for the post-gastrula stages; here, doubtful points
in the interpretation of the external structure have in most cases
been investigated by reference to serial sections.

In preparing the illustrations, composite or ideal figures have
been avoided. Each drawing, unless otherwise specifically
stated, is a faithful representation of an individual embryo; a
sufficient variety of figures has been given to illustrate the most
important deviations from the condition regarded as typical.
All the drawings are the work of the author except figures 268
to 276 which were drawn by Prof. Bashford Dean and with his
generous permission are here published for the first time; figure
203 which was kindly contributed by Dr. L. Hussakoff of the
American Museum of Natural History; and figures 277 to 279
which were drawn by Miss Hedge of Columbia University.

The histological technique employed has already been given
in Part I; it remains to record the methods used in photography.
For embryonic stages, fixation in Solution B (see Part I) followed

458 BERTRAM G. SMITH

by preservation in formalin, gives the best photographic results.
While being photographed, the objects were immersed in water
or formalin. Living larvae were anesthetized with chloretone.
In all cases the exposure was made by daylight.

All the photographs with a magnification of «K 4 were made
with a Bausch and Lomb Zeiss Tessar 72 mm. lens, fitted to a
long bellows Pony Premo No. 6 camera. The camera was fas-
tened in an erect position by means of an improvised wooden
frame. Figures 262, 264, 265 and 266 were taken with a Zeiss
Unar Lens; figures 263 and 267 with a Zeiss Apochromatic Planar.
Seed’s Non-halation- plates were used throughout; they were
developed with Adurol. :

All the negatives are the work of the author except figure 266
which was made by Miss Frances J. Dunbar of the University
of Michigan.

The photographs are untouched, except in a very few cases
' for the purpose of correcting slight defects in the negatives.

VIT. CLEAVAGE

A. Description of cleavage by stages

Stage 1: (figs. 57 to 61 and 204). This stage ischaracterized
by the presence of the first cleavage furrow only. The germinal
area reaches nearly to the equator. |

In artifically fertilized eggs the first cleavage furrow ordinarily
appears about twenty-four hours after fertilization; the time
may vary from eighteen to twenty-eight hours. The furrow
begins as a pit, usually at the animal pole; it lengthens rapidly
at first, then more slowly as it invades the regions of the egg more
heavily laden with yolk. After the first cleavage furrow is well ~
established, it becomes narrow in its middle portion, while still
broad at the ends. ‘The first cleavage furrow of the typical form
becomes superficially complete in Stage 3 (third cleavage).

In the typical condition, the first cleavage furrow passes through
the animal pole, and the division is equal (figs. 57 and 204).
Variations from this condition occur. To test the amount of

EMBRYOLOGY OF CRYPTOBRANCHUS 459

variation, sixty eggs from a single spawning were examined in
the first cleavage stage. In forty-eight cases the condition was
of the typical character described above. In seven cases the
first cleavage furrow was straight, but the cleavage unequal;
figure 58 represents the extreme of this condition. In three
eases, the first cleavage furrow passed through the animal pole,
but its halves met at this pole to form an obtuse angle (fig. 59).

ODOC
ODO

Figs. 57 to 64 Types of first and second cleavage of Cryptobranchus alle-
gheniensis. X 33.

Fig. 57 The first cleavage furrow extends just to the equator, a little further
than is usual before the appearance of the second furrow.

Fig. 62 The first cleavage furrow extends a little below the equator.

Fig. 64 The first cleavage furrow extends alittle below the equator; the second
furrow extends just to the equator.

In two cases the first cleavage furrow formed a semicircle about
the animal pole (fig. 60). In two cases from different spawnings,
neither of which furnished the material for the above data, cir-
cular first cleavage (fig. 61) was found; in each case the animal
pole was excentrically situated within the area bounded by the
cleavage furrow.

Goodale (’11) reports a case of circular first cleavage in Spe-
lerpes; the egg gave rise to a normal embryo. According to
Eycleshymer (’04), in Necturus the cytoplasm is always unequally
divided by the first cleavage, giving rise to blastomeres which
in many cases are decidedly unequal.

460 BERTRAM G. SMITH

Stage 2: (figs. 62 to 64 and 205). The second cleavage furrow
makes its appearance about six hours after the first, which by
this time extends nearly to the equator of the egg. In the stage
represented in figure 205, the first cleavage furrow has just reached
the equator. The second cleavage furrow usually becomes super-
ficially complete in Stage 4 (fourth cleavage), quite uniformly
meeting the first cleavage furrow at right angles at the vegetal
pole.

The earliest indication of the second cleavage furrow is usually
a roughness in the region of the animal pole where the second
groove is to intersect the first. The occurrence of ‘Faltenkranzen’
—a quivering of the surface with the formation of fine radiating
or parallel wrinkles, which extend outward from the cleavage
furrow—is quite marked at the time of the initiation of the second
cleavage furrow. The cause of the formation of similar wrinkles
in the frog’s egg has been investigated by Charles B. Wilson
(96). For some time after its first appearance the second furrow
is much broader, though of course shallower, than the first.

The second cleavage furrows usually depart from the same
point as the first furrow, and proceed vertically, forming a single
straight line at right angles to the first furrow (figs. 62 and 205).
Occasionally the points of departure of the second furrows do
not coincide, as shown in figures 63 and 64. The condition
shown in figure 63 is rarely observed, and is transitional to that:
shown in figure 64, which is quite frequent. That portion of the
first cleavage furrow lying between the points of departure of
the second furrows may be called the polar furrow. As shown
by the individual histories of a large number of eggs, in all cases
in which a polar furrow is present the points of departure of the
second cleavage furrows are separate from the beginning; the
polar furrow at first exists as a part of the straight first cleavage
furrow, but later becomes oblique through the shifting of cells.
In no ease in which the second cleavage furrows have their origin
from the same point in the first cleavage furrow, has there ever
been observed any shifting of cells during this or the following
stage, of such a nature as to produce a polar furrow.

EMBRYOLOGY OF CRYPTOBRANCHUS 461

Comparison with the second cleavage of Necturus as figured
by Eycleshymer (’04), Eycleshymer and Wilson (’10), and in
some unpublished drawings of the living ege by Prof. Bashford
Dean, leads to the conclusion that in Necturus there is much
greater irregularity in the second cleavage than in Crypto-
branchus.

Stage 3: (figs. 65 to 72 and 206). The third cleavage furrows
appear about five hours after the beginning of the second; hence
the interval is shorter than that between the first and second.
At the time when the third furrows begin, the first furrow has
usually reached or passed the equator.

31 3
\
i 3) 3
2 2 2 260 2 r
3
3
65 66 67 68
| fs ee l
3
3
2
2 2 Aur
2
3 {
|
70 71 ie

| |
69
Figs. 65 to 72 Types of third cleavage of Crytobranchus allegheniensis. All
the figures are of the upper hemisphere except figure 66 which represents the lower
hemisphere of the egg shown in figure 65. In no case does any cleavage furrow
except the first reach the lower pole. All the figures are camera drawings of pre-
served material. X 34.

As the cleavage furrows invade the more heavily yolk-laden
lower hemisphere they become comparatively faint except at
their extreme lower ends where they broaden out. During this
_ Stage the second and third cleavage furrows are in general broader
than the first.

At the stage represented in figure 65, when the third furrows
are well established but extend only a short distance from their

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

462 BERTRAM G. SMITH

respective points of origin, the first cleavage furrow has reached
the lower pole where its ends unite. The first cleavage furrow
thus becomes superficially complete, thereby establishing the
holoblastic character of the egg. That there is a strong mero-
blastic tendency is already apparent. In the region of the vege-
tal pole the first cleavage furrow is at first broad, but it later
becomes narrow and faint.

With very few exceptions, the third cleavage furrows depart
from the second furrow, at some little distance from its point of
intersection with the first. In a previous paper (Smith ’06)
it was erroneously stated that the third cleavage furrows usually
depart from the first furrow.

The third cleavage furrows ordinarily begin as two pits in
the second furrow, equidistant from its point of intersection at
the animal pole. From these two pits the third cleavage furrows
ordinarily proceed in an approximately vertical direction (fig.
65), and do not become complete until a later stage (Stage 5).

From the time of the earliest appearance of the third cleavage
furrows, the distances from the first cleavage furrow to their
points of departure from the second remain unaltered; but the
second cleavage furrow, originally straight, often becomes drawn
into a zig-zag line, as shown in the figures.

As will be seen from a study of later stages, the third cleavage
furrowsrarely reach the vegetal pole, but as a rule extend obliquely
in the lower hemisphere to join the first furrow at some distance
from the lower pole (figs. 84 and 91 to 96). Hence the general
statement may be made that the third cleavage furrows are inter-
mediate between a true meridional and a true latitudinal cleavage
but approach more nearly to the former type.

In the typical condition, the cleavage pattern has now lost
its strictly radial, and acquired a biradial symmetry. I have
purposely avoided the use of the word bilateral in this connection,
not only because it does not fit the case so well as the word 67-
radial, but in order to avoid the inference that the condition has
anything to do with the bilateral symmetry of the future embryo.
As will be seen by consulting the figures, this biradial condition
of the cleavage pattern persists in the lower hemisphere through-

EMBRYOLOGY OF CRYPTOBRANCHUS 463

out the late cleavage stages, and in some eggs enables one to
identify the early cleavage furrows even after the beginning of
gastrulation.

Deviations from the type in Stage 3 show a series of conditions
connecting the typical one with a true latitudinal third cleavage.
In such eases the third cleavage furrow proceeds more obliquely,
and at an earlier stage joins the first nearer the animal pole (figs.
68 to 72). In some cases one or more of the third cleavage fur-
rows are truly latitudinal (figs. 70 to 72).

Rare cases occur in which a third cleavage furrow originates
at the animal pole, or from a first cleavage furrow (see fig. 67
for an example of the latter case); occasionally, a third cleavage
furrow may reach the vegetal pole, or unite with a second cleavage
furrow near the vegetal pole (figs. 94 and 95).

In comparing the third cleavage pattern of Cryptobranchus
with that of other forms, one of the most obvious generalizations
brought out is that a vertical third cleavage is characteristic of
heavily yolk-laden and highly telolecithal eggs: e.g., the squid
(Watase 791); Amia (Dean ’96, Whitman and Eycleshymer ’97);
Lepidosteus (Dean 795, Eycleshymer ’99); Acipenser (Dean
95); Ctenolabrus (Agassiz and Whitman ’84); Serranus (H. V.
Wilson 791); Ceratodus (Semon ’00 and ’01); Lepidosiren (Kerr
700 and ’09); Cryptobranchus japonicus (deBussy ’04 and ’05);
and the pigeon (Blount ’07). But the rule is not absolute; con-
cerning the third cleavage of Necturus, Eycleshymer (’04) says:
‘In most cases the cleavage grooves are irregularly formed and
it might be said that the variations are so numerous and so di-
verse that a special description must be written for each egg.”
From this statement and an inspection of his figures (seealso
Kycleshymer and Wilson 710), it appears that a type cannot be
recognized for the third cleavage of this egg; that the irregularity
is greater than in the case of Cryptobranchus and that there is
a more marked tendency for the third cleavage furrows to come
in latitudinally. My material for the very early cleavage stages
of Necturus is too scanty to enable me to form any conclusion
based on direct observations, but some unpublished figures of
the early cleavage of Necturus drawn from the living egg by

464 BERTRAM G. SMITH

Prof. Bashford Dean give confirmatory evidence of the irreg-
ularity of the third cleavage furrows. In Desmognathus the
third cleavage furrows were vertical and regular in the few eggs
studied by Wilder (04); the third furrows depart from an earlier
cleavage furrow at some distance from the animal pole. But
Hilton (04 and ’09), who examined a considerable number of
eggs of Desmognathus in this stage, states that this regular and
vertical form of cleavage occurred in only two or three eggs;
in the others the third cleavage was irregular. In Diemyctylus
(Jordan ’93) there is still greater irregularity in the third cleavage
than is recorded for Necturus or Desmognathus: ‘‘ With the com-
pletion of the second furrow all consistent regularity is at an
end.”

In eggs less heavily yolk-laden, as in Amblystoma (Eycleshy-
mer 95) and the frog, the third cleavage ts latitudinal.

Especially interesting from a comparative point of view are
Budgett’s observations on the cleavage of the crossopterygian
Polypterus as given by Kerr (07). ‘‘From Budgett’s pen
and ink sketch . . . . wecansee that the segmentation
is at first characterized by its almost absolutely equal character.
We may infer with considerable certainty that the two merid-
ional furrows are succeeded by a latitudinal one which is prac-
tically equatorial.” The egg of Polypterus is small, having a
diameter of a little over one millimeter.

In urodeles we find a condition intermediate between the ver-
tical third cleavage characteristic of the fishes generally, and
the latitudinal third cleavage of the anura. In Cryptobranchus
the vertical type prevails; in Desmognathus, Necturus and Die-
myctylus there is increasing irregularity; in Amblystoma the
third cleavage is latitudinal. The possible phylogenetic signifi-
cance of the cleavage of Crytpobranchus will be considered later.

As a rather general rule, in eggs in which the third cleavage
is usually vertical, the third furrows depart from the second
rather than from the first or from the animal pole. As has
already been seen in the case of Cryptobranchus allegheniensis,
this rule is by no means absolute; but in general it applies also
to the squid, and to the teleosts (e.g., Ctenolabrus, Serranus).
DeBussy (’04 and ’05) has described the cleavage stages of

EMBRYOLOGY OF CRYPTOBRANCHUS 465

Cryptobranchus japonicus; his material was meager and lacked
first and second cleavage stages. He states (05) that in the five
eges examined in the third cleavage stage, all the third furrows
are approximately meridional. His figures (04) represent the
third furrows departing most frequently from the first cleavage
furrow, sometimes from the second, sometimes from the animal
pole. In the urodele Hynobius (Kunitomo 710) the cleavage
pattern in this stage resembles that of Cryptobranchus alleghe-
niensis, except, that the third furrows do not so often depart from
the second furrow. In Amia, Whitman and Eycleshymer (’97)
describe the third cleavage furrow as follows:

In the majority of cases they are vertical . . . . They gen-
erally all depart from one or the other of the first two meridionals, thus
giving rise to a distinct bilateral appearance . . . . It oftens

occurs that one or more of the set depart from the first meridional,
while the rest depart from the second, or vice versa.

Stage 4: (figs. 73 to 84 and 207 to 209). This stage is charac-
terized by the appearance of the fourth cleavage furrows, giving,
when complete, sixteen cells. As will appear from the following
observations, the number of micromeres is not constant, but
varies from four to eight.

The fourth cleavage furrows appear about four hours after
the beginning of the third, and about thirty-nine hours after
fertilization. Ordinarily they begin as two grooves, cutting the
first cleavage furrow at right angles, on each side of the second
and a short distance from it (figs. 73 and 74). Thus in position
and direction the fourth cleavage furrows alternate with the
third, which cut the second at approximately right angles. The
fourth cleavage is the first one to cut off micromeres from macro-
meres, and the division is very unequal.

In a given lot of eggs the sixteen-cell stage is reached quite
uniformly at the same time, but with so much variation in the
direction of the fourth cleavage furrows that at first sight no
uniformity is recognizable. By the study of a large number of
eggs the following generalizations are established:

(a) In each quadrant, one micromere is cut off between the
first and second cleavage furrows, giving in each egg a minimum
of four micromeres (fig. 74).

466 BERTRAM G. SMITH

Figs. 73 to 81 Fourth cleavage of Cryptobranchus allegheniensis, X 42. All
the figures are of the upper hemisphere. Figure 73 and 74 are drawn from the
living egg; the others are camera drawings from preserved material. Figures
73 and 74 represent early stages of fourth cleavage; figure 80 is from the same egg
shown in figure 74, representing the condition three hours later. Figure 81 is
drawn from the egg photographed for figure 209.

(b) From the third cleavage furrows the remaining four
parts of the fourth cleavage furrows may continue latitudinally,
forming a complete circle or oval enclosing eight micromeres
(fig. 81); or one or more of these four parts may continue approx-

aie

EMBRYOLOGY OF CRYPTOBRANCHUS 467

imately parallel to the second cleavage furrow, extending verti-
cally and increasing the number of macromeres instead of cutting
off micromeres (figs. 75 to 80). Thus while the total number
of cells is always sixteen, the number of micromeres varies from
four to eight.

We thus get as one extreme type an approximately latitudinal
fourth cleavage furrow; as the other extreme a fourth cleavage
furrow divided into two separate grooves, one on each side of
the second furrow and approximately parallel to it and to each
other. Between these two extremes we find examples of all
possible intermediate conditions.

With regard to the manner of fourth cleavage, eggs of this
stage may be classified into five types, depending on the number
of micromeres present. For the purpose of such a classification,
irregularities in the third cleavage must be allowed for: in cases
where the third cleavage has come in diagonally or latitudinally
to cut off a small cell, such a cell is divided by the fourth cleavage
into two small cells, of which only the one nearer the animal
pole is to be counted as a micromere (fig. 77).

To determine the mode, twenty-five eggs were examined in
the sixteen-cell stage, and the results tabulated as follows:

Nimmmb Eriol mmrcromenresesce wae se eeaeanS if 6 5 4
INumbernlot iGasess-eue- io asec eee eens 4 7 8 3 3

The table shows that the most frequent manner of cleavage
is intermediate between the two extremes described.

In the majority of cases the micromeres are arranged with
considerable regularity ini two parallel rows, separated by the
second cleavage furrow (see especially figs. 78 and 80). This
is the necessary result of the biradial symmetry instituted by
the normal mode of third cleavage, providing there is no exten-
sive shifting of the micromeres. The condition reminds one of
the cleavage pattern of the corresponding stage of the teleost
ege.

Through a shifting of the micromeres, the biradial symmetry
of the cleavage pattern of the blastodisc is usually interfered
with (figs. 75 to 81). In this region, the first and second cleav-

468 BERTRAM G. SMITH

age furrows become irregular and broken to an extent never
observed in earlier stages. Outside of the region of the micro-
meres, the biradial pattern of cleavage is retained.

In this as in other cleavage stages the most recent furrows,
and especially the most recent portions of such furrows, are in
general quite noticeably the widest. This fact once established
may be made use of in connection with other evidence to iden-
tify cleavage furrows. The broadening of the ends of tbe ver-
tical furrows as they invade the lower hemisphere is a fairly con-
stant feature of the cleavage; as shown by unpublished drawings
from living material by Prof. Bashford Dean, it is also well
expressed in the eggs of Necturus.

Psa ears : iT age
82 83 84

Figs 82 to 84 Lower hemispheres of fourth cleavage stages of Cryptobranchus
allegheniensis. All the figures are camera drawings of preserved material. xX
42. ;

Fig. 82. Lower hemisphere of the egg shown in figure 208.

Fig. 83. Lower hemisphere of the egg shown in figure 76. This figure would
serve equally well to represent the lower hemisphere of the egg drawn for figure 75.

Fig. 84 Lower hemisphere of the egg represented in figure 80.

In the majority of eggs of this stage, the second cleavage fur-
rows have reached the lower pole, and the third furrows have
just passed the equator (figs. 82 to 84). The second cleavage
furrows intersect the first at right angles at the lower pole.
For some distance on each side of the pole the second cleavage
furrows are for a time markedly wider than the first. The second
furrows are further distinguished by the fact that the third
furrows run closer to them than to the first. In the latter part
of this stage the third furrows sometimes become complete (fig.
84), as a rule joining the first at some distance from the pole.

EMBRYOLOGY OF CRYPTOBRANCHUS 469

The biradial pattern of cleavage is thus preserved in the lower
hemisphere, and throughout the later cleavage stages affords
a trustworthy means of distinguishing first and second cleavage
furrows in this region.

The fate of the fourth cleavage furrows that proceed vertically
must be studied in later stages. They usually join the second
furrow before reaching the lower pole (figs. 92 and 96).

In a given egg the micromeres vary somewhat in size; but a
comparison of seventeen carefully drawn camera figures, and
the examination of a large number of additional eggs, lead to
the conclusion that in this stage there is no regularity in the
distribution of large and small cells among the micromeres.

DeBussy’s (’04) single figure of the fourth cleavage stage
of Cryptobranchus japonicus shows six micromeres surrounded
by an approximately circular cleavage furrow, and two recent
furrows extending for a short distance vertically.

In Desmognathus, according to Wilder (’04), the fourth cleay-
age is latitudinal; this conclusion was based on the study of mate-
rial very limited in amount. Hilton (’09) states that in a large
number of eggs he has found only a few which exhibit so regular
a type of cleavage as described in Wilder’s eight cell and later
stages.

In Hynobius (Kunitomo ’10), the fourth cleavage furrows are
more uniformly latitudinal than in Cryptobranchus allegheniensis.
In Necturus (Eycleshymer ’04; Eycleshymer and Wilson ’10)
and Diemyctylus (Jordan ’93) a type is no longer recognizable.

In Ceratodus (Semon ’00 and ’01) the fourth cleavage is lati-
tudinal. Amia (Dean ’96; Whitman and Eycleshymer ’97)
and Lepidosteus (Dean ’95; Eycleshymer ’99) resemble the type
with four micromeres described for Cryptobranchus alleghe-
niensis.

Stage 5: (figs. 85 to 96; 210 and 211. This stage is reached
about four hours later than Stage 4. It is characterized by
the presence of the fifth cleavage furrows, giving a maximum of
thirty-two cells, some incompletely divided. More than half
of these cells are micromeres.

470 BERTRAM G. SMITH

Figs. 85 to 90 Upper hemispheres of eggs of Cryptobranchus allegheniensis
in the fifth cleavage stage. All the figures are camera drawings from preserved
material. Figure 86 is drawn from the egg photographed for figure 211, and figure
87 from the egg photographed for figure 210.. X 42.

The careful study of a large number of eggs emphasizes
irregularity in this cleavage and the absence of a well established
type. Two sets of fifth cleavage furrows are often recognizable:
an inner, within the former region of micromeres, and an outer,
just outside of this region. Either set may be, in whole or in
part, vertical, latitudinal or oblique (see especially figure 85
for an example of outer, and figure 89 for an example of inner
latitudinal cleavage furrows). A study of the most regular
cases of cleavage described under Stage 6 shows that in these
eggs the fifth cleavage furrows must have come in with greater
regularity than in any eggs directly observed in the fifth cleavage
stage: these fifth cleavage furrows are almost uniformly vertical,
thus preserving the regular alternation in the direction of the
furrows.

EMBRYOLOGY OF CRYPTOBRANCHUS 471

By the shifting of micromeres the biradial symmetry due to
the manner of third cleavage is usually lost in the blastodise,
and unless the egg has been kept under continuous observation
it becomes in most cases impossible to trace the first and second
cleavage furrows entirely through the region of micromeres.

In preserved material, nuclei are visible from the surface in
some of the micromeres of this and the following stages, indicating
that these cells are becoming flattened out.

As already noted, in this stage if not in the preceding one,
the third cleavage furrows become complete, usually joining the
first at some distance from the pole (figs. 91 to 96). This appar-
ent avoidance of the pole by the third cleavage furrow is doubt-

| Ss. les I
3 3
2E5r5 a
2 4 =
4
om igh 34 Sh elite
91 92
|
3 3 3 { : Bes
4
4 a
2 2 2 D} 9 9/
4 #
3 fo
3 fe 3 |
94 95 96

Figs. 91 to 96 Lower hemispheres of eggs of Cryptobranchus allegheniensis
in the fifth cleavage stage. All the figures are camera drawings from preserved
material. Figure 93 shows a persistent sperm pit (see Part I, Smith ’12). x 42.

Fig. 91 Lower hemisphere of the egg whose upper hemisphere is shown in
figure 87. ;

Fig. 96 Lower hemisphere of the egg whose upper hemisphere is shown in
figure 85. The fourth cleavage furrows have extended further than is usual in
eggs of this stage.

472 BERTRAM G. SMITH

less the mechanical result of the location of the earlier course of
the third furrows nearer to the first than to the second; they
swerve from the vertical toward the nearest existing cleavage
furrow. An analogous pattern may sometimes be observed in
the cracking of the corners of a section of cement walk.

I have observed this biradial cleavage pattern in corresponding
stages of the lower hemisphere of occasional eggs of Necturus;
it is clearly expressed in the cleavage of Desmognathus as figured
by Wilder (’04) and Hilton (04 and ’09).

The same tendency to join the nearest existing vertical furrow
is shown by those fourth cleavage furrows, as a rule not yet com-
plete, that come in vertically. They usually join the second
furrow, at a much greater distance from the lower pole than the
intersection of the third with the first.

In the vicinity of the vegetal pole, both first and second cleav-
age furrows are now only faintly expressed.

In about half the eggs of this stage cell division has proceeded
a little more rapidly on one side of the egg than on the other;
the cells are smaller in surface view, more numerous, and the
cleavage furrows are more uniformly complete (figs. 87, 88 and
210). Thus there is an excentric development of the blasto-
disc, whereby a condition of bilateral symmetry in the cleavage
pattern is produced. This excentric development is a more
constant feature in the stages immediately following. The
question naturally arises whether this bilateral symmetry in
the cleavage pattern has any morphogenetic significance: is it
the outward expression of the establishment of the permanent
bilateral symmetry and antero-posterior differentiation of the
embryo? In other words, does the axis of bilateral symmetry
in the cleavage pattern fall in the median plane of the future
embryo? ‘This question will be considered in a later paper, in
connection with the study of the internal development.

Such an excentric development of the blastodise has been
described for Amblystoma and Necturus by Eycleshymer (’95
and ’04), who cites similar observations on other vertebrates
by various writers. Eycleshymer speaks of a ‘‘second area of
accelerated cell division” as distinguished from the primary area

EMBRYOLOGY OF CRYPTOBRANCHUS 473

of cell division at the animal pole. In Cryptobranchus what
happens seems to be a shifting of the most active area of cell
division to an-excentric position in the blastodisc; hence I have
preferred to speak of it merely as a process of excentric develop-
ment.

No constant relation exists between the axis of bilateral sym-
metry due to excentric development and the original direction
of the first cleavage furrow as shown by those portions of it that
have not undergone shifting. The two may coincide (fig. 88) ;
they may be at right angles to one another; they may be oblique
(fig. 87).

As already indicated, in this stage the cleavage pattern of
Necturus bears a strong resemblance to that of Cryptobranchus
allegheriensis, but there is this marked difference: the third
cleavage furrows of Necturus, when vertical, usually join the
first at a greater distance from the vegetal pole, in the region of
the equator. In most eggs of Necturus examined in this stage
only the first two cleavage furrows extend into the lower hemi-
sphere; these usually meet at right angles at the vegetal pole.
Thus the cleavage of Necturus in this stage seems to show an
even stronger tendency toward the meroblastic condition. But
this is merely a consequence of the tendency for the third cleavage
furrows to come in obliquely or latitudinally; a comparison of
later stages shows that the meroblastic tendency is in reality a
trifle less strongly expressed in Necturus (figs. 107 and 108) than
in Cryptobranchus.

In Amia (Dean ’96; Whitman and Eycleshymer ’97) the fifth
cleavage furrows appear in two sets: an outer set cutting the
eight macromeres latitudinally; and an inner set cutting the
four micromeres in a horizontol plane, hence not visible from
the surface.

Stage 6: (figs. 97 to 102 and 212 to 214). This stage is charac-
terized by the presence of the sixth cleavage furrow, giving a
maximum of sixty-four cells, some of the macromeres being in-
completely divided. Considerably more than half the cells are
micromeres; these occupy an area whose diameter extends over
only about 90° of the circumference of the egg. Hence the mero-

102

et Bee: 97 to 102 Sixth cleavage (Stage 6) of Cryptobranchus allegheniensis.
igures 100 to 102 show upper hemisphere, equatorial view, and lower hemisphere,
respectively, of the same egg. All of the figures are camera drawings from pre-
a material. Figure 98 is drawn from the egg photographed for figure 212.
ts
474

EMBRYOLOGY OF CRYPTOBRANCHUS A475

blastic tendency is strongly expressed. This stage is reached
about four hours later than the beginning of the preceding stage.

A description of a few individual eggs will best indicate the
characteristics of this cleavage.

Out of about fifty eggs studied, the one represented in figure
97 shows the greatest regularity of cleavage in the upper hemi-
sphere. This condition must have been reached by a fairly
constant alternation of vertical and latitudinal cleavage furrows.
This alternation of cleavage furrows carried out with complete-
ness and geometrical precision would give a total of sixty-four
cells, consisting of forty-eight micromeres and sixteen macro-
meres; the micromeres would be arranged in three concentric
rows, each containing sixteen cells. In the egg under consider-
ation, this condition is realized in the outer row of cells, which
is quite regular and contains the theoretical number, sixteen.
But the total number of micromeres is only thirty-nine, hence
a deficiency must exist in the central portion of the blastodise,
or some divisions in this region must have taken place horizon-
‘tally. Sections show that no horizontal divisions ‘have taken
place in this region; on the other hand divisions parallel to the
surface have sometimes occurred in the marginal row of micro-
meres. Therefore cell division is taking place more rapidly in
the marginal than in the central region of micromeres—a condition
which may be the beginning of that accelerated development
of the margin, the later expression of which is almost wholly
internal.

A study of other eggs showing a fairly regular alternation of
cleavage furrows gives additional evidence for this interpretation
(e.g., fig. 100, representing an egg with 39 or 40 micromeres).
While eggs with this degree of regularity in the cleavage pattern
are the exception rather than the rule, it is felt that evidence
derived from them is especially trustworthy; for in such eggs
the equilibrium in the rhythmic alternation of the direction of
cell division has been best maintained, and the rather uniform
lagging-behind of the divisions of the cells in the central area
would seem to be the expression of a normal tendency in the life
of the embryo. In eggs which show disturbances in this equi-

476 BERTRAM G. SMITH

hbrium, discordant factors are more likely to be present to
obscure the normal expression of the course of development.

The sixth cleavage furrows of the outer set, when latitudinal,
divide the macromeres very unequally, cutting off additional
micromeres. The number of micromeres, and the extent of the
blastodise, is increased by such latitudinal divisions; the number
of macromeres is increased by the sixth cleavage furrows only
when these come in vertically.

In a few eggs, as the one shown in figure 99, there is a marked
tendency for the sixth cleavage furrows to come in vertically.
Here, as noted in an earlier stage, the embryo seems to be oscil-
lating between two possible modes of cleavage; but the tendency
to preserve the regular alternation of cleavage furrows is usually
the stronger.

The most marked tendency to vary from the regular pattern
of cleavage occurs along the line of excentric development of the
blastodise (figs. 98 and 212), as described under Stage 5. The
majority of eggs exhibit this tendency in some degree.

We have then, in the cleavage pattern of this stage, two tend-
encies toward differentiation of the blastodise: (a) an acceler-
ated cell division in the marginal portion, pointing toward the
formation. of the germ ring; and (b) an accelerated cell division
about a radius of the blastodise, giving a condition of bilateral
symmetry.

DeBussy’s figure (05, fig. 10) representing the blastodise of
an embryo of Cryptobranchus japonicus with forty micromeres
strongly suggests excentric development; on one side of the first
cleavage furrow only three cleavage furrows reach the equator,
on the other side nine. But the author remarks (p. 530) that
he has observed no secondary center of accelerated cell division
such as has been described by Eycleshymer for Necturus.

A comparison with earlier stages shows that there is an increas-
ing tendency for the micromeres, following a familiar law of
developmental mechanics, to lose their original quetraaen or
triangular outline and become hexagonal.

In the living egg, the roof of the segmentation cavity is some-
what translucent, and spaces communicating with the cavity

EMBRYOLOGY OF CRYPTOBRANCHUS 477

beneath sometimes appear between the cells. Evidently the
roof consists of a single layer of flattened cells; this inference is
confirmed by the study of sections.

On account of the slow cleavage and relative stability of the
macromeres, there is little change in the cleavage pattern of the
lower hemisphere. An advance is shown in that the vertical
fifth cleavage furrows have invaded the lower hemisphere (fig.
102). Those fourth cleavage furrows that proceed vertically are
seldom complete in this stage, but sometimes are found joining
an earlier vertical furrow at a considerable distance from the
vegetal pole.

In this stage the most regular type of cleavage pattern of Cryp-
tobranchus bears a striking resemblance to the corresponding
stage of Amia (Dean 796; Whitman and Eycleshymer ’97).

Since Stage 6 of the egg of Cryptobranchus best serves to illus-
trate the fundamental characteristics of the cleavage pattern,
particularly with regard to the relative size of the micromeres
and macromeres, at this point a comparison may well be made
with the dipnoi and crossopterygi. Thegeneralanuranor urodele
character of the cleavage of the dipnoan egg is apparent in all
existing genera: Ceratodus (Semon ’00 and ’01); Protopterus
(Budgett ’01; Kerr ’09); and Lepidosiren (Kerr ’00, ’01 and ’09).
With respect to inequality in the cleavage, Lepidosiren in partic-
ular closely approaches the condition in Cryptobranchus and
Necturus. The cleavage of Polypterus (Kerr ’07) bears a general
resemblance to that of Amblystoma and the frog.

Stage 7: (figs. 103 to 105; 215 and 216). This stage is charac-
terized by a doubling of the number of cells found in the preceding
stage, and by a slight extension of the region occupied by the
micromeres. The stage is reached about four hours later than
Stage 6.

Figures 103 and 104 show a fairly representative egg in this
stage. The cells in the region of the animal pole are markedly
larger than the other micromeres. This condition may be due
to one or both of two factors: (a) the flattening of the cells com-
posing the roof of the segmentation cavity; (b) a slower rate of
division in these cells, as noted in Stage 6. There is marked

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

478 BERTRAM G. SMITH

104

105 106

Figs. 103 to 105 Surface views of eggs of Cryptobranchus allegheniensis in
Stage 7, showing cleavage furrows. Figures 103 and 104 represent upper and
lower hemispheres, respectively, of the same egg. Camera drawings from pre-
served material. X 7.

Fig. 106 Equatorial view of an egg in Stage 8, showing cleavage pattern.
Camera drawing from preserved material. X 7. :

activity in cell division in an area excentrically situated, though
this is not so apparent in the particular egg under consideration
as in some other eggs in the same stage. In the lower hemisphere,
the biradial character of the cleavage pattern is preserved and
accentuated.

In the egg represented in figure 105 we see the beginning of
a process of fundamental importance in the further history of

EMBRYOLOGY OF CRYPTOBRANCHUS 479

the embryo—the phenomenon of immigration of cells from the
single-layered roof of the segmentation cavity. In a surface view,
it is evident that some of the cells in the excentrically situated
area of most rapid cell division are partially submerged. They
are not merely smaller superficially than the other micromeres,
but are sunken below the general surface and present the appear-
ance of being crowded inward. Their condition will be further
described in a consideration cf the internal development; their
later history and fate form an important phase of the process
of embryo-formation.

In most eggs of this stage, at the margin of the blastodise
oblique furrows (probably fifth cleavage furrows) occasionally
cut off cells intermediate in size between micromeres and macro-
meres. In the lower hemisphere some recent furrows, presum-
ably fifth cleavage furrows, usually extend well toward the vicin-
ity of the vegetal pole. The macromeres are, as a rule, long,
narrow and wedge-shaped. On account of the segregation of
most of the protoplasm within the region of the blastodisc, all
latitudinal divisions of the macromeres are very unequal, cutting
off new micromeres instead of increasing the number of macro-
meres. In the living egg, the roof of the segmentation cavity
still appears somewhat translucent, and spaces sometimes occur
between these cells; but neither of these conditions is so marked
as in the preceding stage.

Stage S: (figs. 106 and 217). This stage is reached about
twelve hours later than Stage 7; it is best described by reference
to the figures. The micromeres have become much more numer-
ous and smaller; there is a slight extension of their area. There
is a more gradual transition, or gradation in the size of the cells,
between micromeres and macromeres. In the lower hemisphere
the fifth cleavage furrows have, as a rule, become complete; they
rarely reach the lower pole, but join an earlier furrow at some
distance from the vegetal pole. Biradiality of the cleavage pat-
tern still enables one, as a rule, to distinguish first and second
cleavage furrows in this hemisphere.

In the upper hem'sphere the excentric area of accelerated cell
division noted in the preceding stages is usually quite marked.

ASO BERTRAM G. SMITH

In preserved material the nuclei of the micromeres are often
easily distinguishable in surface views

In the living egg, the roof of the segmentation cavity has be-
come quite opaque, and the cells are compactly arranged. During
the latter part of this stage a translucent condition begins to
appear at the animal pole, indicating a thinning-out of the cells
in this region, as in Stage 6; but this time the cells form a firm
tissue, with no spaces between them.

In Necturus the cleavage furrows in the region of the vegetal
pole are fainter than in Cryptobranchus; this condition is reversed

107 108

Figs. 107 and 108 Advanced cleavage stage of Necturus maculosus. . Two

views of a single egg, photographed after preservation. X 4.

in the upper hemisphere, where the micromeres are outlined
far more boldly in Necturus (figs. 107 and 108) than in Crypto-
branchus (both statements refer to preserved material, fixed by
the same method).

Stage 9: (figs. 109, 110 and 218). This stage is reached about
nineteen hours later than the preceding stage. Individual micro-
meres in the region of the animal pole are barely visible to the
naked eye. The zone of transition between micromeres and
macromeres has become broader and more marked. An excen-
trically situated area of accelerated cell division in the micromeres
is only occasionally found in surface views of this stage.

EMBRYOLOGY OF CRYPTOBRANCHUS 481

In the living egg, the roof of the segmentation cavity appears
as a translucent tissue throughout a circular area about 40 degrees
in diameter in the region of the animal pole. This indicates a
decided thinning-out of the cells of this region.

Biradiality of the cleavage pattern of the lower hemisphere
still enables one to distinguish in many embryos, though not
in all, the first and second cleavage furrows. Usually two or
three cells quite small in surface view occur at the vegetal pole;
they are quite characteristic of this and the following stage, but
are sometimes found in the preceding stage. At the vegetal

109 mile

. Figs. 109 and 110 Stage 9 of Cryptobranchus allegheniensis. Equatorial
view and lower hemisphere of different eggs, showing cleavage pattern. Camera
drawings from preserved material. X 7.

pole the cleavage furrows, both in living and preserved material,
are sometimes both broad and deep, forming quite noticeable
fissures; a similar condition is common in Amblystoma (Eycles-
hymer 795). In Cryptobranchus this condition is in marked con-
trast to the stage immediately preceding, in which the furrows
in this region were faint. In Necturus, on account of the varia-
bility of the third cleavage furrows, the biradial pattern of the
macromeres is not so clearly expressed as in the egg of Crypto-
branchus.

Stage 10: (figs. 111, 112 and 219). This stage, reached a day
or two later than Stage 9, immediately precedes the beginning
of gastrulation. The micromeres at the upper pole are invisible

482 BERTRAM G. SMITH

to the naked eye, and barely distinguishable with the magnifi-
cation used for photographing (x 4). The area occupied by
the micromeres extends approximately to the equator, though
the broad zone of transition makes it difficult to define. In the
vicinity of the vegetal pole, the cleavage furrows have again
become faint; in many cases, In preserved material, they are
distinguishable as fine lines lighter in color than the general sur-
face, rather than as actual grooves. For the accurate study of
these furrows in this and the following stage, a binocular micro-
scope is usually required. When their boundaries are distinct,
on account of their large size the macromeres are readily seen
with the naked eye.

jun
att:

ht:

co SS

Figs. 111 and 112 Lower hemispheres of two eggs of Cryptobranchus alleghe-
niensis in Stage 10, showing cleavage pattern. The embryo shown in figure 112
is slightly older than the one represented in figure 111. Camera drawings from
preserved material. In each egg, the lower pole as determined by gravity lies
at the center of the figure; the vegetal pole, at the intersection of the first two
cleavage furrows, is slightly above this point. The upper part of each figure rep-
resents the side on which the blastopore is to appear. X 7.

Usually, the cleavage pattern of the lower hemisphere retains
enough of its earlier bilateral symmetry to enable one to distin-
guish first and second cleavage furrows. ‘The vegetal pole, since
it occurs at the intersection of the first two cleavage furrows,
may in most cases still be determined quite accurately and con-
veniently by means of the cleavage pattern. As shown in figures
111 and 112, the vegetal pole is excentrically located in the area
occupied by the macromeres; a more rapid multiplication of
cells has occurred on one side of this area, so that on this side

EMBRYOLOGY OF CRYPTOBRANCHUS 483

the micromeres and transitional cells approach nearer the vegetal
pole. A meridian drawn through the vegetal pole and the center
of the area occupied by the macromeres defines the axis of excen-
tricity; this axis bears no constant relation to the first cleavage
furrow and the axes of biradial symmetry determined by the
early cleavage furrows. The biradial symmetry of the cleavage
pattern is of course somewhat disguised by the more rapid multi-
plication of cells at one end of the axis of excentricity.

In this stage occurs a slight tilting of the morphological axis
of the egg within a meridional plane determined by the axis of
excentricity, so that the vegetal pole no longer coincides with
the lower pole as determined by gravity. The vegetal pole is
slightly uptilted on the side where the more rapid multiplication
of cells occurs, hence the meridian defining the axis of excentricity
passes also through the new pole determined by gravity. This
new pole at first lies intermediate between the vegetal pole and
the center of the area occupied by the macromeres; in later
stages, through continued tilting of the egg in the same direction,
it comes to lie beyond this center. Throughout the ensuing
stages we must distinguish between the morphological axis of
the egg as determined by the animal and vegetal poles, and the
vertical axis determined by gravity. The method of locating
the vertical axis, and exact measurement of the amount of rota-
tion, will be given in the following stage.

If the egg be sectioned along the axis of excentricity the internal
structure, to be described later, shows that this axis lies in the
sagittal plane of the embryo; the side on which the small cells
approach nearer to the vegetal pole is the one on which the blasto-
pore is to appear. Thus the excentric position of the vegetal
pole within the area occupied by the macromeres enables one to
orient the egg with reference to future body regions.

In perhaps the majority of cases, the transition from large
to small cells is more evenly graded on the side where it occurs
nearest to the vegetal pole, than on the opposite side where it
is characterized by a rather abrupt line of demarcation (figs.
111 and 112). This feature, when present, gives a true bilateral
symmetry to the cleavage pattern of the lower hemisphere; the
axis of this bilateral symmetry coincides with the axis of excen-

484 BERTRAM G. SMITH

tricity previously defined, hence lies in the sagittal plane of the
embryo. The excentric position of the vegetal pole within the
area occupied by the macromeres, and the bilateral character
of the cleavage in this region, are more marked in many eggs
taken inmediately after the beginning of gastrulation; these
features are usually better expressed than in the eggs shown in
figures 111 and 112, which were chosen because the distinctness
of the early cleavage furrows enabled them to be drawn with the
camera lucida. Schultze (’00, Taf. 11, fig. 12) has described a
similar bilaterality in the late cleavage of the lower hemisphere
of the frog’s egg.

The question of the possible relation of the excentric and bilat-
eral development of the lower hemisphere just described, to the
excentric development of the blastodise noted in previous stages,
will be discussed in a later paper.

In the living egg, the roof of the segmentation cavity, though
apparently thin, is not quite so translucent as in the preceding
stage. It is, however, noticeably more translucent on the side
toward the future blastopore, and on this side the transition to
the opaque yolk cells is more abrupt.

During the late stages of cleavage, a tendency toward fading
out of the cleavage furrows in the vicinity of the vegetal pole
has been noted. In some individual cases this process has gone
so far that the earlier cleavage furrows are lost. to view, even
when searched for with the binocular microscope. This tendency
may be interpreted as due to a difficulty in sustaining the holo-
blastic method of cleavage in an egg so heavily laden with yolk.
In the corresponding stages of Necturus, this tendency is even
more marked. My study of the cleavage pattern of the lower
hemisphere of the late segmentation stages of both Cryptobran-
chus and Necturus has been confined to preserved material, but
Professor Dean informs me that he has noticed this merging of
lower blastomeres in the late segmentation stages of the living
eggs of Necturus. My Necturus material is not so favorable
as the egg of Cryptobranchus for the study of the cleavage pattern
in this stage, so a detailed comparison. will not be attempted.

In the very late cleavage stages of Cryptobranchus japonicus,
Ishikawa (’08 and ’09) describes a shallow furrow (‘Scheidewand-

EMBRYOLOGY OF CRYPTOBRANCHUS 485

furche’ or ‘septal furrow’) bounding the posterior margin of the
roof of the segmentation cavity, parallel to the future blastopore.
Such a furrow does not normally occur in this stage of Crypto-
branchus allegheniensis, but a similar furrow makes its appear-
ance shortly after the beginning of gastrulation.

B. Summary

The cleavage is holoblastic, but with great inequality in the
size of the micromeres as compared with the macromeres.

Biradial symmetry of the cleavage pattern begins with the
third cleavage stage. In the upper hemisphere, as a consequence
of the shifting of the micromeres, this biradial symmetry is lost
at about the fifth cleavage stage. In the lower hemisphere,
because of the slow cleavage and stability of the macromeres,
it persists until after the beginning of gastrulation, and in some
eggs enables one to distinguish first and second cleavage furrows
even after the blastopore has appeared.

An excentrically situated area of unusually small micromeres
is apparent in surface views of most eggs in Stages 5 to 8 inclu-
sive; the cleavage pattern of such eggs thus possesses bilateral
symmetry.

In the sixth cleavage stage there is more rapid cell division
in the marginal region of the micromeres than in the region of
the animal pole. The later expression of this tendency is almost
wholly internal.

In Stages 7 and 8 some of the superficially smaller micromeres
are becoming submerged through a process of immigration.

In the later cleavage stages there is a tendency for the cleavage
furrows to become less distinct than formerly in the region of
the vegetal pole, indicating a difficulty in sustaining the holo-
blastic character of the cleavage in an egg so heavily laden with
yolk; the same tendency is observed in Necturus.

In late segmentation immediately preceding gastrulation the
cleavage pattern enables one to predict the side on which the
blastopore is to appear; the egg undergoes a slight rotation on
a horizontal axis at right angles to the median plane.

486 BERTRAM G. SMITH
VIII. GASTRULATION AND EARLY FORMATION OF THE EMBRYO

A. Description by stages

Stage 11: (figs. 113 to 137 and 220 to 222). This stage extends
from the time of the first appearance of the blastopore as a short
horizontal groove until its ends meet to form a complete circle.
In eggs kept in their natural environment, gastrulation begins
about seven days after fertilization and two days after the begin-
ning of Stage 10.

MH

ed
nena

0 08 p

ee5eee iz
AD

113 a

Figs. 118 and 114 Lower hemispheres of two eggs of Cryptobranchus alle-
gheniensis in an early gastrula stage, showing cleavage furrows. The vertical
axis as determined by gravity lies at the center of each figure; the vegetal pole,
at the intersection of the first two cleavage furrows, is about 7 degrees above the
vertical pole. In figure 113 the first cleavage furrow lies approximately in the
median plane of the gastrula; in figure 114 it is at right angles to this plane. Cam-
era drawings, finished under the binocular, from preserved material. X 8.

The blastopore is first distinguished as a shallow irregular and
broken horizontal groove two or three millimeters in length.
lying about 15 degrees below the equator. It occurs at the upper
limit of transitional cells between micromeres and macromeres,
and its immediate site is distinguished by a rather abrupt demar-
cation between micromeres and distinctly larger transitional
cells. The groove is started, not by a lining-up of cells and the
union of cleavage furrows, as described by Eycleshymer (’95)
for Amblystoma, but by the sinking-in of groups of entire cells

EMBRYOLOGY OF CRYPTOBRANCHUS : 487

at intervals along a narrow zone several cells in width; hence
from its very beginning the process is not a splitting-apart of
cells, but invagination. The groove soon becomes continuous
and deepens by the inturning of cells along both margins.

After the groove has reached a length of three millimeters or
more, the process of invagination becomes accompanied by one
of overgrowth or epiboly: the dorsal lip grows slowly down over

oe

9

SED,
0 8

ries,
(YAN

Uy)
8
Ly
‘A
WY

oe
ity

oo
oe

He
te

BY
voit

ARR 28,
96102108). “
Nepean esate

8,
3

+
WH
QT (> Mg AIL AT el cg
VEO v9. ate ay

Fig. 115 Lower hemisphere of a gastrula of Cryptobranchus allegheniensis,
in aslightly later stage than the preceding, showing the lining-up of the cells within
the horns of the blastopore. Freehand drawing from a photograph of preserved
material. X 10. :

the cells transitional between micromeres and macromeres (figs.
113 to 115). As shown in figure 115, the transitional cells just
within the horns of the blastopore are elongated as if compressed ;
here the cells line up and lengthen out at right angles to a line
connecting the horns of the blastopore.

After the blastoporic groove has attained the form of a semi-
circle (fig. 133), a zone of rather abrupt demarcation between
micromeres and transitional cells completes the circle begun by
the crescentic blastopore; this zone marks the site of the ventral
lip of the blastopore. A little later, the blastoporic groove ex-
tends rapidly along this line of demarcation, becoming an almost

48s. BERTRAM G. SMITH

perfect circle, and enclosing a broad horseshoe-shaped band of
transitional cells, within which lie the macromeres (fig. 138).

It has already been noted that a slight rotation of the egg on
a horizontal axis has taken place in Stage 10, so that it is now
necessary to distinguish between the morphological axis of the
egg and the vertical axis determined by gravity, since the two
no longer coincide. In the study of gastrulation this rotation
must be taken into account, and some means must be found for
measuring it. In studying the morphological features of the
egg (position of blastopore, etc.) in their relation to the vertical
axis, two general methods have been used: (a) the living egg,
placed in a small vial of water, has been studied in side view and
measurements made against a protractor used as a background;
and (b) for accurately locating the vertical axis I have devised
the following apparatus: a glass disc such as is used for an ocular
micrometer was marked in the center with a small dot; a circle
with a radius of 4 mm. was then drawn about this dot as a center.
When this dise is placed in the eyepiece of a low-power micro-
scope used in studying the eggs, the circle is just large enough
to enclose the image of anegg. When the egg, immersed in water
in a watch glass, is accurately placed so that its image is enclosed
by the circle, the dot lies over the upper pole of the vertical axis;
this point is then marked by puncturing with a hot needle. The
operation was first tried on living eggs, which were then fixed
for further study; but since with the living egg even a small
puncture in this region usually causes the embryo to collapse
during the subsequent process of fixation, in general this method
is less satisfactory with living than with preserved material.
On account of the usually perfect preservation of the form of
the egg by the fixing fluid employed, the results obtained by
marking preserved material seem fairly trustworthy, especially
when spherical eggs are selected and a large number used. The
position of the upper vertical pole, thus marked, gives a reference
point for correlating the morphological features of the egg with
the vertical axis; the measurements were made by means of
camera drawings.

EMBRYOLOGY OF CRYPTOBRANCHUS 489

In making these measurements, it is especially necessary to
guard against using eggs with an unusually large yolk plug,
since this is one of the commonest abnormalities. Moreover,
even in perfectly normal eggs there is considerable variation in
the position of the blastopore, so that a large number of eggs
must be studied and averages taken. The results obtained by
the two methods agree closely.

Since the blastopore, at the time of its first appearance is: only
about 15 degrees below the horizontal equator and approximately
parallel to it, the blastopore at first forms an are of an imaginary
circle whose diameter, measured along a meridian of the egg,
is about 150 degrees. At the time when the blastopore has
reached the form of a semicicle, this diameter measures about
125 degrees; when the blastopore has become a complete circle
the average diameter, in normal embryos, is only 94 degrees.
Therefore the crescentic blastopore forms an arc of a circle of
steadily diminishing diameter; the lips of the blastopore, and
the entire germ ring (to be described in a later paper), contract
as they progress slowly downward over the egg.

In preserved material the cleavage pattern of the macromeres
is still fairly well defined; by means of careful study with a binoc-
- ular microscope it is usually possible to distinguish first and
second cleavage furrows (figs. 113 and 114). This enables a
direct comparison to be made between the direction of the first
cleavage furrow and the median plane of the gastrula; this point
will be discussed in a later paper. The identification of early
cleavage furrows in this region is furthermore of importance in
enabling one to determine the position of the vegetal pole, since
this is located at the intersection of the first two cleavage furrows.
Measurements show that at the time when the blastopore is first
clearly established, the vegetal pole lies, on the average, 68
degrees below it, and 7 degrees above the lower pole of the verti-
cal axis. At the time when the blastopore becomes a semicircle,
the vegetal pole lies only 32 degrees below its dorsal lip; when
the blastopore first becomes a complete circle the vegetal pole
lies only 26 degrees below the dorsal margin of the yolk plug.
During this time continued rotation of the egg has brought its

JOURNAL, OF MORPHOLOGY, VOL. 23, NO. 3

EMBRYOLOGY OF CRYPTOBRANCHUS 491

morphological axis to an angle of 44 degrees from the vertical;
the ventral lip of the blastopore now lies about 24 degrees beyond
the lower pole of the vertical axis. These changes are set forth
diagrammatically in figures 134 to 137.

Two quantitative results of considerable importance are
brought to light through the study of these data: (a) the dorsal
lip of the blastopore has grown downward over the yolk cells for
a distance of about 42 degrees; (b) the egg has rotated in the
opposite direction about 37 degrees from its position at the begin-
ning of gastrulation, making a total rotation of 44 degrees. At
first, overgrowth is more rapid than rotation; at the time when
the blastopore has reached the form of a semicircle its dorsal lip
is 43 degrees below the horizontal equator. Later, rotation is
more rapid than overgrowth, and at the time when the blastopore
has become a complete circle its dorsal lip has been carried back
to a position 20 degrees below the horizontal equator, only 5
degrees lower than its original position in space.

The changes in the upper hemisphere visible from the surface
during the establishment of the blastopore are remarkable, since
they afford clues to many important processes within. In the
living egg especially, because of the translucent character of
the upper hemisphere, one is able to get total views of many
phases of gastrulation, such as could not be obtained from serial
sections except by means of reconstructions. Except where
otherwise noted, the following description is based on the study
of the living egg.

At the very beginning of the process of gastrulation, the nearly
transparent roof of the segmentation cavity is of quite uniform

Figs. 116 to12l Stage 11 (gastrula) of Cryptobranchus allegheniensis. Cam-
era drawings from preserved material. In all the figures, the upper vertical pole
as determined by gravity lies toward the top of the page; blp., dorsal lip of the
blastopore; f., fenestra (roof of the blastocoele differentiated into a window-like
structure); s.f., septal furrow. X 7}.

Fig. 116 Lateral view of an early gastrula stage. The sharp differentiation
of the fenestra is rather precocious in this egg.

Figs. 117 to 119 Lateral views of a characteristic series of later embryos.

Figs. 120 and 121 - Antero-ventral views showing stages in the disappearance
of the fenestra. Figure 120 is from the egg drawn for figure 119.

492 BERTRAM G. SMITH

extent about the animal pole as a center, covering an area about
140 degrees in diameter. As gastrulation advances this clear
area becomes encroached upon at its posterior margin (figs. 122
and 123) by the extension of the opaque material. Meanwhile
the boundary of the roof of the blastocoele becomes more sharply
defined; before the upgrowth of the postero-dorsal opaque region
has reached the animal pole the margin of the blastocoele roof
is usually bounded by a sharply defined furrow, the ‘septal fur-
row’ of Ishikawa (see below)—a characteristic and almost unique
feature of the gastrulation of Cryptobranchus. The precise
stage at which this furrow appears varies considerably in different
eggs; figure 116 shows a case of unusually early appearance,
figures 117, 125 and 126 a stage in which it is usually well estab-
lished. Moreover, the distinctness of this groove varies greatly,
particularly in eggs of different spawnings; in some lots of eggs
the groove is established early and is very sharply marked, while
in occasional lots of eggs it is almost absent.

The septal furrow appears first at the posterior margin of the
roof of the segmentation cavity, then extends gradually around
to its anterior margin; in its appearance and manner of extension

Figs. 122 to 133 Stage 11 (gastrula) of Cryptobranchus allegheniensis. Free-
hand drawings of the living eggs, viewed by both transmitted and reflected light;
the proportions of the various parts are checked by comparison with camera draw-
ings of preserved material. The drawings are oriented with respect to the
vertical axis determined by gravity. The roof of the segmentation cavity is
nearly transparent; the roof of the gastrocoele is quite translucent, or slightly
opaque in the regions containing mesoderm; heavily yolk-laden regions are
decidedly opaque; bc., roof of blastocoele; blp., dorsal lip of the blastopore; f.,
fenestra (roof of the blastocoele differentiated into a roof-like structure); gc.,
gastrocoele; m., region containing mesoderm. X 5.

Figs, 122 to 124 Upper hemisphere, lateral view and lower hemisphere of an
egg in the beginning gastrula stage.

Figs. 125 and 126 Upper hemisphere and lateral view of an egg a little later
than the preceding.

Figs. 127 to 129 Postero-dorsal view, upper hemisphere and lateral view of
an egg slightly later than the preceding.

Fig. 130 Upper hemisphere of a slightly later egg.

Figs. 130 to 188. Upper hemisphere, postero-dorsal view and lower hemisphere
of an egg near the close of Stage 11 (shortly before the appearance of the neural
groove).

EMBRYOLOGY OF CRYPTOBRANCHUS

494 BERTRAM G. SMITH

it somewhat resembles a blastopore (fig. 221). Later, the groove
becomes faint at its posterior margin, very pronounced at its
antero-ventral margin (fig. 222). The area enclosed by the groove
diminishes in size with its forward movement; it also becomes
almost transparent. Since throughout the remainder of its
history this area strikingly resembles a window, I shall refer to
it as the fenestra.

In material fixed in a modification of the bichromate-acetic-
formalin mixture (see Smith 712, Section III, Solution B) con-
taining twice the usual amount of potassium bichromate, the
fenestra is cut up into small polygonal areas separated by furrows
that greatly resemble cleavage furrows (figs. 116 to 120 and 222;
ef. Ishikawa, ’08 and ’09). These polygonal areas do not repre-
sent single cells; each comprises a group of several cells. The
phenomenon is not entirely an artifact, since it often appears,
though faintly, in the living egg. By this method of fixation
the septal furrow is likewise accentuated. d

Before describing the further history of the fenestra it is desir-
able to direct attention to some other changes in the upper hemis-
phere as observed in the living egg.

About the time that the fenestra becomes limited to the ante-
rior half of the upper hemisphere by the upgrowth of the poste-
rior margin of the opaque region, a translucent area, the roof
of the gastrocoele, appears in this region of upgrowth (figs. 125
and 126). This translucent area is at first crescent-shaped; it
is separated from the more transparent fenestra by an opaque
band which is the outward expression of the septum separating
the gastrocoele from the blastocoele.

As soon as the septum has advanced into the anterior half of
the upper hemisphere, the translucency of the roof of the gas-
trocoele extends backward almost to the blastopore—evidently
by the deepening of the gastrocoele in this region, admitting
light. Meanwhile each postero-lateral margin of this region
becomes bordered with a faint band of a slightly more opaque
character—an effect due largely to the early mesoderm (figs.
127 to 129), though the entoderm is also concerned in producing
it.

EMBRYOLOGY OF CRYPTOBRANCHUS 495

Later changes are concerned with the forward, or rather ven-
trad, progress of the septum and the increase in the extent of
the translucent roof of the gastrocoele, with a correlated ventrad
movement of the fenestra and a diminution of its area; there
is a slight increase in the extent and opacity of the mesoderm
(figs. 118 to 120 and 1380 to 132). The fenestra finally closes
just below the horizontal equator on the ventral side of the egg
(fig. 121). The changes in the position and extent of the fenestra
are shown diagrammatically in figures 134 to 137.

The foregoing detailed account of the progress of the septum
as viewed from the exterior in the living egg of Cryptobranchus
clears up whatever doubt may exist as to the significance of the
‘shadowy area’ described in the gastrula of Spelerpes by Goodale
(11) and confirms his suggestion as to the nature of this area.

Ishikawa (’08 and ’09) describes in the early gastrula of Cryp-
tobranchus japonicus a furrow bounding the roof of the blasto-
coele at its posterior margin, which he calls the ‘Scheidewand-
furche’ or ‘septal furrow.’ As compared with the furrow of
similar nature described above for C. allegheniensis, it is earlier
in. making its appearance, since it antedates the blastopore.
The area later enclosed by this furrow has been named by Ishikawa
the ‘Keimhohlensegment’ or ‘blastocoele-segment’; judging from
his figures its later history is much the same as that of the cor-
responding structure, which I have preferred to call the ‘fenestra,’
in. Cryptobranchus allegheniensis.

The only mention of similar structures which I can find in the
literature on other forms is a description by Hatta (’07) of a
groove which he calls the ‘boundary groove’ in the gastrula of
Petromyzon. As compared with the septal furrow of Crypto-
branchus this groove is greatly exaggerated in Petromyzon, con-
stricting the egg so that in some cases it assumes an hour-glass
form. :

As suggested by Hatta, the boundary groove or septal furrow
is passive in origin, and a product of gastrulation. Similar con-
ditions have produced it in two such widely separated forms as
Cryptobranchus and Petromyzon; in each case the egg contains
considerable yolk, and the roof of the blastocoele is unusually

496 BERTRAM G. SMITH

136 137

Fig. 134 Diagram of an egg of Cryptobranchus allegheniensis in the beginning
gastrula stage viewed from the lateral aspect, showing average amount of rotation,
and the positions of the beginning blastopore and the septal furrow; blp., blasto-
| pore; m, morphological axis; s.f., septal furrow; v, vertical axis determined by
gravity; v.p., vegetal pole.

Fig. 185 Similar diagram of a gastrula at the time when the blastopore reaches
the form of a semicircle. Lettering as before.

Fig. 136 Similar diagram of a gastrula at the time when the blastopore first
becomes a complete circle. Lettering as before.

Fig. 187 Combination of the preceding diagrams. The egg is shown in the
position assumed at the close of the period considered. The black band indicates
the amount of overgrowth of the dorsal lip of the blastopore (42 degrees); blp.,
blp’, and blp?, mark the successive positions of the dorsal lip of the blastopore;
s.f. and s.f.’, successive positions of the septal furrow; f, position of the vanishing
fenestra. Other lettering as in the preceding figures. :

EMBRYOLOGY OF CRYPTOBRANCHUS 497

thin. As stated by Ishikawa, the polygonal figures formed on
the surface of the blastocoele-segment (fenestra) are perhaps due
to the pressure which produces the gradual diminution of its
area; but the cells of the fenestra are not compacted together to
any considerable extent, since the gastrocoele roof and wall merely
grow under them.

In view of the later history, it is evident even from surface
views that in the stage shown in figures 116 and 123 the forma-
tive material for the embryo is mainly concentrated in the equa-
torial region as a broad band or zone of cells, wider in its poste-
rior portion. As will be shown in the description of the internal
structure, this equatorial zone as distinguished in surface views
is only roughly comparable to the germ ring of fishes.

My material is lacking for the study of the early gastrula
stages of Necturus; late gastrula stages differ from Cryptobran-
chus chiefly in that the blastopore earlier becomes a complete
circle. In Spelerpes, according to Goodale (’11), no ventral lip
is formed to the blastopore.. As compared with urodele and
anuran eggs in general, the blastopore of Cryptobranchus is late
in closing; in its mode of gastrulation the egg of Cryptobranchus
approaches more nearly the type observed in meroblastic eggs.

Stage 12: (figs. 138 to 150 and 223 to 225). This stage is char-
acterized by the presence of the neural groove and is terminated
by the appearance of the neural folds. The neural groove appears
about three days after the beginning of gastrulation.

At the time of the earliest indications of the neural groove,
the blastopore has just become a complete circle. At the close
of the preceding stage it had a diameter of about 94 degrees;
it now rapidly becomes smaller, so that before the appearance
of the neural folds its diameter averages about 26 degrees (figs.
138 to.145).

During the early part of this stage the yolk plug is characterized
by a broad crescent-shaped or horseshoe-shaped area of smaller
cells lying ventrad and laterad to the macromeres (figs. 138 and
139). Along the lateral line of transition between the macromeres
and these smaller cells, the cells appear compressed and exhibit
a tendency to line up and merge their cleavage furrows (figs.

498 BERTRAM G. SMITH

138 139

Figs. 138 and 139 Posterior views of embryos of Cryptobranchus allegheniensis
in Stage 12, showing the condition of the blastoporeé and the cleavage furrows of
the yolk plug. Camera drawings from preserved material. The embryos are
not accurately oriented with respect to the vertical axis determined by gravity.
ts

Fig. 138 Showing condition shortly after the appearance of the neural groove.

Fig. 139 A little later than the preceding.

138 to 142; cf. figs. 113 to 115). Toward the close of this stage,
the smaller cells become completely overgrown by the ventral
and lateral lips of the blastopore, leaving only the larger ones
exposed (fig. 144); evidently overgrowth is now proceeding more
rapidly at the ventral than at the dorsal lip of the blastopore.
At the close of the stage the greatly reduced yolk plug lies entirely
on the postero-dorsal side of the lower pole of the vertical axis.

During the earlier part of the stage under consideration the
closing fenestra often persists as a pit or small tract of distinct

Figs. 140 to 145 Dorsal views of embryos of Cryptobranchus allegheniensis
in Stage 12, showing a series of stages in the development of the neural groove.
Camera drawings from preserved material. The embryos are not oriented with
respect to the vertical axis determined by gravity. X 7.

Fig. 140 Showing earliest appearance of the neural groove.

Fig. 141 Slightly later than the preceding.

Fig. 142 Slightly later than the preceding, showing segmented neural groove.

Fig. 143 Slightly later, segmented neural groove. See also figure 225 from
the same embryo.

Fig. 144 Late neural groove.

Fig. 145 Showing the condition of the neural groove at the time of the first
faint indications of the neural folds.

499

‘

CRYPTOBRANCHL

rY OF

OG

EMBRYOL

14]

140

145

144

500 BERTRAM G. SMITH

furrows at the equator on the antero-ventral side of the egg.
It usually disappears by the time the neural groove is well estab-
lished.

In preserved material, the roof of the gastrocoele is consid-
erably paler than the remaining surface of the egg; during the
latter part of this stage the neural plate is usually differentiated
as a spatulate area extending from the dorsal lip of the blastopore
to a little distance in front of the upper pole of the vertical axis,
and distinguishable through the greater whiteness of its surface
(see especially figs. 223 and 224).

The dorsal lip of the blastopore is not a perfect arc of a circle,
but is somewhat incurved on each side of a forward-extending
notch in the median line (figs. 138 to 145).

At the time of its first appearance, the neural groove occurs
as a distinct furrow extending from the notch in the dorsal lip of
the blastopore forward in the median line for a distance of about
60 degrees; the anterior half is much broader and deeper than
the posterior half (fig. 140). In a slightly later stage, the neural
groove has extended to a total length of about 95 degrees but
is nowhere so deep as in the anterior half during the preceding
stage (fig. 141). It is now a rather shallow groove, narrow in
its posterior portion, wider and more broken by occasional deeper
depressions or fissures in its middle and anterior parts. These
early transverse furrows do not occur at very regular intervals,
and are probably only incidental to the process of infolding of
the tissues.

A little later, the neural groove becomes decidedly deeper in
its middle portion (fig. 142). The change is not uniform through-
out this region, but instead there is a series of three or four large
pits or depressions at fairly regular intervals, giving a segmented
appearance to the groove. Sometimes this segmented condition
is very marked; it has been repeatedly observed in living mate-
rial. Gradually thesegmented region, though less sharply marked
becomes more extensive than before (fig. 148); it is best seen in
living material viewed by transmitted light, when the neural
groove appears made up of a regular succession of alternate
light and dark areas. Shortly before the appearance of the neural

EMBRYOLOGY OF CRYPTOBRANCHUS 501

folds, the neural groove becomes conspicuous in its anterior as
well as its middle portion by the broadening and deepening of
the former region; at the same time the posterior end becomes
fainter (fig. 144).

At the time of the first faint indications of the neural folds,
the neural groove is both broad and deep throughout its entire
length but especially in its anterior portion (fig. 145). During
these later changes in the breadth and depth of the neural groove
there has been very little increase in length; at the close of the
period considered it has a length of about 105 degrees and extends
from the dorsal lip of the blastopore nearly to the upper pole of
the vertical axis.

According to Griggs (710), in Amblystoma the first groove to
appear in the median line of the neural plate is not the neural
groove, properly speaking; this appears later on the same site.
There first appears a ‘posterior germinal depression’ which does
not reach the blastopore; a little later an ‘anterior germinal de-
pression’ is formed, which is discontinuous with the earlier groove.
In a later stage both give place to the neural groove which is
a different structure on the same site.

In Cryptobranchus the earliest groove to appear in the median
line of the neural plate extends forward without break from the
dorsal lip of the blastopore. The marked depression shown in
figure 140 probably corresponds to the ‘posterior germinal depres-
sion’ in Amblystoma; the later depression in the anterior portion
of the neural groove perhaps corresponds to the ‘anterior germinal
depression,’ but it is at no time sharply separated from the
remainder of the groove. The later history of the groove will
be given in the following stages and should be consulted in this
connection; but it may here be stated that after a careful study
of both surface views and serial sections I have come to the con-
clusion that the differences in the grooves appearing early and
late in the median line of the neural plate of Cryptobranchus
are differences in degree, not in kind, hence I have used the term
‘neural groove’ throughout.

In living material the embryo may be viewed by transmitted
light. During the early part of this stage (figs. 146 and 147)

502 BERTRAM G. SMITH

the broad lateral bands lying in the posterior part of the egg at
some distance from the median line are more marked than in
the preceding stage. The study of sections shows that they are
due to the combined optical effect of the mesoderm and an
unusually thick region of the entoderm. The neural groove is par-
ticularly translucent. The greatly reduced blastocoele persists
in the region of the equator on the antero-ventral side of the egg;
the center of its external wall is marked by a pit, the vestige of
the fenestra. During the latter part of Stage 12 (fig. 148) the

146 147
Figs. 146 and 147 A living egg of Cryptobranchus allegheniensis in an early
neural groove stage, viewed so far as possible by transmitted light. Figure 146
shows the upper hemisphere, figure 147 a postero-dorsal view. 7.

lateral bands are obscured by the thickening of the neural plate;
in the central portion of the neural groove there usually appear
a series of translucent pits arranged at regular intervals. The
pit marking the site of the closing fenestra has disappeared, but
there usually remains a translucent area indicating a vestige of
the blastocoele; this area is often imperfectly separated from
the translucent roof of the gastrocoele.

In this stage it is usually impossible to identify cleavage furrows
in the yolk plug, but the stability of the larger cells and the fact
that they remain longest exposed afford a means of locating
approximately the vegetal pole. We have seen that in the pre-
ceding stage the vegetal pole was situated a little above the center
of the area of largest macromeres; at the close of Stage 12 these

EMBRYOLOGY OF CRYPTOBRANCHUS 503

largest macromeres are the only ones exposed, and we may feel
quite sure that the vegetal pole lies in their midst and probably
very near the center of the greatly diminished yolk plug. The
morphological axis of the egg is thus approximately determined
by a line passing from the center of the yolk plug through the
center of the egg, and at the close of Stage 12 this axis makes an
angle of 52 degrees with the vertical—showing that rotation has
proceeded 8 degrees further than in the preceding stage (fig. 149).

Fig. 148 Upper hemisphere of a living egg of Cryptobranchus allegheniensis
viewed so far as possible by transmitted light, shortly before the appearance of
the neural folds. X 9.

It will be recalled that the dorsal lip of the blastopore first
appears, on the average, 68 degrees above the vegetal pole (fig.
134), and that the ventral lip is first formed about 68 degrees
from the vegetal pole on the opposite side (fig. 136). Hence
the dorsal and ventral lips are formed respectively at approxi-
mately equal distances from the vegetal pole, though the ventral
lip is formed much later than the dorsal. We may now compute
the amounts of overgrowth of the dorsal and the ventral lips
respectively: at the close of Stage 12 the dorsal lip has overgrown
the yolk for an average distance of 55 degrees, and the ventral
lip has advanced toward it through an are of about 55 degrees;

504 BERTRAM G. SMITH

149 150

hig. 149 Diagram of an egg of Cryptobranchus allegheniensis at the close of
Stage 12, showing the amount of rotation and the position of the yolk plug; m,
morphological axis; v, vertical axis determined by gravity; y.p., yolk plug. The
cross indicates the position of the anterior end of the neural groove.

Fig. 150 Combination of figure 144 with some features of figure 137, showing
successive positions of the blastopore; blp and blp’ respectively indicate early
and late positions of the dorsal lip of the blastopore. The black band at. the right
of the figure indicates the amount of overgrowth (55 degrees) of the dorsal lip of
the blastopore; the dotted line indicates the amount of overgrowth (55 degrees)
of the ventral lip.

that is to say, the distances are approximately equal (fig. 150).
By far the greater amount of overgrowth of the dorsal lip occurred
during the preceding stage, hence it is clear that during the pres-
ent stage overgrowth is taking place much more rapidly at the
ventral than at the dorsal lip of the blastopore.

Stage 13: (figs. 151 to 164 and 226 to 228). The most conspic-
uous changes during this stage are those concerned with the
formation of.the neural folds and the segmentation of the neural
plate. The neural folds begin to form about one and one-half
days after the appearance of the neural groove. During the
progress of this stage the neural groove becomes most conspic-
uous in an anterior and a posterior portion, separated by a middle
region. in which it is comparatively faint (see especially figs. 151
to 155).

The early stages in the formation of the neural folds are shown
in figures 151 to 158 and need no further description, save to men-

EMBRYOLOGY OF CRYPTOBRANCHUS 905

155 156 157 158

Figs. 151 to 158 Stage 13. A series of embryos of Cryptobranchus alleghe-
niensis showing early stages in the formation of the neural folds and the segmenta-
tion of the neural plate. Camera drawings, finished under the binocular, from
preserved material. I, II and III indicate the earliest transverse grooves; 1,
2,3, grooves appearing a little later, numbered consecutively and not in the order
of appearance. X 6.

tion that the surface just outside of the neural folds becomes
very much roughened and traversed by fissures parallel to the
folds, indicating stresses and the rapid shifting of material.
During the later part of this stage a pair of less conspicuous
transverse folds appear lateral to the anterior end of the neural
plate (fig. 158); the sivnificance of these folds has not yet been
determined with certainty (but see Stage 15).

The first transverse groove to cross the neural plate is shown
in figure 152. A little later, two transverse grooves appear in
rapid succession posterior to it. These first three transverse

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

506 BERTRAM G. SMITH

grooves are equidistant, and so distinct that they may readily
be seen with the naked eye; since they regularly appear in the
same order and position in different embryos, and persist through-
out the further history of the open neural plate, they serve as
trustworthy landmarks during the following stages. In the
figures they are numbered with Roman numerals. By following
their history through later stages they have been traced to the
region. of the medulla oblongata of the adult brain; consequently,
at least all that portion of the neural plate in front of Groove III
' belongs to the cephalic plate.

The early segmentation of the cephalic plate in front of Gros
I will now be considered. There first appears a transverse groove
dividing this region into two portions of which the posterior is
slightly the smaller (figs. 155 and 156); the anterior of these
areas is then crossed by two more grooves (figs. 157 and 158),
while the posterior area is for the present doubtfully segmented.
The smaller transverse grooves occurring in various parts of the
cephalic plate are irregular in position and probably are of no
segmental value; most of them disappear in later stages. Those
grooves in front of Groove I which are regarded as of metameric
value are numbered with Arabic numerals, consecutively and
without regard to the order of appearance.

The question. naturally arises whether these early transverse
divisions of the cephalic plate are neural in origin or secondarily
produced by the segmentation of the underlying mesoderm.
This question has not yet been thoroughly investigated by the
study of sections, but the results of a preliminary examination
favor the idea that in front of Groove I at least, they are primarily
neural structures; the mesoderm, particularly in front of Groove
I, is at this time quite thin as compared with the neural plate,
and hardly capable of producing the modifications of the latter
layer.

Since Grooves I, II, III, ete. (see also Stage 14) are produced
in regular order from before backward there is ground for sus-
picion that they are intimately connected with the formation of
the mesoblastic somites. In view of the fact that the segmenta-
tion of the region immediately in front of Groove I is late in

EMBRYOLOGY OF CRYPTOBRANCHUS 507

appearing and seldom clearly expressed (Stage 14), we must be
on our guard against a possible discontinuity or difference in
kind between the segmentation of the anterior and the posterior
regions of the cephalic plate. These points can be settled only
by a careful study of sections of eggs that have first been described
externally; but from surface views alone we are justified in claim-
ing that we have in the open cephalic plate transverse divisions
which may be homologized in different embryos, and which are
probably of true metameric value; hence they may be of use in

solving the vexed problem of the segmentation of the vertebrate
head.

Fig. 159 A living embryo of Cryptobranchus allegheniensis in the early part
of Stage 13, viewed in direct sunlight, and so far as possible by transmitted light.
From a freehand sketch of the upper hemisphere. X 10.

A pair of depressions just within the neural folds near the ante-
rior end of the cephalic plate probably indicate the anlage of the
optic vesicles (cf. Eycleshymer ’95; Locy, 795).

Some features of this stage are best brought out by the study
of living material; for-this purpose embryos have been examined
in direct sunlight. As shown in figure 159 a transverse opaque
band early appears directly in front of the neural plate in the
median region; in position and appearance it reminds one of the
ectamnion of the chick (Lillie ’08, pp. 138 and 139). The neural

508 BERTRAM G. SMITH

folds are conspicuous at an earlier stage in living than in preserved
material. In embryos later than the one figured, transverse
furrows in the neural plate appear as described in preserved
material.

During Stage 13 the blastopore nearly closes, then makes
little advance in this respect during the next two stages. Varia-
tions in the degree of reduction of the blastopore during these
three stages are so great that this structure cannot be used as
a character for classifying embryos into stages.

As shown in figures 160 to 162, during Stage 13 the blastopore
changes from a diamond shape to that of an anchor; the forward-

160 161 162

Figs. 160 to 162 A series of embryos of Cryptobranchus allegheniensis in Stage
13, showing changes in the size and form of the late blastopore. Camera drawings
from preserved material. X 5.

projecting part is derived through an exaggeration of the notch
previously noted in the dorsal lip of the blastopore. The lappets
lying on each side of this median notch of the blastopore are
continuous with the neural folds; thréugh their apposition the
dorsal part of the yolk plug becomes closed over. Thus the
extreme posterior end of the embryo is undoubtedly formed by
a process of concrescence. As shown in later stages, the ventral
part of the blastopore ‘becomes reduced to a transverse slit (figs.
177 and 178); during this process the yolk plug usually becomes
entirely withdrawn into the egg, but a small mass of yolk some-
times persists at the surface. The late history of the blastopore
is much the same in Cryptobranchus japonicus, as described by
Ishikawa (’08).

EMBRYOLOGY OF CRYPTOBRANCHUS 509

At the close of Stage 13 the neural grove has reached a length
of about 124 degrees; its anterior end usually lies quite accurately
at the upper vertical pole, while the neural folds extend about
16 degrees in front of it. We have seen that the closing blasto-
pore marks the approximate position of the vegetal pole; this
pole has now rotated a total distance of 56 degrees from the ver-

tical axis.

-_

163 164

Fig. 163 Diagram of an embryo of Cryptobranchus allegheniensis at the close
of Stage 13, showing the position of the neural plate and neural folds with refer-
ence to the morphological and the vertical axes; m, morphological axis; v, vertical
axis.

Fig. 164 Diagram showing the position of the embryonic body of Crypto-
branchus allegheniensis, and illustrating some features of embryo-formation;
a to b (72 degrees), portion of the embryo formed in situ; 6 to c (60 degrees), por-
tion formed by overgrowth of the dorsal lip of the blastopore, with the possibility
of concrescence; c to d (roughly estimated at 16 degrees), portion undoubtedly
formed by concrescence; d to e (60 degrees), distance traveled by the ventral lip
of the blastopore. Other lettering as in the preceding figure.

We now have sufficient data for a statement of the position
of the embryonic body on the egg, and for pointing out certain
features of its mode of formation (figs. 163 and 164). About
72 degrees of the anterior end of the embryo is formed in situ.
About 60 degrees is formed in connection with the overgrowth
of the dorsal lip of the blastopore; in this case there is the possi-
bility of concrescence through the apposition of material on each

510 BERTRAM G. SMITH

side of the median notch, which may be shifted toward the median
line during the process of overgrowth. This point can be definitely
settled only by experiment; but in the absence of experimental
data we can say that there is no positive evidence of such a proc-
ess taking place, while certain considerations weigh against it.
For in certain observed cases rapid shifting of material is accom-
panied by a roughening of the surface with the formation of par-
allel fissures, as in the region just outside of the neural folds
during their formation and progress toward the median line.
There is an entire absence of any such feature in the dorsal lip
of the blastopore.

A region at the posterior end of the embryo, which is roughly
estimated at 16 degrees, is formed through the concrescence of
the lateral and ventral lips of the blastopore. A part of this
material has been brought through a distance of 60 degrees by
the overgrowth of the ventral lip of the blastopore; it will be
‘observed that this distance equals that of the overgrowth of the
dorsal lip of the blastopore.

At the close of Stage 13, when the embryonic body is for the
first time clearly indicated, it has a total length of about 148
degrees. The posterior end is formed around the vegetal pole;
the anterior end lies about 40 degrees from the animal pole.
Hence the statement made in Part I (Smith 712) to the effect
that the axis of polarity of the late ovarian egg defines the prin-
cipal axis of the embryo is not quite accurate; but the embryo is .
formed almost wholly in a hemisphere of the egg lying to one
side of the axis of polarity. A review of its history shows that
the embryo is formed almost entirely out of material derived
from a band of cells lying in the equatorial region of the late
blastula and early gastrula, and that this band of cells is narrow
on the ventral, broad on the dorsal side of the egg (figs. 116 and
123).

Goodale (’11), after reviewing the literature of the subject
in connection with his own work on Spelerpes, concluded that

The amphibian embryo develops almost entirely in a vertical half

of the egg, the tail appearing near the lower pole, while the anterior
end of the body developsin greater or less degree in the upper hemisphere,

EMBRYOLOGY OF CRYPTOBRANCHUS o11

depending upon the particular species. The position of the head of
the embryo seems correlated with the length of the embryo, so that
the longer the embryo, the higher up on the egg it develops.

The terms ‘upper’ and ‘lower’ are evidently here used in the
sense of animal and vegetal, that is to say, with reference to those
points on the surface of the egg which were in the vertical axis
of the egg before it commenced to rotate; therefore the results
obtained with Cryptobranchus fall in line with the general state-
ment quoted. The results of Goadale on Spelerpes and my own
work on Cryptobranchus agree closely in locating the posterior
end of the embryo at the vegetal pole; it is worthy of note that
this conclusion was reached independently and by entirely differ-
ent methods.

Stage,14: (figs. 165 to 179 and 229 to 232). This stage is reached
about one day later than the beginning of Stage 13. Since at:
this time scarcely any two embryos agree in the rate of develop-
ment of homologous regions of the body, it is impossible in this
stage to make a close classification. In marking off this stage
from the one following, the principal character considered is
the approach of the neural folds toward the median line.

Figures 165 to 176 represent twelve embryos that illustrate
the principal changes in the antero-dorsal region during this
stage. It will be seen that there is a progressive addition of
transverse grooves posterior to the three that first appeared. In
front of Groove I the cephalic plate is traversed primarily by six
grooves; of these Groove 1, which was noted in the preceding
_ stage, has a very transitory existence and in most cases is lost
in Stage 14; likewise the median portion of Groove 2 has often
disappeared. Moreover in this or the following stage Groove
4 disappears, following a marked depression and perhaps sub-
mergence of the segment between it and Groove 3.

A significant relation exists between Grooves I, IT, III, ete. and
the intersomitic grooves which now appear just gutside the neural
folds; by an inspection of figures 165. to 176 it will be seen that
in all cases these are in direct apposition. Since the mesoblastice
somites are the most characteristically segmented structures of
the vertebrate body, it follows that the true segmental units of

173 174 175 “176

Figs. 165 to 176 Antero-dorsal views of embryos of Cryptobranchus alleghe-
niensis in Stage 14, showing especially the segmentation of the neural plate. Cam-
era drawings finished under the binocular, from preserved material. X 6.

The earliest transverse grooves to cross the neural plate are numbered with
Roman numerals in the order of appearance; in front of Groove I the transverse
grooves are numbered with Arabic numerals consecutively without regard to the
order of appearance. Figure 166 is from the embryo photographed for figures 229
and 230; figure 176 is from the embryo photographed for figures 231 and 232.

512

EMBRYOLOGY OF CRYPTOBRANCHUS 513

the open neural plate, in this region at least, are the divisions
between grooves—that is, the ridges rather than the depressions,
for the former are in line with the body somites. If, as appears
likely, there is continuity between the structures of the anterior
and posterior regions of the cephalic plate, then the rule may
be extended to include the entire neural plate. As thus defined,
there are seven segments—‘neuromeres’—in front of Groove I;
posterior to this groove an undetermined number of segments
also belong to the head.

In the early stages of the formation of the neural folds trans-
verse grooves are sometimes found in them, continuous with
the transverse grooves of the neural plate (see especially figs. 154,
165 and 172). In such cases the neural fold is marked by an
outer as well as an inner notch, both in lHne with the transverse
furrow of the neural plate. This condition is only temporary
and apparently it is transitional to a later phase in which the
inner notch grows at the expense of the outer one, until an outer
convexity of the fold appears opposite the inner concavity (see
especially fig. 168). In the region of the body somites these
outward flexures thus lie in line with the intersomitic grooves
as well as with the transverse grooves of the neural plate. This
condition is seldom so well expressed as in the embryo shown in
figure 168; the convolutions of the neural folds are often irregular
and bear no. definite relation to the segments. But it is fairly
certain that in all cases where the neural folds are well upraised
and flexures occur which are segmentally arranged, the outward
flexures lie opposite the transverse furrows and not opposite
the ridges between them. Moreover, in sagittal sections the
transverse grooves on the external surface of the neural plate
are found to correspond to ridges on the internal surface.

Those who have described segmental structures in the neural
folds or closed neural tube have, as a rule, accepted Orr’s (’87)
definition of the segmental units or neuromeres as outward flex-
ures of the neural folds. But'if the above considerations be well
founded, the true segments are to be sought rather in the seg-
ments between the transverse grooves of the neural plate, and
in the inward flexures of the neural folds. In other words neu-

514 BERTRAM G. SMITH

romeres are the transverse ridges on the inside rather than on
the outside of the brain. To a limited extent this view coincides
with that of Kupffer (’85 to ’93), who maintained that the true
neuromeres are the transverse divisions of the open neural plate
rather than the later appearing structures in the neural folds.
Most of the features of this stage thus far described have been
observed in living as well as in preserved material. The liter-
ature on the early development of the central nervous system
has recently been. reviewed by Griggs (’10) ; a more comprehensive

177 178 j 179

Figs. 177 to179 Camera outlines of embryos of Cryptobranchus allegheniensis
in Stage 14, drawn from preserved material. X 5.

Figs. 177 and 178 Posterior views showing late blastopore.

Fig. 179 Lateral view showing position of the embryonic body at the close of
Stage 14. The egg is shown in its natural position with respect to the vertical
axis, which passes in the plane of the paper parallel to its lateral margins. The
embryo proper has a total length of about 155 degrees. This figure and figure
166 are drawn from the same egg.

survey of the earlier work on the segmentation of the vertebrate
head is given by Loey (’95). Ishikawa (’08) has described seg-
mental divisions in the open neural plate of Cryptobranchus
japonicus. |
During this stage, if not already in the preceding stage, the
anterior or dorsal part of the blastopore becomes closed over,
while the ventral part persists as a transverse crescentic slit
(figs. 177 and 178). At the close of Stage 14 the embryo has .
increased slightly in length (fig. 179); it now extends over about
155 degrees of the surface of the egg. This increase in length

EMBRYOLOGY OF CRYPTOBRANCHUS Way

183 184 185

Figs. 180 to185 Antero-dorsal views of embryos of Cryptobranchus alleghe-
niensis in Stage 15. Camera drawings finished under binocular, from preserved
material. Figure 182 is drawn from the embryo photographed for figure 234;
figure 184 is drawn from the embryo photographed for figure 233. 6.

of the embryo involves a noticeable increase in the antero-poste-
rior dimension of some of the neuromeres.

Stage 15: (figs. 180 to 189; 233 and 234). This stage is reached
about eighteen hours later than the beginning of the preceding
stage.

In the following account, each neuromere is designated by
the number of the groove bounding it on the posterior side.
Neuromeres 1 and 2 have usually coalesced; neuromere 4 dis-
appears during this, if not in the preceding stage. More definite
swellings now occur in neuromeres 1, 2, 3 and 5; the region between
Grooves 5 and I is less clearly segmented and is usually somewhat
depressed. The outlines of the neural folds in the head region |

516 BERTRAM G. SMITH

now suggest the definitive primary divisions (forebrain, midbrain
and hindbrain) of the embryonic brain.

The various structures of the neural plate have not yet been
followed into the definitive divisions of the embryonie and adult
brain; but the preliminary examination of some later embryos
dissected by splitting them in the median line with a razor shows
that the transverse divisions in the neural plate persist for some
time after the closure of the neural folds. Neuromeres in the
closed neural tube arealso often apparent from the surface. Hence
it is easy to judge approximately concerning the fate of individ-
ual neuromeres of the cephalic plate, but to avoid possible error
it seems best to defer a definite statement until the internal his-
tory of the brain has been more carefully studied.

The pair of folds which in the preceding stages extended trans-
versely on each side of the cephalic plate now slant backward
(see especially fig. 185); the appearance, particularly in living
material, suggests that they are in some way concerned with
the origin of the vascular bands which in later stages extend
along each side of the yolk sac and give rise to the omphalo-
mesenteric or vitelline veins (fig. 192).

The transverse opacity in front of the neural plate is conspic-
uous in living material viewed by transmitted light (fig. 186),
but is not apparent in surface views of preserved material.

The anterior part of the blastopore is now normally closed
over, and the posterior or ventral part is reduced to a transverse
slit (figs. 188 and 189). Apparently the middle portion of this
transverse slit never becomes completely closed, but in later
stages persists as the anal or cloacal opening. ‘The embryonic
body has elongated so that it now extends over about half the
circumference of the egg (fig. 187).

For the study of transverse divisions in the open neural plate,
Necturus is not nearly so favorable as Cryptobranchus. In
Necturus the blastopore (figs. 268 to 279) closes much earlier
than in Cryptobranchus. Moreover in Necturus the closure
of the blastopore is often practically complete; in many specimens
preserved at the time of the closure of the neural folds, scarcely
more than a vestige of the blastopore is visible from the surface.

EMBRYOLOGY OF CRYPTOBRANCHUS one

Fig. 186 Antero-dorsal view of a living embryo of Cryptobranchus allegehe-
niensis in Stage 15, viewed mainly by transmitted light. From a freehand sketch.
x 10.

187 188 189

Fig. 187 Lateral view of an embryo of Cryptobranchus allegheniensis in Stage
15, showing the position of the embryonic body. The egg is shown in its natural
position with respect to the vertical axis which passes in the plane of the paper
parallel to its lateral margins. Camera drawing from preserved material. X 5.

Figs. 188 and 189 Posterior views of embryos of Cryptobranchus alleghenien-
sis in Stage 10, showing the form of the late blastopore. Camera drawings from
preserved material. X 5.

Figs. 183, 187 and 188 All drawn from the same embryo.

Figs. 185 and 189 Drawn from the same embryo.

518 BERTRAM G. SMITH

In these later stages, the blastopore is doubtless often indistin-
guishable in living material (figs. 274 to 276).

A triradiate form of the blastopore is not so frequently found
in Necturus; in Cryptobranchus japonicus (Ishikawa ’08), it often
occurs. A general resemblance may be noted between the blas-
topore of the urodeles cited and that of the dipnoans (Ceratodus,
Semon ’01; Protopterus and Lepidosiren, Kerr ’09).

B. Summary

Gastrulation involves a combination of the processes of invag-
ination or emboly, and overgrowth or epiboly.

During gastrulation the roof of the segmentation cavity be-
comes very thin, and is bounded superficially by a sharp furrow,
the ‘septal furrow’ of Ishikawa.

On account of the translucent character of certain parts of
the egg, many of the internal changes concerned with gastrulation
can be followed quite satisfactorily in living material.

For some time after the beginning of gastrulation, the vegetal
pole may be located through the intersection of the first and
second cleavage furrows.

During gastrulation and the formation of the neural groove
and neural folds the egg rotates on an axis at right angles to the
median plane so as to bring the morphological axis at an angle
of 56 degrees from the vertical.

The dorsal lip of the blastopore is formed about 68 degrees
above the vegetal pole; the ventral lip is formed much later at
an equal distance on the other side of the vegetal pole. Since
the closing blastopore lies approximately at the vegetal pole,
overgrowth proceeds through equal distances on the dorsal and
the ventral sides of the egg. During the early part of gastrula-
tion, before the ventral lip is formed, overgrowth takes place
rapidly and extensively at the dorsal lip; after the blastopore
has become a complete circle, overgrowth takes place very slowly
at the dorsal lip, very rapidly at the ventral lip.

_Not until after the neural folds are well formed is the yolk
plug completely overgrown; as compared with Necturus and

EMBRYOLOGY OF CRYPTOBRANCHUS 519

most amphibian eggs the blastopore is very late in closing. In
late stages the blastopore has the form of an anchor or an inverted
T; the posterior transverse portion remains longest as an open
slit, and the center of this transverse portion never completely
closes but persists as the anal or cloacal aperture.

The posterior end of the embryo forms approximately at the
vegetal pole. At the time when the neural folds are first formed
the embryo has a total length of about 148 degrees, hence its
anterior end does not reach the animal pole. About 72 degrees
of the anterior end of the embryo (nearly half its total length)
is formed in situ; about 60 degrees posterior to this is formed by
overgrowth with the possibility of conerescence. Only a very
small part at the posterior end, perhaps 16 degrees, is formed by
the meeting of the lateral and ventral lips of the blastopore; this
part is undoubtedly formed by concrescence.

From the time of its first appearance the neural groove is con-
tinuous with a median notch in the dorsal lip of the blastopore.
There is evidence that the neural groove early acquires a segmented
structure.

Transverse grooves, definite in number and location, cross
the neural plate, dividing it into true segments or neuromeres.
In the region. of the mesoblastic somites the transverse grooves
of the neural plate are in line with the intersomitic grooves, and
the neuromeres are in line with the somites. Segmental flexures
of the neural folds sometimes occur; in these cases the outward
flexures of the neural folds are in line with the transverse grooves,
and the inward flexures are in line with the neuromeres.

At the time of the closure of the neural folds, the embryo has
increased in length so that it extends over about one-half of the
circumference of the egg.

1X. DEVELOPMENT AFTER THE CLOSURE OF THE NEURAL FOLDS
A. Description by stages, to the time of hatching

Most of the important features of the later external develop-
ment are sufficiently illustrated by the photographs. Only a
brief account is here necessary and this will deal principally with

520 BERTRAM G. SMITH

observations on living material. Some comparisons with Nec-
turus have been given in a previous paper (Smith ’11 a). Late
stages of Cryptobranchus japonicus have been figured by Ishikawa
(04 and ’08) and de Lange (’07).

Stage 16: (figs. 235 to 237). This stage is reached about eight-
een. hours later than the beginning of Stage 15. It is character-
ized by closed neural folds which are still more or less separated
by a median groove. Ganglionic ridges are forming at the sides
of the brain. The blastopore is no longer a transverse slit, but
a small round orifice which probably represents the definitive
cloacal opening. Up to this time the great majority of the eggs
have retained the ‘vitelline membrane.’ During Stage 16 or
slightly later this covering usually becomes ruptured as a conse-
quence of the growth of the embryo and is finally cast off.

During the gastrula and open neural groove stages, careful
observations have been made to test the presence of cilia on the
ectoderm, with absolutely negative results. Currents of water
produced by ciliary motion. may be detected through the move-
ments of yolk particles within the vitelline membrane when this
is present, or by means of powdered carmine added to the water
in cases where the vitelline membrane has been shed. At the
time when the neural folds are closing, cilia are present on the
sides of the body and the ventral surface of the yolk sac, but are
absent from the neural folds. The general direction of the ciliary
currents is toward the posterior end of the body.

Stage 17: (figs. 238 to 242). This stage is reached about a
day later than Stage 16. The neural folds are definitely closed
and the head well upraised. The optic vesicles are indicated
by slight paired expansions of the anterior part of the brain.
In some embryos the anlage of the pronephros is apparent through
an elevation of the overlying ectoderm. Cilia are absent from
the dorsal surface above the neural tube but are quite generally
present elsewhere and are particularly strong or numerous on
the dorsal surface of the body, lateral to the neural tube. In
general the beat of the cilia along the sides and ventral surface of
the body is backward.

EMBRYOLOGY OF CRYPTOBRANCHUS yA

The embryo is still erect (i.e., with the dorsal surface upper-
most). In this position it has been observed, in many cases,
to rotate slowly on a vertical axis. To test the direction of rota-
tion a large number of embryos were placed separately in watch
glasses and individual records made. Out of sixteen embryos
that showed rotation, only two moved in a clockwise direction,
the other fourteen in an anti-clockwise direction. The rotation
is, of course, caused by the cilia. The direction of rotation in
this stage can hardly be explained as the result of a tendency
for the embryos to lean to one side oftener than to the other, for,
as will be shown in a later stage, the facts are otherwise. Possi-
bly more extended observations would show more equality in
the results; or there may be a uniform asymmetry in the distri-
bution, or in the rate or direction of beating, of the cilia of the
ventral surface of the yolk sac.

Stage 18: (figs. 243 and 244). This stage is reached about
twenty-four hours after the beginning of Stage 17. It is charac-
terized by a prominent outstanding head with marked cephalic
flexure and distinct optic vesicles, and by the presence of the
pronephros and the first definite indications of the budding tail.
During the latter part of this stage the mandibular arch is usually
recognizable.

Patches of cilia are now distributed over the entire surface:
the beat of the cilia is in general backward and the currents are
much the same as figured in the next stage (fig. 190).

During this stage the embryo topples over from its erect posi-
tion so as to fall to one side, on which it lies throughout several
succeeding stages until spontaneous movements enable it to change
its position. An incidental result of this position is to bring a
larger areaof the ciliated surface into contact with the substratum:
as a consequence of this and of the stronger ciliation, rotation
of the embryo is now of more marked occurrence. The most
rapid motion observed was performed by an embryo that com-
pleted a rotation in just two minutes.

The functional value of the ciliary motion is at least two-fold:
(1) it bathes the surface of the embryo with currents of water
which are subservient to respiration; and (2) rotation of the

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

BPs BERTRAM G. SMITH

embryo, when it occurs, serves to prevent adhesion of the embryo
to the envelope with consequent abnormalities.

The ciliation and rotation of the frog embryo have been
described by various writers, notably Assheton (’96). Piersol
(09) bas deseribed rotation in the embryo of Plethodon.

Stage 19: (figs. 190 and 245 to 248). This stage begins about
twenty-four hours later than Stage 18. It is characterized by
from two to three distinct gill invaginations, a budding tail, a
very marked outward expression of the pronephros (see especially
figs. 245 and 246), and beginning lateral vascular bands, the
anlage of the vitelline veins (see especially fig. 245). In addition

190 191

Fig. 190 Diagram of an embryo of Cryptobranchus allegheniensis in Stage
19, showing the direction of the water currents produced by cilia.
Fig. 191 Same as figure 190, for Stage 21.

to the cephalic flexure there is a slight cervical flexure which
reaches its maximum in this stage. About sixteen to twenty
mesoblastic somites are apparent in surface views.

In the living embryo, the lateral vascular bands are conspicuous
structures, for they are pink with blood; but they are not yet
differentiated into true veins. Overlying the upper part of the
yolk sac, they extend from the heart region longitudinally on
each side of the body and meet posteriorly a little below the tail.
During this stage and the stages immediately following, they
shift slowly toward the ventral surface. There is considerable
variation in the position of thisband in embryos that are otherwise
in the same stage. <A similar vascular area has been figured for
Cryptobranchus japonicus by de Lange (07).

EMBRYOLOGY OF CRYPTOBRANCHUS 523

In this stage practically all embryos are found lying on either
the right or the left side. To determine the relative number of
cases of each, one hundred and nine embryos, taken at random
from several different spawnings, were examined; fifty were found
to lie upon the right side, fifty-nine upon the left side. These
numbers are approximately equal, hence the occurrence follows
the laws of chance.

Rotation of the embryo now occurs in a greater proportion of
cases than in the preceding stage, though of the embryos studied
hardly more than one in ten has been observed to rotate. It
is noticeable that in some cases the direction of rotation is clock-
wise, in other cases anti-clockwise. In eleven embryos that
showed rotation eight moved in an anti-clockwise direction; all
these were lying on the left side. Three embryos rotated in a
clockwise direction; of these two were lying on the right side,
one on. the left side. Hence as a general rule the direction of
rotation is correlated with the position of the body. The actual
direction is not what one would expect if the beat of the cilia
were directly backward in all parts of the body; but by an inspec-
tion of figure 190 it will be seen that in the head region the beat
is ventrad, in the posterior region dorsad. The ciliary currents
are alike on the two sides of the body.

Stage 20: (figs. 192 and 249 to 251). This stage is reached
about two days later than the beginning of Stage 19. Through
the loss of the cervical flexure the head is now brought more nearly
in. line with the body, but on the other hand there is an increase
in the cephalic flexure. Five gill invaginations are usually visible,
the two posterior ones being sometimes indistinct. There are
distinct nasal pits. The roof of the medulla is becoming thin and
transparent. From twenty-five to thirty mesoblastic somites
are apparent in surface views. The tail undergoes a decided
increase in size, and is slightly flexed ventrally. The front limb
anlagen appear during the latter part of this stage.

The lateral vascular bands are not yet differentiated into veins,
but during the latter part of this stage some small veins have
been observed, in living material, extending vertically from the
vascular bands for a short distance above them (fig. 192). No

=P 197

Figs. 192 to 197 Development of the vitelline veins of Cryptobranchus alle-
gheniensis. All the figures are from living material and, with the exception of
the veins in figure 192, are drawn with the aid of acamera. X 5.

Fig. 192 Stage 20. The stippled area indicates the lateral vascular band.
The right and left sides of the embryo are practically alike.

Figs. 193 and 194 Right and left views of an embryo in Stage 22.

Fig. 195 Lateral, slightly ventral, view of an embryo nearly ready to hatch.

Fig. 196 Ventral view of an embryo nearly ready to hatch.

Fig. 197 Ventral view of a larva about ten days after hatching.

524

EMBRYOLOGY OF CRYPTOBRANCHUS 529

movement of the blood has been observed in this stage. The
embryos are ciliated, and undergo rotation, much as in the pre-
ceding stage. Muscular movements do not ordinarily occur,
but have been observed when the embryos were placed in fixing
fluids.

Stage 21: (figs. 191 and 252 to 255). This stage is reached
about four days later than the beginning of Stage 20. It is char-
acterized by the presence of three budding external gills, and
small front limb rudiments. The tail is longer, and has a more
decided ventral flexure, than in the preceding stage.

The dorsal surface of the embryo is now for the first time
slightly pigmented. The pigmentation begins in that portion
of the ectoderm overlying the nervous system, and gradually
extends downwards over the sides of the body. During this
and the following stages (McGregor ’97), the pigment cells in
the region of the mesoblastic somites are grouped metamerically.

Numerous veins now branch off dorsally from the lateral vas-
cular bands. In the bands themselves, the two main trunks
of the omphalo-mesenteric or vitelline veins, one on each side
of the yolk sac, are being differentiated, but they rarely become
complete before the next stage. In general the differentiation
of the vitelline veins has gone further on the side of the body that
happens to be uppermost. During the latter part of this stage
the heart is pulsating regularly, with about twenty-five to forty
beats per minute; the blood sometimes surges back and forth
in the principal veins overlying the yolk sac, with a slight excess
in the forward movement, but the vitelline circulation is not
yet completely established.

The distribution of cilia remains much the same as in the pre-
ceding two stages, but the ciliary currents move more uniformly
toward the posterior end of the body (fig. 191).

Spontaneous muscular movements, consisting chiefly of a
bending of the body laterally into a U shape, now occur, but the
embryo is as yet unable to turn over.

Stage 22: (figs. 193, 194 and 256 to 259). This stage begins
about five days later than Stage 21; its most distinctive charac-
teristic is that the external gills have rudimentary branches and

526 BERTRAM G. SMITH

are pink with blood. The tail is rapidly increasing in size and
beginning to straighten out. The front limb rudiments now
take the form of conspicuous outstanding lobes, the distal ends
not yet divided into digits. ‘Toward the close of this stage the
rudiments of the hind limbs appear. The entire dorsal surface
of the body is well sprinkled with pigment cells; the eyes are
becoming deeply pigmented. The lateral line system is well
developed. During this stage occurs the rapid formation of
the gular folds.

The lateral vascular bands no longer appear as such, but on
their site are differentiated the two main trunks of the vitelline
veins (figs. 193 and 194). As compared with the earliest position
of the lateral vascular bands, the vitelline veins lie considerably
nearer the ventral surface of the yolk sac. While the vitelline
system of veins is primarily a paired one, almost from the begin-
ning one side is usually found better developed than the other.
The heart now contracts at the rate of about forty to sixty beats
per minute, and the blood pulsates regularly through the vitelline
veins.

The cilia are especially well developed on the external gills;
here the ciliary currents are strongest. Spontaneous muscular
movements now occur at frequent intervals. The movements
consist of jerking the head from side to side; wriggling; reversal
of the laterally curved position of the body by turning over;
swimming movements by means of which the embryo butts
against the envelope; and swimming in a circle. .The functional
value of these movements seems to be to afford exercise for the
developing muscles. Embryos removed from the capsules at
this stage make practically the same movements; they are unable
to progress in a straight line and are incapable of prolonged swim-
ming movements.

During the later stages of development before hatching, the
water in which the embryos are kept has a pronounced ‘fishy’
odor. yes

Stage 23: (figs. 195, 196, 260 and 261). ‘The limits of this
stage are fixed to include the time of hatching. In a given lot
of embryos hatching does not fake place all at once, but extends

EMBRYOLOGY OF CRYPTOBRANCHUS Eh

over a period of about a week, with a corresponding variation
in the degree of development of different individuals at the time
of escape from the envelopes. On the average, hatching occurs
about two weeks later than the beginning of Stage 22, and about
six weeks after fertilization. Previous to the hatching period,
the envelopes become much softened and considerably enlarged
by the absorption of water, making room for the growing embryo.
The latter usually escapes by pushing (‘worming’) its way through
the envelope, leaving a smali round hole; in some cases it bursts
the envelope by means of wriggling movements.

The newly hatched larva measures about 23 to 25 mm. in
length. Very noticeable is the retention of a large yolk sac with
conspicuous bright-red vitelline veins; the bushy external gills
are pink with blood. In proportion to size of body, the tail is
much larger than in the adult. The dorsal surface of the body
and the sides of the tail are well pigmented, but in general the
larva of Cryptobranchus, like that of Necturus, is pale as com-
pared with amphibian larvae that develop from a pigmented
egg exposed to the light. The ventral surface is lacking in pig-
ment, leaving the abdominal region yellow from the presence of
yolk, and the throat region transparent. The heart can be read-
ily observed without dissection. The anterior limb rudiment
is provided with two digits. In most specimens the body somites
are plainly visible, but they do not show well in the photographs.
On account of its large size, graceful outlines and bright colors,
the newly-hatched larva is a striking and beautiful object.

In the resting position, the larva lies on its side, turning occa-
sionally from one side to the other. The newly hatched larva is
able to swim rapidly in a straight line for a short distance, using
the tail as a propeller. The larvae avoid the light, and are posi-
tively rheotactie.

The vitelline veins (figs. 195, 196 and 261) have shifted further
toward the mid-ventral line; on one side of the body these veins
are well developed, on the other side they show arrest of devel-
opment with signs of atrophy. The heart now beats about sixty
to seventy times per minute.

528 BERTRAM G. SMITH

Aération of the blood is afforded, not only by the external gills,
but by the capillaries lying close to the surface over all the body.
On account of its great exposed surface, the tail may be of espe-
cial importance as a respiratory organ.

As in the preceding stage, cilia persist over the entire surface
of the body, and the ciliary currents are strongest in the vicinity
of the gills.

B. Larval development, and the metamorphosis

The changes in the form of the body, and the gradual increase
in size, during the first year of larval development, are shown
by the photographs (figs. 262, to 267).

Year-old larvae reared in the laboratory reach a length of
about 5 to 7 em.; three two-year-old specimens reared in the lab-
oratory measured respectively, after preservation in alcohol, 7
cm., 8 em. and 9.5 em. Near the close of the second summer
these latter specimens lost their external gills. The few speci-
mens with external gills taken in August from their natural
environment (they were found under small flat stones in shallow
water) measured as follows: 6.4 cm., 6.8 cm., 7.0 cm., 7.3 em.,
7.7 cm., 12.0 em., 12.3 cm. It will be noticed that these speci-
mens sort into two lots, one lot containing those with body lengths
ranging between 6 and 8 cm., the other lot containing specimens
approximately 12 em. in length. Though this data is rather
meager, the rather considerable gap between the two lots suggests
that we are dealing with larvae of the first and second summers,
respectively. In comparing the larger larvae taken from their
natural environment with the two-year-old specimens reared in
the laboratory, allowance must be made for the fact that the
latter were measured after being shrunken by preservation in
alcohol; moreover the body form of the laboratory specimens
seems to be shorter and stouter than the normal. Specimens
with a body length of 14 cm. and more, taken from their natural
environment, have invariably lost their external gills. The com-
bined evidence from specimens reared in the laboratory and

EMBRYOLOGY OF CRYPTOBRANCHUS 529

those taken from their natural habitat indicates that the meta-
morphosis occurs at the end of the second year.

At the time of hatching, the embryo retains a supply of yolk
sufficient to last it for several months; the mouth is still quite
ventrally situated. So far as its method of nutrition is concerned,
during this period the young Cryptobranchus is an embryo rather
than a larva. Gradually the yolk disappears, and the mouth
assumes a terminal position. Specimens reared in the laboratory
begin to take food about two to four months after hatching; they
must be fed individually with bits of scraped beef. No notice
is ordinarily taken of the food unless it is moved about imme-
diately in front of the animal and preferably a little to one side of
the mouth. Some of the specimens take food more readily, and
grow more rapidly, than others. One lot of larvae, reared in
the laboratory, ate young frog tadpoles. A 12 em. specimen
taken from its natural habitat ate a large Corydalis larva; another
newly captured 12 cm. specimen regurgitated a partly digested
6 em. larva of its own kind.

During the first month after hatching, the vitelline veins of
one side of the yolk sac degenerate, while those of the other side
shift to a more nearly median position (fig. 197). Degeneration
of the right or the left vitelline vein takes place in about an equal
number of cases. It has already been noted that in those stages
when the embryo lies continuously on one side the vitelline veins
are best developed on the uppermost side; furthermore that
the embryo falls on the right or the left side in about an equal
number of cases. The facts strongly suggest ‘that the position
of the embryo during the period when the vitelline veins are
developing is the factor that determines on which side the vitel-
line vein shall persist; but since making these observations I
have had no opportunity to put the matter to a rigid test.

With the reduction of the yolk sac, the vitelline circulation
suffers a corresponding diminution in extent; during the late
stages of this process, through the increasing thickness and opac-
ity of the ventral body wall the vitelline veins are somewhat
obscured.

530 BERTRAM G. SMITH

In the free-swimming larva, the heart beats more rapidly than
was the case during embryonic development.

Withina week ortwo after hatching, the rapid growth of the front
limb rudiments enables the larva to support itself in the normal
position of the adult. One month after hatching, the front limbs
have increased decidedly in length and possess the full number
of digits (four). The form and position of the front limbs adapt
them for use as paddles; by means of a simultaneous backward
stroke they aid the larva in getting a quick start for swimming.
The posterior limbs develop more slowly; at this time they are
relatively short and as a rule possess but three digits, though in
some cases the full number (five) are present. Ten weeks after
hatching in all cases both pairs of limbs possess the full number
of digits and are used in walking in the same manner as in the
adult. The limbs are broad and flat, and in swimming at a
moderate rate of speed are used as paddles. After the sixth month
of larval development the posterior limbs surpass the anterior in
size and strength. In the two-year-old specimens reared in the
laboratory the limbs appeared weak and poorly developed as com-
pared with newly captured specimens of the same age.

As in the adult, the tail is the principal organ of locomotion
during rapid swimming. Up to five or six months after hatching
the tail remains much larger in proportion to body-size than in
the adult. In the year-old larva the tail is proportionally much
smaller than in specimens ten weeks after hatching. In the
two-year-old specimens reared in the laboratory the tail is smaller
than in newly-captured specimens of the same age.

‘One month after hatching, pigmentation is greatly advanced
and extends over the external gills; when viewed from above
the larva is now nearly black. The ventral surface remains
white and nearly transparent in the throat region, yellow in the
region of the yolk sac. Six months after hatching, the dorsal
surface shows large dark spots of unusually dense pigment, which
are characteristic of all the later stages; the ventral surface is
slightly pigmented. In the year-old larva, the dorsal surface
is still nearly black, but with a few scattering inconspicuous
yellow spots; the abdomen is grayish and the throat region almost

EMBRYOLOGY OF CRYPTOBRANCHUS 53k

white. In the two-year-old larva the general color effect is not
so dark; the larva is taking on the variegated color pattern of
the post-larval and adult stages.

In the month-old larva, shedding of the cuticle was observed
for the first time. From this time on, the water of the aquarium
becomes almost cloudy with detached portions of epidermis.

Soon after the hatching period, cilia disappear from the general
surface of the body, but persist on the gills where a strong eddying
current of water is produced. The gills remain ciliated until
at least six months after the hatching period; at this point obser-
vations were necessarily discontinued.

As early as five months after hatching and frequently there-
after, larvae have been observed to come to the surface for air
and to give up large bubbles of air from the mouth, indicating
that the lungs are functional.

During the first year of larval life, the ventral portions of the
three gill arches bearing the external gills are expanded into
leaf-like plates, partly covered by the opercular lateral portions
of the gular fold. At this time there are three gill openings.
Toward the end of the first year the median portion of the gular
fold disappears, but the lateral portion extends as a small opercu-
lar fold above the root of the anterior external gill. During
the second sunimer the opercular fold extends dorsally farenough
to cover the bases of all three external gills. With the loss of
the external gills the opercular flap persists and partly roofs
over a shallow cavity containing the leaf-like plates previously
mentioned.

C. Post-larval stages

Sexual maturity is attained with a length of at least 30 cm.
for the male, and 35 em. for the female. During August, speci-
mens 14 to 20 em. in length have been very rarely taken, while
specimens measuring from 20 cm. upwards are very plentiful ;
this suggests that the young ordinarily reach a length of at least
20 cm. at the end of the third vear, and that they do not ordinarily
become sexually mature until the end of the fourth year.

532 BERTRAM G. SMITH.

The immature post-larval stages resemble in coloration the
young adults described in Part I. Aside from the loss of the
external gills, the most conspicuous changes as compared with
the larvae are. a slight progressive dorso-ventral flattening of
the body, and the gradual development of the folds of the skin
which are so prominent in adult and especially in very large
and presumably old specimens. A transverse slit-like spiracular
opening is bordered anteriorly by the small opercular flap and
posteriorly by a similar but still smaller fold of skin. The spiracle
leads to a small cavity containing a single persistent gill opening
bordered by two leaf-like plates; the other gill-lamella is greatly
reduced or has disappeared.

D. Summary

The vitelline membrane is shed about the time of the closure
of the neural folds.

Shortly after the closure of the neural folds, the entire surface
of the embryo becomes ciliated. The beat of the cilia is in gen-
eral toward the posterior end of the body; currents of water are
produced which are subservient to respiration. After the appear-
ance of the external gills, ciliary currents are especially strong
in their immediate vicinity. Soon after the hatching period,
the cilia disappear except on the external gills; here they persist
until at least six months later.

Soon after the closure of the neural folds, the embryo falls on
one side, where it lies until, in a much later stage, it is able to
turn over through muscular activity. The embryo falls on the
right or the left side in about an equal number of cases.

Rotation of the embryo, due to the beating of the cilia, com-
mences before the embryo has fallen on its side but is more pro-
nounced afterward. In most cases rotation proceeds in a clock-
wise direction when an embryo is lying on its right side, in an
anti-clockwise direction when it is lying on its left side. Rotation
may be of service in preventing adhesion of the embryo to the
capsule.

The vitelline veins develop as paired structures along the
sides of the yolk sac, and shift gradually toward the ventral

EMBRYOLOGY OF CRYPTOBRANCHUS 533

median line. The vitelline veins develop more rapidly on that
side of the yolk sac which happens to be uppermost. Soon after
the hatching period, the vitelline veins of one side degenerate,
while those of the other side reach their fullest development;
the veins of the right or the left side persist in about an equal
number of cases. Probably the position of the embryo during
the stages when it lies continuously on one side is the factor that
determines which set of vitelline veins shall gain the ascendancy.

The newly hatched larva retains a supply of yolk sufficient
to last it from two to four months.

The tail of the early larva is proportionally much larger than
in the adult.

Pulmonary respiration is established about five months after
the hatching period.

The metamorphosis takes place at the end of the second vear.

Sexual maturity is attained, probably at the end of the fourth
year, with a body length of at least 30 em. for the male and 35
em. for the female.

X. TIME RECORD

For the characteristics of the different stages reference is made
to the text and illustrations, especially the photographs. For
methods used in obtaining the time record, see the introduction
(Section VI). Table 1, page 534.

XI. ABNORMALITIES

The present section deals primarily with abnormalities found
in embryos taken from their natural environment, or kept under
conditions as nearly normal as possible.

1. Large yolk plug

In the early stages of development the most common abnormal-
ity is the presence of an unusually large yolk plug. Examples
are shown in figures 198 and 199, though these are far from rep-
resenting extreme cases. In well-marked examples the blasto-
pore forms a complete circle only a little below the equator, and

534 BERTRAM G. SMITH

TABLE 1

STAGE TIME AFTER FERTILIZATION * INTERVAL

24 hours (18 to 28)

1
vy 30 hours (24 to 36) 6 hours (5 to 7)
3 35 hours (29 to 44) 5 hours (4 to 6)
4 39 hours 4 hours (3 to 5)
5 43 hours 4 hours
6 47 hours 4 hours
7 51 hours | 4 hours
8 63 hours 12 hours
9 33 days | 19 hours (14 to 24)
10 5 days 36 hours (24 to 48)
11 7 days (6 to 8) 48 hours
12 10 days | 72 hours
13 113 days | 36 hours
14 | 123 days | 24 hours
15 | 13; days 18 hours
16 14 days 18 hours
17 15 days | 24 hours
18 | 16 days 24 hours
19 17 days 24 hours
20 19 days | 48 hours
21 23 days 4 days
22 | 4 weeks 5 days
23 6 weeks 2 weeks

Metamorphosis: Two years after fertilization.

a large yolk plug persists long after the closure of the neural
folds; such embryos usually produce normal larvae, though spina
bifida is an occasional result. In extreme cases the blastopore
may form entirely above the equator; such embryos die before
reaching the larval stage.

This abnormality appears most frequently in eggs that have
been kept in jars of water during warm weather, and especially
in material that has been shipped long distances. Probably it
may be brought about by a variety of unfavorable conditions:
heat, lack of oxygen, mechanical agitation, and injurious sub-
stances (e.g., cinders) in the water. It is readily produced by
treatment with sodium chlorid; very marked cases of spina bifida
may result.

EMBRYOLOGY OF CRYPTOBRANCHUS Ray)

198 199

Figs. 198 Postero-dorsal view of a gastrula of Cryptobranchus allegheniensis

with an unusually large yolk plug. The blastopore extends around the egg as
a complete circle, but only the dorsal portion shows distinctly in the photograph.

From preserved material. 4.

Fig. 199 Posterior view of an embryo shortly before the closure of the neural
folds, showing persistent yolk plug. Photographed from preserved material.
x 4.

2. Exovate abnormality

A rather common abnormality originating at the time of the
formation of the neural groove may be called the exovate abnor-
mality. A small protuberance on the site of the closed fenestra
takes the form of a spherical exovate; in some cases this reaches
a diameter about half as great as that of the egg. In most cases
the extra-ovate remains connected with the egg by a very narrow
stalk; the malformation is entirely extra-embryonic, the pro-
truded yolk is gradually absorbed and the egg produces a normal
larva. In other cases the protuberance takes the form of a dome-
like swelling and increases in size until the egg collapses.

3. A double embryo

On September 27, 1906, I found a nest containing embryos
in an advanced stage of development, and among them the double

536 BERTRAM G. SMITH

embryo shown in figure 200. The total bulk of this double em-
bryo is about equal to that of an ordinary single embryo from
the same spawning. That the occurence of such a monstrosity
in this species is exceedingly rare in nature is shown by the fact
that during the past seven years I have collected many thousands
of embryos from nests, yet in no other instance have I found an
abnormality even approaching the one under consideration.

Fig. 200 Double embryo of Cryptobranchus allegheniensis. Photographed

after preservation in formalin. X 4.

The precise manner in which the double embryo originated
is problematical. Single capsules sometimes contain two eggs
in close contact; but the small size of the double embryo pre-
cludes the possibility that it was formed by the union of two
such eggs. Through treatment with sodium chloride I have
produced embryos in which the brain differentiated without
closure of the neural folds, but no true double embryos were
obtained by this means; moreover in nature it is improbable
that unusual chemical influences should affect a single ege.

It is more likely that the abnormality was brought about by
mechanical means. <A partial separation of the first two blasto-

EMBRYOLOGY OF CRYPTOBRANCHUS Sow

meres (e.g., through the egg becoming lodged in a crevice) might
in some cases result in the formation of a double embryo (Spe-
mann ’01 to 703). But owing perhaps to the heavily yolk-laden
character of the egg, and the slowness with which the first furrow
cuts its way through the yolk, my attempts to produce a double
embryo in this manner have failed. When a fine silk thread or
a hair was tied about the egg in the two-cell stage so as to con-
strict it in the plane of the first cleavage furrow, the egg usually
burst before reaching the gastrula stage. In a single case the
egg lived until after the formation of the neural folds, but devel-
oped a single embryo with its principal axis at right angles to
the constricting cord.

A more probable explanation is that in the two-cell stage the
egg became inverted and remained for some time in this position
subject to the rearrangement of contents through the disturbing
influence of gravity acting in a direction opposite to the normal
(Schultze ’95). But the whole question is complicated by the
more fundamental problem of the determination of the median
plane of theembryo. My experiments, to be described in a later
paper, show that in Cryptobranchus the first cleavage furrow
tends to form at right angles to the direction of entrance of
the sperm (as in Triton, but not as in the frog where the first
cleavage tends to coincide in direction with the pathof the sperm).
If in Cryptobranchus, as in the frog, the entrance path of the
sperm les approximately in the median plane of the embryo,
then the conditions necessary for the production of a double
embryo by the separation of the first two blastomeres must be
of very exceptional occurrence.

4, Spiral-tailed monsters

In the fall of 1910, some embryos shipped by express to the
University of Wisconsin and reared there in city water, acquired,
in about 8 per cent of their number, the abnormality shown in
figures 201 and 202. From the condition of the tail in late larval
stages, specimens affected with this malformation may be desig-
nated as ‘spiral-tailed monsters.’ At first this peculiarity seemed

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

ww t
>

BERTRAM G. SMITH

merely the persistence of an embryonic condition; for in early
stages the tail is always strongly flexed and in these particular
cases it failed to straighten out. But as development progressed
the tail became twisted into a pronounced spiral. The illustra-
tions show the extreme condition; cases occur forming a series 1n-
termediate between this and the normal. In some cases the back
is arched or humped. So far as ean be judged from surface views,

Fig. 201 Spiral-tailed monster of Cryptobranchus allegheniensis at the time

of hatching. Photographed from the living embryo anaesthetized with chlore-
tone. XX 3.

the entire abnormality seems to be brought about by the con-
stricting effect of a band of tissue lying in the ventral median line.

As a result of this malformation, the larva is compelled to le
on one side, and can swim only in a grotesque fashion, with back-
ward circling movements. Affected larvae take no food, and
die shortly efter using up their supply of yolk.

An analysis of the water was secured but the subject has not
vet been further investigated. Stockard (’06, p. 119) noted
the occurrence of a similar abnormality in Fundulus embryos.

EMBRYOLOGY OF CRYPTOBRANCHUS 939

5. Critical periods in the life history

In closing this section it remains to note that there are three
stages in the life history characterized by unusual mortality—
nature’s examinations occur in these stages. These critical
periods of development are:

Fig. 202 Spiral-tailed monster of Cryptobranchus allegheniensis two months
after hatching. Photographed from the living larva anaesthetized with chlore-
tone. X 3.

(a) The late blastula. Many eggs seem unable to form a gas-
trula.

(b) The hatching period. Many embryos seem to lack the
strength necessary to escape from the gelatinuous envelope,
and die in the capsule.

(c) The period of change in the method of nutrition. After ex-
hausting their supply of yolk, many larvae refuse to take food,
or seem unable to set up digestive processes, and die of starvation,

540 BERTRAM G. SMITH
XII. PHYLOGENY

We are here concerned with (a) the origin of the amphibia in
general, (b) the origin of the urodeles, and (c) the interrelation-
ships of the urodeles witb special reference to Cryptobranchus.
The evidence may be classified as (a) anatomical, (b) paleontolog-
ical, and (c) embryological. In the following survey, no attempt
is made to keep these lines of evidence strictly separate; for
paleontology is simply an extension of comparative anatomy to
fossil forms, with the added element of sequence in time; while
the development during the late embryonic and larval stages
gives the key to many adult structures that could not otherwise
be homologized.

It is not within the scope of this paper to enter extensively
into anatomical and paleontological questions; but as a check
on possible generalizations derived from the study of amphibian
embryology the writer has devoted considerable time to a study
of the paleontological material in the American Museum of Nat-
ural History, and has endeavored to acquire some degree of
familiarity with the results of modern research in this field. In
this work he has been greatly aided by Dr. W. K. Gregory, who
has in progress a detailed review of the origin of the amphibia
(see abstracts in Science, Gregory ’11 a, ’11 b and 712).

The tetrapoda form a coherent group. The gap between this
group and the fishes is one of the weak places in vertebrate phylog-
eny, and it is here that the idea of continuity in the descent of
the higher vertebrates has been most often attacked; some have
maintained a diphyletie origin for the tetrapods and fishes.

The amphibia undoubtedly include the most primitive known
tetrapods. It is on this account that the amphibia possess a
peculiar interest from a phylogenetic point of view; the problem of
the origin of the amphibia is the problem of the origin of the
tetrapods in general.

The fundamental unity of the gnathostome type leads us to
look among the true fishes for the nearest living or fossil repre-
sentatives of the ancestral stock of the amphibia. For in fishes
and tetrapods the following features, as well as many others,

EMBRYOLOGY OF CRYPTOBRANCHUS 041

are homologous: the chief divisions of the brain; the cranial
nerves; the eye muscles and their innervation; the chondrocra-
hium arising from trabeculae and parachordals, and including
olfactory, optic and auditory capsules; the visceral arches; and
the essential structures of the circulatory system. If we compare
only the amphibia with the fishes the range of resemblances
becomes still greater; all the important structures are in essential
agreement except those concerned with the outer halves of the
paired extremities.

The argument for a diphyletic origin of the pisces and amphibia
is based upon (a) the lack of homology in certain elements of
the skull, and (6) the difficulty that is experienced in deriving
the amphibian limb from the fin of any known fish. But the
resemblances in the skull bones become very close when we con-
sider fossil forms, and the trend of increasing knowledge is in
the direction of more complete homology rather than the reverse.
Moreover we find difficulties almost as great in homologizing
cranial elements in different fishes; for these structures are plastic
and exposed to environmental influences, and with the radical
change from an aquatic to a terrestrial habitat we should expect
them to be profoundly influenced. The endoskeleton of the
paired appendages presents us with a problem of greater difficulty
but here, too, we have to deal with structures that we should
expect to be greatly modified in connection with the change of
habitat. In view of the wide range of resemblances in important
structures one is hardly inclined to consider seriously the idea of
a diphyletic origin for the fishes and amphibia.

The amphibians must have descended from some fish having
scales with the potentiality of fusing into bony plates. The
dermal bones forming the roof of the skull must have been arranged
in pairs on each side of a median suture; for this is the condition
found in the most primitive known amphibians (e.g., Branchio-
saurus). The ancestral form must be sought in some fish having
the endoskeleton of the paired fins widely protruded from the
body, and with pectoral and pelvic members similar. Such
fins functioned primitively as paddles; with the adoption of a
terrestrial method of locomotion, by creeping or crawling on a

542 BERTRAM G. SMITH

more or less solid substratum, the endoskeletal elements of the
limbs become greatly strengthened and progressively longer
in order to lift the body from the ground. Such a progressive
elongation of the limb bones, particularly the proximal elements,
may be traced in both fossil fishes and amphibians; in the latter
the pelvic girdle is also found becoming definitely articulated
with the axial skeleton. The ancestral form must have been
short -bodied; for in many groups of animals a progressive elonga-
tion of the body, culminating in eel-like forms, is found to accom-
pany a degenerate structure scarcely capable of giving rise to
higher forms (Gregory 07, Appendix I).

The lack of scales with the potentiality of fusing into bony
plates is alone sufficient to exclude the elasmobranchs from the
immediate ancestry. For affinities ancestral to the amphibia
most authors have looked to the crossopterygii or the dipnoi.
Both have dermal bones, and both fulfil the requirement regarding
fins with widely protruded basal lobes and with endoskeletal
elements from which the framework of true limbs might be
derived.

At first sight the dipnoi seem best to bridge the gap between
fishes and amphibia. For the lung-fishes have survived by
virtue of an approach to the tetrapod type, enabling them to
exist during periods of drought. But various considerations
derived from the study of paleontology and comparative anatomy
make it probable that these terrestrial adaptations were independ-
ently acquired, and that the dipnoi were already too highly spe-
cialized in other respects to give rise to theamphibia. Themosaic
of small bones forming the greater part of the roof of the skull,
particularly in the fossil representatives of this group (e.g., Dip-
terus), and the usual occurrence of one to several large median
elements in this region, make it difficult or impossible to homol-
ogize the dermal bones of the skull with those of amphibia. The
characteristic dentition is far removed from that of the amphibia.
Marginal teeth, with exceptions in the cases of some very early
forms (e.g., Phaneropleuron), are lacking; in the later forms
the loss of maxillae, premaxillae and nasals shows a progressive
tendency toward degeneration in these regions. The concen-

EMBRYOLOGY OF CRYPTOBRANCHUS 543

tration of the teeth into tritoral clusters in the roof and floor of
the mouth is not in itself an unfavorable feature, for vomerine
teeth occur in the amphibia; but in all except certain very early
fossil forms (e.g., Uronemus) the fusion of these teeth into large
dental plates with grinding ridges has gone too far to give rise
to the condition found in the amphibia. So far as the paleonto-
logical and anatomical evidence is concerned, the known facts tend
to exclude the dipnoi from the direct ancestry amphibia, of the
yet do not wholly preclude the possibility that future discov-
eries may supply us with more favorable material amongst early
fossil forms. In so far as the terrestrial adaptations of the dipnoi
resemble those of the ampbibia, the case may be one of parallelism
or convergence; their more fundamental resemblances indicate
that they are not very far removed from a common ancestry.
Turning to the crossopterygil we find more favorable anatom-
ical and paleontological grounds for comparison with the am-
phibia. The dermal elements forming the roof of the skull occur
in paired series; a large number of cranial bones may be definitely
homologized with those of the amphibia (see especially Baur ’96;
Moodie ’08 a; for materials for further comparison see Goodrich
09, and Zittel 711). It is difficult to believe that identical rela-
tions in so many bones could have been independently evolved.
With regard to the fins, we find examples of a bifurcated type of
endoskeleton that makes a more favorable starting-point for a
tetrapod limb than the archipterygial type of the dipnoi; e.g.,
see fig. 203 for the pectoral fin of Sauripterus, and Goodrich
09, p. 275, fig. 244, for the pelvic fin of Eusthenopteron; the
pectoral fin of Eusthenopteron (Goodrich ’09, p. 282, fig. 252)
is not quite so favorable. But both of these forms belong to the
Rhizodontidae, whose skull is not so favorable for comparison
with the: amphibia as the skull of some other crossopterygians;
in no one form do we find all the conditions ideal for the deriva-
tion of the tetrapod type. The occurrence in Polypterus of two
kinds of ribs, both the ventral or pleural ribs characteristic of
the teleostomi and dipnoi, and the dorsal ribs characteristic of
elasmobranchs and tetrapods, is a point emphasized by Baur

544 BERTRAM G. SMITH

Fig. 203 Endoskeleton of the pectoral fin of Sauripterus taylori Hall, a cross-
opterygian from the upper Devonian. One-third natural size, linear reduction.
From a drawing by Dr. L. Hussakoff of the American Museum of Natural History.

(96) as favoring a crossopterygian rather than a dipnoan ances-
try for the amphibia.

It is upon embryological grounds that the strongest case has
been made out for the derivation of the amphibia from the dipnoi;
the known facts of development indicate the common origin
and later separation of these two groups. In so far as this view
is based upon a study of the early stages, the evidence may be
dismissed with the remark that the early development of the
crossopterygian Polypterus (Kerr ’07 a) resembles that of the
anura and the urodeles quite as much as does the early develop-
ment of the dipnoi (Semon ’00 and ’01; Budgett ’01; Kerr ’00,
’01 and ’09); furthermore that these early stages are of very little
value in connecting up. the great groups of vertebrates. But
some marked resemblances between dipnoi and amphibia in
the later stages of development cannot be disregarded. Kellicott
(05, a and b), on the basis of a detailed study of the circulatory
system of Ceratodus, came to the conclusion: ‘The resemblance
in the vascular and respiratory systems between Ceratodus, the

EMBRYOLOGY OF CRYPTOBRANCHUS 945

most primitive of the dipnoi, and the amphibia, especially the
urodeles, are numerous and important, and cannot be explained
as parallelisms.’’ In this connection it is important to compare
the development of the circulatory system of Polypterus. From
the account given by Kerr (’07), it appears that in Polypterus
the vascular system, particularly because of the presence of only
a single pair of aortic arches, is decidedly less amphibian in
character than that of Ceratodus.

In the present state of our knowledge it is impossible to reach
an unqualified decision of the question under consideration.
In weighing the evidence one should not forget that in numerous
“eases where anatomical and embryological evidence have come
into conflict in deciding questions of the phylogenetic relationships
of the larger groups of animals, it is the embryological evidence
that has had to give way, and that recent anatomical evidence
has to give place to paleontological. Whatever light may be
shed by future discoveries on the question of the derivation of
the amphibia from the crossopterygii or the dipnoi, it is clear
that the point of origin is not far from either stock; in other words,
that the three lines of descent have separated from a common
stem at no very great intervals.

Concerning the immediate ancestry of the living amphibia
we have detailed evidence only in the case of one group, which
fortunately for our purpose is the caudata or urodela. According
to Moodie (’08 b) the urodeles are descended from the branchio-
sauria, a group of primitive extinct amphibia from the carbon-
iferous and Permian. These are small, short-tailed amphibians
with broad heads. The skull is slightly more complex than in
the urodeles, and there is a dermal exoskeleton consisting of
rows of thin semi-cycloid scales, especially on the flanks and under
side of the body. External gills are present in the larvae. The
view that these forms are ancestral to the urodeles is based on
a detailed comparison of the structure of the skull, the structure
and form of the vertebrae and the ribs, the number of digits,
the arrangement of the phalangeal elements, the character of
the pectoral and pelvic girdles, the distribution of the lateral
line system, the structure and form of the long bones, and finally

546 BERTRAM G. SMITH

the shape of the body. According to Moodie the caudata are
degenerate branchiosaurians and the changes which have taken
place in the exoskeleton are mostly brought about by the loss of
certain parts. The urodele skull is especially degenerate in the
occipital and temporal regions; it may be derived from the
skull of Branchiosaurus (for which see Credner ’81 to ’90) by the
loss of the dermoccipitals (supraoccipitals of Moodie ’08; post-
parietals of Zittel 711), supratemporals (Zittel ’11), postfrontals,
postorbitals, sclerotics, epiotics (tabulares of Zittel ’11), jugals
and quadratojugals. The urodele skull has also in many cases
become narrow in connection with a general elongation of the
body—a degenerate feature. Hand in hand with the loss of
dermal elements in the skull has gone the loss of the exoskeleton
of the body.

The paleontological history of the apoda is unknown; but exist-
ing members of the group are in certain respects more primitive
than the urodeles and more nearly allied to the stegocephali.
Thus the hyoid and branchial apparatus is more primitive than .
that of any other recent ampbibia; dermal scales are present
which are proebably homologous with those of the stegocephali.
As in the urodeles, the skull shows degeneration in the loss of
certain bones; but the epiotics (tabulares) are retained, and occa-
sionally the postfrontals and the lacrimals. A second row of
teeth is sometimes present on the mandibles. Aside from some
very degenerate features, in otber respects the apoda are highly
specialized, indicating, as in the case of the anura, a line of descent
separate from the urodeles. Yet there are some very suggestive
resemblances. Attention has already been directed (Smith 712)
to the marked similarity in structure of the egg envelopes of.
Ichthyophis to those of Cryptobranchus and Amphiuma, and
to the likeness in the brooding habits; but in Ichthyophis the
female protects the eggs, in Cryptobranchus the male. In their
embryological development the apoda show many points of sim-
ilarity to the reptiles.

With this background, we now come to the question of the
interrelationships of the urodeles. In particular we are concerned
with the phylogenetic position of the aquatic urodeles (the peren-

EMBRYOLOGY OF CRYPTOBRANCHUS 547

nibranchs and the derotremes?) as related to the land-living
salamanders.

The most prevalent view has been that the aquatic urodeles
are the most primitive. Parker and Haswell (’97, vol. 2, p. 291)
have said: ‘‘The perennibranchiate urodeles are undoubtedly
the lowest of existing amphibia; they lead up, through such
forms as Amphiuma, with persistent gill slits but deciduous gills,
to the land salamanders, in which a purely terrestrial form is
assumed.’ In a group standing between fishes and the typically
terrestrial vertebrates, it is natural to regard the aquatic forms
as transitional to the terrestrial. In fishes there are usually
five branchial arches, and the gill slits remain open throughout
life. The land-living salamanders have in the adult state only
two branchial arches (Parker ’77; Wiedersheim ’77; Cope ’89),
and the gill slits are open only to the end of larval life. In the
aquatic urodeles there are usually three or four branchial arches
(as in the larvae of the terrestrial forms), and the gill slits usually
remain open throughout life—conditions intermediate between
fishes and salamanders. In view of the occurrence of external
gills in all larval forms, the persistence of such gills in the peren-
nibranchs might, on the recapitulation theory, be regarded as
a primitive character.

In the opinion of various authors, the above interpretation
represents a short-sighted view of the matter. Boas (’81) was
the first to assert that the perennibranchs are larvae that have
lost the ability to transform; this conclusion was reached as a
result of a comparative study of the circulation. Cope (’85)
described the retrograde metamorphosis of Siren, and concluded
that the present Sirens are descendants of a terrestrial type.
Gadow (’01, p. 66) suggested a terrestrial ancestry for the uro-
deles. Kingsley (’01) says: ‘‘The salamandrina form the central
urodele stem, and the perennibranchs and derotremes have
been derived from this stem by degeneration and the retention
of larval characters.’’ Kingsbury (05), on the basis of a compar-

2 The classification of Stannius (’56) is here followed, as it seems best adapted
for the purposes of this discussion.

D548 BERTRAM G. SMITH

ative study of the cranial elements, came to the tentative con-
clusion that Necturus is a permanent larva. The generalization
formulated by Boas has recently been reiterated and expanded
by Versluys (09); briefly stated, his views are as follows:

The great resemblance of the perennibranchs to salamander
larvae is only a consequence of the fact that the first are also
larvae, but larvae which no longer come to full development as
their ancestors did; the metamorphosis is imperfectly undergone
or wholly omitted. Nevertheless these larvae become sexually
mature, as in neoteny. In the course of time, in adaptation to
their aquatic habitat, they have become degenerate in many re-
spects. The derotremes are salamanders that have become fixed
in the transitional stage or metamorphosis. Thus the aquatic
urodeles have a terrestrial ancestry; they are forms which have
reverted to an aquatic mode of life.

The probable course of events giving rise to the perennibranchs
may be described as follows: While the mature salamanders
are constructed after the fashion of a land animal, their larvae
live in water and in the course of time have become more and
more adapted to aquatic life. They have extended their larval
organization and thus increased the difference, which must be
overcome by metamorphosis, between the larva and the grown-
up animal. In time some organs show arrested development
or degeneration of such asort that the larvae cannolonger develop
into land-living salamanders; they remain life-long water dwell-
ers. The condition is one of fixed neoteny (Boas 796). Accord-
ing to this view the perennibranchs are disconnected from one
another and have evolved as neotenic larvae from different
salamanders.

Confirmatory evidence for this view comes from the studies
of Emerson (’05) on Typhlomolge. In some details of its struc-
ture this animal shows a remarkable similarity to the larvae of
Spelerpes ruber. Probably Typhlomolge has been derived from
the neotenic larva of some salamander of the family (Plethodonti-
dae) to which Spelerpes belongs, adaptations to a subterranean
water life having been added.

EMBRYOLOGY OF CRYPTOBRANCHUS D49

The derotremes also are not primitive urodeles. Since they
show a mixture of the characters of the larvae and grown-up
salamanders, one cannot derive them from typical larvae. Pre-
sumably the derotremes are descended from typical salamanders
which have returned completely to aquatic life. Thus the meta-
morphosis, which serves to adapt the larvae to land life, has lost
its biological meaning; it extends over a longer time, and finally
sexual maturity overtakes the yet imperfectly built animal. Some
organs complete the metamorphosis, others retain wholly or
in part the larval condition. So in its skull Cryptobranchus
(Reese ’06) resembles the grown-up salamander; in its circulation
it retains larval characters. Versluys believes Cryptobranchus
to be descended from the amblystomidae.

If we take into account only living forms, much of the data
thus far considered in seeking a solution of this general question
of the relationship of the aquatic and land-living urodeles may
be read either way. The key to the situation lies inthe compar-
ison of the structures of existing urodeles with those of fossil
forms. We have seen that the extinct ancestor of the urodeles
was probably an animal whose skull was more complete than
that of any living urodele. In Necturus (Kingsbury ’05), there
is a considerable reduction in the number of cranial bones, com-
parable to the condition in the larval Spelerpes, and contrasting
with the condition in the adult Spelerpes. If the stem-form of
the urodeles was an animal like Branchiosaurus, then so far as
the cranial elements are concerned the salamander, and not the
perennibranch, is the more primitive form. There are five
branchial arches in Branchiosaurus—-a condition which does
not exclude either view of the interrelationships of the urodeles.
Branchiosaurus had external gills only in the larval stage; this
indicates that the condition found in the perennibranchs is proba-
bly not in the ancestral line of the caducibranchs (salamandrina).

Von Eggeling (’11), from a comparative study of the histogen-
esis of the skeleton of the limbs of urodeles, favors the view that
the perennibranchs, derotremes and siredon are derived from
the caducibranchs.

550 BERTRAM G. SMITH

With the transition from the water to the land, or vice versa,
one might expect important modifications in the sense organs.
In the amphibia this expectation is best realized in the auditory
organs, which in the form of a columellar apparatus fitting into a
fenestra ovalis here occur for the first time in the vertebrate
series. From a study of the modifications of this apparatus we
may well hope to obtain information concerning the more inti-
mate phylogenetic relationships within the group. According to
the recent work of Kingsbury and Reed (’09) on the urodeles,
there is close correlation between the type of auditory apparatus
present and the habits (aquatic, semi-aquatic, terrestrial, burrow-
ing, ete.) of the animals. From their results one may note many
resemblances between the adults of the typically aquatic forms
and the aquatic larvae of the terrestrial forms, as well as con-
trasts between both of these and the adult terrestrial forms.
Concerning the phylogenetic relationships as indicated by the
auditory apparatus Reed (09) says:

Cryptobranchus is the most generalized. The amblystomidae are
intermediate between Cryptobranchus and all other groups. The ple-
thodontidae and desmognathidae are departures from the Amblys-
toma stem while from these the sirenidae and Amphiuma seem to be
degenerated. Diemyctylus and Triton are identical with regard to
these ear structures and differ from all others. They are to be consid-
ered the most specialized. Between Diemyctylus and Triton on the

one hand and the amblystomidae on the other Salamandra stands inter-
mediate, resembling more strongly the amblystomidae.

If the generalized condition is really a primary one, then so
far as the evidence from this single character is concerned, Cryp-
tobranchus is one of the most primitive of urodeles; the evidence
is not in line with the hypothesis of Versluys. But one should
seriously consider whether the correlation between the sound-
transmitting organs and the environmental relations is not too
close to make the character of much phylogenetic value.

Osborn (’88) states that in Cryptobranchus we have the most
primitive type of brain thus far observed among the ampbibia.
But it is not clear that the simple condition found is necessarily
primitive.

EMBRYOLOGY OF CRYPTOBRANCHUS ao

Whipple (06), from a comparative study of the ypsiloid appa-
ratus in urodeles reaches the following conclusions:

(a) “That forms with lungs but without vestiges of an ypsiloid
apparatus, and with no evidence of degeneration in the pelvic
region (e.g., Necturus) are neither degenerate forms, nor perma-
nent larvae of any of the salamandrina.”’

(b) “That the presence of a functional ypsiloid apparatus
in Cryptobranchus indicates that Cryptobranchus les near the
line of descent of the salamandrina.”’

In summing up the facts for and against the hypothesis of
the phylogenetic relationships of the perennibranchs, derotremes
and salamandrina as outlined by Versluys, it seems to me that
the arguments in favor of the hypothesis are founded on charaec-
ters of greater phylogenetic value. In the reptiles and mammals,
land forms are always primitive, aquatic forms secondary (Os-
born ’02). To the writer the evidence seems convincing in favor
of a similar view for the recent urodeles. It might be added
that pentadactylous limbs, more or less perfectly developed in
fossil as well as recent amphibia, were undoubtedly produced
in connection with terrestrial habits; and it should be emphasized
that the forms ancestral to the present-day aquatic urodeles were
probably not purely terrestrial, but passed through an aquatic
larval stage, as In Branchiosaurus and most of the living sala-
manders.

It remains to consider briefly the phylogenetic value of some
of the faets concerning the life cycle of Cryptobranchus that
are embodied in the present contribution, and to discuss their
bearing on the subjects just treated from a historical point of
view. In an investigation thus far confined mainly to the exter-
nal features of development, manifestly little more than a begin-
ning can be made in such an interpretation.

The repeated failure of embryological generalizations to solve
some of the larger phylogenetic problems led to a widespread
reaction against the earlier too sweeping conclusions based on
the recapitulation theory. The reproductive processes, while so
fundamental, are very plastic, modified in closely related species
and even changing somewhat in the same species kept under

ay BERTRAM G. SMITH

different conditions. <A strict application of the biogenetic law
would imply that the early stages of development are exclusively
palingenetic; but from the earliest stages we find adaptations
(e.g., the presence of yolk) that are prospective in their signifi-
‘ance and have to do with distinctively larval phases, while the
larval characters may be highly coenogenetic. It is the failure
to distinguish between palingenetic and coenogenetic features
of development that is mainly responsible for bringing the recap-
itulation theory into disrepute. With a clearer recognition of
the limitations in this field of study, we may yet question whether
the reaction against the validity of the recapitulation theory
has not gone too far. As compared with anatomy and paleontol-
ogy, embryology is doubtless of little service in connecting up the
great groups of animals; yet it is indispensable in the solution of
many special problems.

In a comparative study of the breeding habits (Part I, Smith
12) we must remember that we are dealing with characters that
occur very late in the ontogeny, having to do with the adult
rather than with the developing animal. Consequently such
characters are of no more phylogenetic value than the habits
in general and the family, generic and specific morphological
characteristics of the adult; they are exceedingly plastic and of
value for comparison only within a very limited range of forms.

A phylogenetic interpretation of the methods of fertilization
in urodeles has been given in a previous paper (Smith ’07 b); it
is here referred to with the remark that in view of the conclusions
reached in the present paper regarding the trend of evolution
in the urodeles, the series should be reversed. If we could go
back far enough in the phylogeny of the vertebrates, doubtless
we should find external fertilization to be the primitive condition
(e.g., as in Amphioxus); but it is entirely possible that in Crypto-
branchus the method is secondarily acquired.

As we should expect, the brooding habit of Cryptobranchus
is very similar to that of the closely-related Amphiuma. The
absence of a brooding habit in Necturus is noteworthy. Brood-
ing habits very similar to that of Cryptobranchus are found in
Desmognathus and Plethodon, but with this important differ-

EMBRYOLOGY OF CRYPTOBRANCHUS 908

ence: in these latter forms the female, not the male, cares for
the eggs.

The egg capsules of the urodeles, as in other groups, show
generic and even specific differences (e.g., the specific differences
in the egg masses of Amblystoma, Smith ’11b). This is what
might be expected, since the capsules are the product of the soma,
not of the germ cells; the facts are in no way incompatible with
von Baer’s law. ‘The close resemblance between the egg envel-
opes of Cryptobranchus and Amphiuma accords with the sys-
tematic relationship. Amongst terrestrial urodeles there is no
close approach to the type found in Cryptobranchus; but in
general we should seek for affinities in forms having the egg
capsules more or less independent in the cluster, connected by
stalks (e.g., as in Desmognathus, Wilder ’99), rather than in
forms in which the individual capsules are surrounded by a com-
mon jelly mass, as in Amblystoma.

The origin of the follicle cells of Cryptobranchus has been
traced (Part I) from the epithelial cells of the ovarian wall.
Hence these cells belong to the soma, and the rather marked
differences between the follicle cells of Cryptobranchus and
Necturus are of value for comparison only with a very limited
range of forms.

Taking up the history of the egg proper, we first note that
the progressive change of the ovarian egg from analecithal through
an isolecithal to a telolecithal stage is a recapitulation of a very
ancient series of events in the phylogeny. The actual amount
of yolk present is a coenogenetic character, for the yolk content
changes greatly within nearly related forms.

The factors that determine the amount of yolk present in the
eggs of different species are complex and do not readily fall under
any single law. Protection of the eggs through nesting and
brooding habits makes possible a reduction in their number,
enabling the female to endow each egg with a larger store of
yolk, thereby giving the young a better start in life. Such a
store of yolk allows development to go further before the young
animal is cast upon its own resources, so that the necessity for
peculiarly larval adaptations is minimized. In purely terrestrial

JOURNAL OF MORPHOLOGY. VOL. 23, NO 3

554 BERTRAM G. SMITH

non-placental forms a large amount of yolk is certainly necessary ,
for the food (e.g., insect larvae) of the young is larger and less
easy to capture than is the case with aquatic larvae. On the
other hand a large store of yolk is common in the eggs of fishes.
The most we can infer is that the presence of a large amount
of yolk in the egg is one of the conditions that makes possible
the invasion of the land; it cannot be said that the presence of
an unusual amount of yolk in the egg of Cryptobranchus is
evidence either for or against a terrestrial ancestry.

Among the known eggs of urodeles that of Cryptobranchus
is probably the most heavily yolk-laden; the egg of Necturus
contains nearly as much yolk. In the salamandrina one notes
the large amount of yolk in the eggs of Desmognathus (Wilder
04; Hilton ’04 and ’09); Spelerpes (Goodale ’11) and Plethodon
(Piersol ’09).

The absence of pigment in the egg of Cryptobranchus is cor-
related with the nesting habits, whereby the eggs are protected
from the light. Similar conditions are found in Necturus, Des-
mognathus, Plethodon and Spelerpes.

The occurrence of a protoplasmic mantle and cytodise in the
late ovarian egg of Cryptobranchus parallels a condition in the
teleost egg which reaches its full expression just before fertiliza-
tion; in Cryptobranchus this condition is transient. The segre-
gation of a definite layer of cytoplasm close to the surface of the
blastodise in Cryptobrancbus shortly after fertilization suggests
a parallel with the marked increase in thickness of the germinal
disc of the teleost egg immediately after fertilization; Professor
Dean informs me that he has observed a similar phenomenon
in ganoids.

The fundamental features of the early embryonic development
have to do with the building up of the very general structures
and body relations common to all vertebrates. In their most
general aspects cleavage and gastrulation are extremely palin-
genetic phases of development. But we find secondarily imposed
on the essential features of these processes, modifications which
are highly coenogenetic and of adaptive significance mainly
for the embryo and larva (Lillie ’98), rather than for the adult.

EMBRYOLOGY OF CRYPTOBRANCHUS 555

These stages seem to possess few characters intermediate between
the extremes indicated, consequently they tell us little about
the relationships of the larger groups; but the coenogenetic char-
acters, which are plastic and vary widely within a limited range,
may be of value for determining relationships within these groups.
The most striking of these coenogenetic characters are correlated
with the presence of yolk (Conklin ’07).

In the highly telolecithal and beavily yolk-laden egg of Crypto-
branchus, we may interpret holoblastic cleavage as a persistent
primitive character. For were the ancestral form one with mero-
blastic cleavage, we should hardly expect the holoblastic method
to arise under such unfavorable conditions.

Comparisons between the early cleavage furrows of different
urodeles seem justified on the ground that we are comparing cells
of the same generation; but in view of the highly indeterminate
character of the cleavage (Jordan and Eycleshymer ’94) such
comparisons do not take us very far. Mechanical factors
doubtless play a part (McMurrich ’94), but these mechanical
factors are conditioned by the organization of the egg, which is
hereditary.

There is close correlation between the method of third cleavage
and the yolk content. A vertical third cleavage is characteristic
of heavily yolk-laden and highly telolecithal eggs; a latitudinal
third cleavage is found rather in eggs with yolk both smaller
in amount and with a lesser degree of segregation from the cyto-
plasm. Morgan (’93) has shown that in teleost eggs from which
yolk has been experimentally removed, the third cleavage often
comes in latitudinally, yet the eggs produce perfect embryos.
Marked variation in the direction of the third cleavage furrows
occurs in eggs in which the conditions are intermediate in char-
acter; the egg is oscillating between two possible modes of
cleavage. A vertical third cleavage is characteristic of the egg
of Cryptobranchus; it is found less uniformly in the eggs of
Desmognathus; in Necturus and Diemyctylus the third cleav-
age is irregular; in Amblystoma it is latitudinal. So far as this
character is concerned Cryptobranchus lies nearest to Desmog-
nathus and is most remote from Amblystoma.

556 BERTRAM G. SMITH

The important features of the later stages of the cleavage
have to do with processes that are not well expressed in the
superficial cleavage pattern: migration of cells and the various
processes of differentiation that lead up to gastrulation and early
embryo-formation.

The occurrence of a septal furrow in the gastrula stages of two
such widely-separated forms as Cryptobranchus and Petromyzon
is a remarkable case of convergence in purely embryonic charac-
ters. The septal furrow and fenestra are the mechanical product
of gastrulation by invagination and epiboly in a heavily yolk-
laden egg with a very thin roof totheblastocoele. Sofarasknown
these features of gastrulation in Cryptobranchus are unique
among urodeles; but there is evidently an approach to this con-
dition in Spelerpes, since the egg is heavily yolk-laden and during
gastrulation the blastocoele roof becomes quite thin (Goodale’11).

The study of the later embryonic and larval stages is as yet too
superficial to furnish much data for phylogenetic generalization;
yet for this purpose the late stages will probably prove of greater
value than the earlier ones (Wilson ’98, p. 23). It should be par-
ticularly noted that the larval Cryptobranchus reaches an age
of two years before transforming—evidence of a retarded meta-
morphososis. Reasons have already been given for ee that
the metamorphosis is incomplete.

The study of the breeding habits, the organization of the eggs,
and the early course of development lead us to look among the
land-living salamanders for affinities to Cryptobranchus—more
particularly to forms like Desmognathus, Spelerpes and Plethodon.
Considerable evidence from comparative anatomy, particularly
with regard to skull structure, will be found to barmonize with
this view. Desmognathus in particular is suggestive; according
to Kingsbury (’02) it is semi-aquatic in its habits, living at the
edges of swiftly running brooks. It conceals itself under stones
at the edge of the stream or in its immediate vicinity, and here
its unpigmented and heavily yolk-laden eggs are laid. There
is a brooding habit, though in this case the female guards the
eggs; we have noted some marked similarities to Cryptobranchus
in the early development. Yet we are hardly warranted in con-

EMBRYOLOGY OF CRYPTOBRANCHUS 557

cluding that the relationship is very close. In particular we
should expect the immediate ancestral stock of Cryptobranchus
and the closely related miocene fossil Andrias scheuchzeri (Gadow
’01, p. 84) to consist of larger animals.

According to the view adopted in this paper, the urodeles,
very remotely descended from aquatic stock, are primarily ter-
restrial, but with aquatic larvae. On the land they were unsuc-
cessful in the struggle for existence in the open; they took refuge
in sheltered situations, and for this they have paid the penalty
of degeneration. Yet, in the main, the result of the arrest of a
typical terrestrial adaptive radiation has been the retention of.
primitive characters. Some became secondarily aquatic; this
is one phase of the tendency toward secluded habits, and involves
the retention of larval ¢haracters. Added to these more conspic-
uous peculiarities we find a great variety of special adaptations to
a retired mode of life, some of them correlated with the defense-
less condition of the animals. Of all these changes, reversion
to an aquatic mode of life is a factor which, cutting across many
lines of descent, has done most to disguise the real relationships,
and of this we have a conspicuous example in Cryptobranchus.

558 BERTRAM G. SMITH

BIBLIOGRAPHY

A few titles are included that are not directly referred to in the text.

Agassiz, L., AND WuiTmMan, C. O. 1884 On the development of some pelagic
fish eggs. Proc. Amer. Acad. Arts and Sciences, vol. 20.

ASSHETON, RicHarD 1896 Notes on the ciliation of the ectoderm of the amphib-
ianembryo. Quart. Jour. Micros. Sci., vol. 38.

Baur, G. 1896 The stegocephali. A phylogenetic study. Anat. Anz., Bd. 11.

Buount, Mary 1907 The early development of the pigeon’s egg, with especial
reference to the supernumerary sperm nuclei, the periblast and the
germ wall. Biol. Bull., vol. 13.

Boas, J. E. V. 1881 Ueber den Conus Arteriosus und die Arterienbogen der
Amphibien. Morph. Jahrb., Bd. 7.

; 1896 Ueber Neotenie. Festschr. f. Carl Gegenbaur, Bd. 2.

Braver, Aucust 1899 Beitrige zur Kenntniss der Entwicklung und Anatomie
der Gymnophionen. II. Die Entwicklung der aussern Form. Zool.
Jahrb., Abth. Anat., Bd. 12.

Bupcert, JoHN Samuret 1901 On the breeding habits of some West African
fishes, with an account of the external features in the development
of Protopterus annectens, and a description of the larvae of Polypterus
lapradei. Trans. Zool. Soc. London, vol. 16, part 2, August. Also
Budgett memorial volume, part VI (Kerr ’07).

Conkuin, E. G. 1907 The embryology of Fulgur: A study of the influence of
yolk on development. Proc. Phila. Acad. Sci., July.

Corr,E.D. 1885 The retrograde metamorphosis of Siren. Amer. Nat., vol. 19.
1889 The batrachia of North America. Bull. U. 8S. Nat. Mus., no. 34.

CREDNER, HERMANN’ 1881-1890 Die Stegocephalen und Saurier aus dem roth-
liegenden des Plauen’schen Grundes bei Dresden. R. Friedlander
und Sohn, Berlin.

Dean, Basnurorpd 1895 Early development of gar-pike and sturgeon. Jour.
Morph., vol. 11.

1896 The early development of Amia. Quart. Jour. Micros. Sci., vol. 38.
pE LanGcgE, Dan., Jr. 1907 Die Keimblitterbildung des Megalobatrachus max-
imus Schlegel. Anat. Hefte, Bd. 32.

pE Bussy, L. P. 1904 Eerste Ontwikkelungsstadién van Megalobatrachus
maximus Schlegel. Tijsdchrift der nederlandsche Dierkundige Ver-
eeniging, Bd. 8.

1905 Die ersten Entwicklungsstadien des Megalobatrachus maximus.
Zool. Anz., Bd. 28.

Eaceiine, H. von 1911 Der Aufbau der Skeletteile in den freien Gliedmassen
der Wirbeltiere. Untersuchungen an urodelen Amphibia. Gustav
Fischer, Jena. (Reviewed by Bardeen in Science, September 29).

Emerson, E. T. 1905 General anatomy of Typhlomolge rathbuni. Proc.
Boston Soc. Nat. Hist., vol. 32.

EycLesHyMer, ALtBerT C. 1895 The early development of Amblystoma, with
observations on some other vertebrates. Jour. Morph., vol. 10.
1898 The location of the basis of the amphibianembryo. Jour. Morph.,
vol. 14.

EMBRYOLOGY OF CRYPTOBRANCHUS 559

EYcLESHYMER, ALBERT C. 1899 The cleavage of the egg of Lepidosteus osseus.
Anat. Anz., Bd. 16.

1902 The formation of the embryo of Necturus, with remarks on the
theory of concrescence. Anat. Anz., Bd. 21.
1904 Bilateral symmetry in the egg of Necturus. Anat. Anz., Bd. 25.

EyYcLEsSHYMER. A. C., and Wiuson, James M. 1910 Normal plates of the devel-
opment of Necturus maculosus. Keibel, Normentafeln zur Entw.
der Wirbeltiere, Heft 11. Gustav Fischer, Jena.

Gapow, Hans 1901 Amphibia and reptiles. Cambridge Nat. Hist., vol. 8.

GormTtTE, ALEXANDER 1875 Die Entwicklungsgeschichte der Unke. Mit einem
Atlas. Leopold Voss, Leipzig.

G6éup1, Emit A. 1899 Uber die Entwicklung von Siphonops annulatus. Zool.
Jahrb., Abth. Syst., Bd. 12.

GoopaLE, H. D. 1911 The early development of Spelerpes bilineatus Green.
Amer. Jour. Anat., vol. 12.

GoopricH, E. 8. 1909 Vertebrata craniata. Part IX of Lankester’s treatise
on zoology. Adam and Charles Black, London.

Greoory, W. K. 1907 The orders of teleostomous fishes. A preliminary review
of the broader features of their evolution and taxonomy. Annals N.
NWerAcaderscies volumes leenonmos
191la The limbs of Eryops and the origin of paired limbs from fins.
Secretary’s report of the meeting of the New York Academy of Sciences,
Section of Biology, Feb. 13. Science, vol. 33, Mar. 31.
1911b Further notes on the evolution of paired fins. Secretary’s
Report of the meeting of the New York Academy of Sciences, Section
of Biology, Oct. 16. Science, vol. 34, no. 886, Dec. 22.

1912 The origin of paired limbs. Secretary’s report of the meeting
of the New York Academy of Sciences, Seni of Biology, Dec. 11,
1911. Science, vol. 30, Apr. 12.

Grices, Leranp 1910 Early stages in the development of the central nervous
system of Amblystoma punctatum. Jour. Morph., vol. 21, no. 3.

Harta, 8. 1907 Gastrulation in Petromyzon. Jour. Coll. Sci., Tokyo.

Hertwic, Oscar 1906 Handbuch der vergleichenden und experimentellen
Entwicklungslehre der Wirbeltiere. Gustav Fischer, Jena.

Hitt, CHartes 1899 Primary segments of the vertebrate head. Anat. Anz.,
Bd. 16.

1900 Developmental history of the primary segments of the vertebrate
head. Zool. Jahrb., Abth. fiir Anatomie und Ontog., Bd. 18.

Hinton, Wm. A. 1904 Segmentation of the ovum of Desmognathus fusca. Amer.
Nat., vol. 38.

1909 General features of the early development of Desmognathus
fusca. Jour. Morph., vol. 20.

Ikepo, 8S. 1902 Contributions to the embryology of the amphibia: The mode
of blastopore closure and the position of the embryonic body. Jour.
Coll. Sci., Tokyo, vol. 17, Part IT.

{surkawa, C. 1904 Beitriige zur Kenntniss des Riesen-Salamanders (Megalo-
batrachus maximus Schlegel). Proc. Dept. Nat. Hist., Tokyo Impe-
rial Museum, vol. 1, no. 2.

560 BERTRAM G. SMITH

IsHikawa, C. 1908 Ueber den Riesen-Salamander Japans. Mitteilungen der
deutschen Gesellschaft fiir Natur-und Vélker-kunde: Ostasiens. Bd.
ieee
1909 Note on the gastrulation of the giant salamander, Megaloba-
trachus sieboldii. Proceedings of the seventh international zoological
congress, Boston, Aug. 19-24, 1907.

JorDAN, Epwin O. 1893 The habits and development of the newt. Jour. Morph.,
vol. 7.

JorpDAN, E. O. and EyctesHyMER, A.C. 1894 On the cleavage of the amphibian
ovum. Jour. Morph., vol. 9.

Kewuiicott, W. E. 1905a The development of the vascular and respiratory
systems in Ceratodus. N. Y. Acad. Sci. Memoirs, vol. 2, Part TV.
1905b The development of the vascular system of Ceratodus. Anat.
Anz., vol. 26.

Kerr, J. GRAHAM 1900 External features in the development of Lepidosiren.
Philos. Trans..Royal Soc., vol. 192, B.

1901 The development of Lepidosiren paradoxa. II. With a note
upon the corresponding stages in the development of Protopterus annec-
tens. Quart. Jour. Micros. Sci., vol. 45.

1907a The development of Polypterus senegalus Cuv. Budgett mem-
orial volume (Kerr, ’07b).

1907b . The work of John Samuel Budgett. Cambrige Univ. Press.
1909 Normal plates of the development of Lepidosiren paradoxa
and Protopterus annectens. Keibel, Normentafeln zur Entw. der
Wirbeltiere, Heft 10. Gustav Fischer, Jena.

Kinessury, B. F. 1902 The spermatogenesis of Desmognathus fusca. Amer.
Jour. Anat., vol. 1.

1905 The rank of Necturus among tailed batrachia. Biol. Buil., vol.

Kinessury, B. F. and Reep, H. D. 1909 ‘The columella auris in amphibia.
Jour. Morph., vol. 20.

Kinasuey, J. 8. 1899 Text-book of vertebrate zoology. Henry Holt and Co.,
New York.

1901 Amphibian studies. Abstract of paper read before the American
Morphological Society. Biol. Bull., vol. 2.

iKunitomo, Kanar 1910 Ueber die Entwicklungsgeschichte des Hynobius
nebulosus. Anat. Hefte, Beitr. und Ref. Anat. und Entw., Bd. 40.

Kurprrer, C. 1885 Primire Metamerie des Neuralrohrs der Vertebraten. Sit-
zungsber. der mathem.-physik. Kl. der K. B. Akad. d. W. Miinchen.
1892 Entwickelungsgeschichte des Kopfes. Merkel und Bonnet,
Ergebnisse der Anatomie, Bd. 2.

1893 Studien zur vergleichenden Entwicklungsgeschichte des Kopfes
der Kranioten. Heft 1. Miinchen und Leipzig.

1903-1905 Die Morphologie des Centralnervensystem. Handbuch
der vergl. und exp. Entwickelungslehre der Wirbeltiere, Hertwig, ’06.

Linuiz, Frank R. 1898 Adaptation in cleavage. Woods Hole biological lec-
tures, 1897-1898.

1908 Development of the chick. Henry Holt and Co.

EMBRYOLOGY OF CRYPTOBRANCHUS 561

Locy, W. A. 1895 A contribution to the structure and development of the ver-
tebrate head. Jour. Morph., vol. 11.

McCiEnpDOoN, J. F. 1910 The development of isolated blastomeres of the frog’s
egg. Amer. Jour. Anat., vol. 10.

McGreeor, J. H. 1897 An embryo of Cryptobranchus. Anat. Anz., Bd. 13.

McMorricu, J. P. 1894 Cell division and development. Woods Hole biologi-
cal lectures.

Moopiz, Roy L. 1908a The lateral line system of extinct amphibia. Jour.
Morph., vol. 19.
1908b The ancestry of the caudate amphibia. Amer. Nat., vol. 42.

Moraan, T. H. 1893 Experiments on teleost eggs. Anat. Anz., Bd. 8.

1904 Die Entwicklung des Froscheies. Wilhelm Engelmann, Leipzig.

Orr, H. 1887 A contribution to the embryology of the lizard. Jour. Morph.,
vol. 1.

Ossporn, H. F. 1888 A contribution to the internal structure of the amphibian
brain. Jour. Morph., vol. 2.

1902 The law of adaptive radiation. Amer. Nat., vol. 36.

Parker, W.K. 1877 On the structure and development of the skull in the uro-
delous amphibia. Part I. Philos. Trans. Royal Soc., vol. 167, Part IT.

ParkKER, T. JEFFERY, and HASWELL, WM. A. 1897 A text-book of zoology, vol.
2. The Macmillan Co.

Pizrsot, W. H.*1909 The habits and larval state of Plethodon cinereus ery-
thronotus. Trans. Canadian Institute, vol. 8, Part IV.

Reep, H.D. 1909 Systematic relations of the urodela as interpreted by a study
of the sound-producing organs. Science, Apr. 30.

Reese, Albert M. 1906 Anatomy of Cryptobranchus allegheniensis. Amer.
‘Nat., vol. 40. i

Roginson, ARTHUR, and ASSHETON, RicHAarD 1891 The formation and fate
of the primitive streak, with observations on the archenteron and germ-
inal layers of Ranatemporaria. Quart. Jour. Micros. Sci., N.8., vol. 32

ScHuLTzE, O. 1895 Die kiinstliche Erzeugung von Doppelbildung bei Frosch-
Larven mit Hilfe abnormer Gravitationswirkung. Archiv Entwm.,
Bd. 1.

1900 Ueber das erste Auftreten der bilateralen Symmetrie im Verlauf
der Entwickelung. Archiv fiir mikr. Anat., Bd. 55.

Semon, R. 1900 ‘Die Furchung und Entwicklung der Keimblitter bei Ceratodus
forsteri. Aus Semon, zoologische Forschungensreisen in Australia
und dem Malayischen Archipel, Bd. 4. Gustav Fischer, Jena.

1901 Normentafeln zur Entw. des Ceratodus forsteri. Keibel, Nor-
mentafeln zur Entw. der Wirbeltiere, Heft 3. Gustav Fischer, Jena.

Smitu, Bertram G. 1906 Preliminary report on the embryology of Crypto-
branchus allegheniensis. Biol. Bull., vol. 11.

1907 a The life history and habits of Cryptobranchus allegheniensis.
Biol. Bull., vol. 18.

1907 b The breeding habits of Amblystoma punctatum Linn. Amer.
Nat., vol. 41.

1911 a The nests and larvae of Necturus. Biol. Bull., vol. 20.

562 * BERTRAM G. SMITH

SmirH, Bertram G. 1911 b Notes on the natural history of Amblystoma
jeffersonianum, A. punctatum and A. tigrinum. Bull. Wis. Nat. Hist.
Soc., vol. 9.
1912 The embryology of Cryptobranchus allegheniensis, including
comparisons with some other vertebrates. Part 1: Introduction; the
history of the egg before cleavage. Jour. Morph., vol. 23.

SpeMann, H. 1901-1903 Entwickelungsphysiologische Studien am Triton-Hi.
I, II, III. Archiv Entwm., 12, 15, 16.

Stannius, H. 1856 Handbuch der Zootomie. Berlin.

StrockarD, CHaArRues, R. 1906 The development of Fundulus heteroclitus in
solutions of lithium chlorid, with appendix on its development in fresh
water. Jour. Exp. Zool., vol. 3.

Verstuys, J. 1909 Die Salamander und die urspriinglichsten vierbeinigen
Landwirbeltiere. Naturwissenschaftlichen Wochenschrift, Neue Folge,
BdeSa Nias:

WataskE, S. 1891 Studies on cephalopods. Jour. Morph., vol 4.

Wuippue, Inez L. 1906 The ypsiloidapparatus of urodeles. Biol. Bull., vol. 10.

WuitMan, C. O. and EyctesHyMer, A. C. 1897 The egg of Amia and its cleav-
age. Jour. Morph., vol. 12.

WrepEerRsHEIM, R. 1877 Das Kopfskelett der urodelen. Morph. Jahrb., Bd. 3.

Wiper, Harris HAWTHORNE 1899 Desmognathus fusca and Spelerpes bilineat-

us. Amer. Nat., vol. 33.
1904 The early development of Desmognathus fusca. Amer. Nat.,
vol. 38.

Wiutson, Cuas. B. 1896 The wrinkling of the frog’s eggs during segmentation.
Amer. Nat., vol. 30.

Writson, E. B. 1898 Cell lineage and ancestral reminiscence. Woods Hole
biological lectures.

Witson, H. V. 1891 The embryology of the sea bass (Serranus). Bull. U.S.
Fish Comm., vol. 9.

ZrrveL, Karu A. von 1911 Grundziige der Paliontologie. R. Oldenbourg
Miinechen und Berlin. :

PLATE 3

EXPLANATION OF FIGURES
(Cryptobranchus allegheniensis)

All the figures are from preserved material, fixed in bichromate-acetic-formalin
excepting the eggs shown in figures 210 and 212 which were fixed in formalin. The
animal hemisphere is shown in every case. X 4. ;

204 Stagel. First cleavage.

205 Stage2. Second cleavage.

206 Stage 3. Third cleavage.

207 Stage 4. Fourth cleavage.

208 Stage 4. Fourth cleavage. Figure 82 is drawn from the same egg.

209 Stage 4. Fourth cleavage. Figure 81 is drawn from the same egg.

210 Stage 5. Thirteen micromeres. Figure 87 is drawn from the same egg.

211 Stage 5. Seventeen micromeres. Figure 86 is drawn from the same egg.

212 Stage 6. About thirty-two micromeres. Figure 98 is drawn from the
same egg.

213 Stage 6. About thirty-two micromeres.

214 Stage6. About thirty-six micromeres.

215 Stage 7. About sixty-four micromeres.

564

EMBRYOLOGY OF CRYPTOBRANCHUS PLATE 3

BERTRAM G. SMITH

205 206

208 209

213

THE JOURNAL OF MORPHOLOGY, VOL. 23, NO. 3

PLATE 4

EXPLANATION OF FIGURES

(Cryptobranchus allegheniensis)

All the figures are from preserved material fixed in bichromate-acetic-forma-
lin: - x4.

216 Stage 7.5. Equatorial view.
217 Stage 8.0. Upper hemisphere.
218 Stage 9.5. Equatorial view.

On

219 Stage 10.5. Equatorial view of an embryo nearly ready for gastrulation.

220 Stage 11. Early gastrula, showing the dorsal lip of the blastopore.

221 Stage 11.5. Antero-dorsal view of a gastrula, showing the fenestra.

222 Stage 11.6. Anterior view of a gastrula, showing the fenestra.

223 Stage 12.4. Dorsal view showing early neural plate and neural groove.

224 Stage 12.4. Posterior view of the same embryo as in the preceding
figure, showing the yolk plug and a part of the neural plate and neural groove.

225 Stage 12.5. Dorsal view showing the neural groove and the dorsal lip
of the blastopore. Figure 143 is drawn from the same embryo.

226 Stage 13.4. Dorsal view showing early neural folds.

227 Stage 13.5. Posterior view showing the form of the late blastopore, and
the posterior part of the neural plate.

sa

EMBRYOLOGY OF CRYPTOBRANCHUS PLATE 4

BERTRAM G. SMITH

218

222 223 22¢

225 226 227

567

PLATE 5

EXPLANATION OF FIGURES

(Cryptobranchus allegheniensis)

Bichromate-acetic-formalin fixation. > 4.

228 Stage 13.5. Posterior view showing the form of the late blastopore, and
the posterior part of the neural plate.

2299 Stage 14.1. Dorsal view showing neural folds and the early segmentation
of the neural plate. Figure 166 is drawn from the same embryo.

230 Stage 14.1. Postero-dorsal view of the same embryo as in the preceding
figure, showing the late blustopore and the segmentation of the neural plate.

231 Stage 14.9. Posterior view, showing the late blastopore. Figure 176
is drawn from the same egg.

232 Stage 14.9. Dorsal view of the same embryo as in the preceding figure,
showing the segmentation of the neural plate.

233 Stage 15.4. Dorsal view. Figure 184 is drawn from the same egg.

234 Stage 15.8. Dorsal view. Figure 182 is drawn from the same egg.

235 Stage 16.0. Dorsal view showing closing neural folds.

236 Stage 16.8. Dorsal view.

237 Stage 16.8. Antero-dorso-lateral view of the same embryo as in the pre-
ceding figure.

238 Stage 17. Antero-dorsal view (inverted with respect to the natural posi-
tion).

939 Stage 17. Lateral view.

568

EMBRYOLOGY OF CRYPTOBRANCHUS PLATI

BERTRAM G. SMITH

228 230

Pane fd 233

234 Fas Fas:

239

569

PLATE 6

EXPLANATION OF FIGURES

(Cryptobranchus allegheniensis)

Bichromate-acetic-formalin fixation. 4.

240
241
242
243
244
245
246
247
248
249

Stage 17.
Stage 17.
Stage 17.
Stage 18.

Stage 18.
Stage 19.

Stage 19.
Stage 19.

Stage 19.

Stage 20.

Ventral view of the embryo shown in the preceding figure.
Dorso-lateral view of the embryo shown in the preceding figure.
Dorsal view of the embryo shown in the preceding figure.
Ventral view.

Lateral view of the embryo shown in the preceding figure.
Lateral view.

Dorsal view.

Lateral view of the embryo shown in the preceding figure.
Lateral view.

Lateral view.

570

EMBRYOLOGY OF CRYPTOBRANCHUS PLATE 6

BERTRAM G. SMITH

247

248 249

PAWS 7

EXPLANATION OF FIGURES

(Cryptobranchus allegheniensis)

Bichromate-acetic-formalin fixation. 4.

250
251

D259)

Dye
253
25

255
256
257

ol

Stage
Stage
Stage
Stage
Stage
Stage
Stage
Stage

20.
20.
Zi.
21.
21.
VANE
22.

99
Ziad s

Lateral view.

Ventral view of the embryo shown in the preceding figure.
Ventral view.

Lateral view.

Dorsal view of the embryo shown in the preceding figure.
Ventral view of the embryo shown in the preceding figure.
Dorsal view.

Ventral view of the embryo shown in the preceding figure.

EMBRYOLOGY OF CRYPTOBRANCHUS PLATE 7

BERTRAM ¢ SMIIH

PLATE 8

EXPLANATION OF PLATES
(Cryptobranchus allegheniensis)

Bichromate-acetic-formalin fixation except for figures 260 and 261 which were
made from living specimens anaesthetized with chloretone. 4.

258
259
260
261

Stage 22. Lateral view.

Stage 22.5. Lateral view.

Stage 23. Dorsal view of an embryo ready to hateh.

Stage 23. Lateral (slightly ventral) view of a newly-hatched larva.

EMBRYOLOGY OF CRYPTOBRANCHUS PLATE 8
BERTRAM G. SMITH

Z256

229

260

261

miedo

040

PLATE 9

EXPLANATION OF FIGURES

(Cryptobranchus allegheniensis)

All the figures are natural size.

262 Living larvae reared in the laboratory, anaesthetized with chloretone
and photographed about five weeks after hatching.

263. Larva reared in the laboratory, killed in bichromate-acetic-formalin
about two months after hatching and preserved in formalin. The specimen is
slightly shortened through the action of the fixing fluid.

264 and 265. Two viewsof a larvareared in the laboratory, anaesthetized with
chloretone and photographed about ten weeks after hatching.

266. Larva reared in the laboratory, anaesthetized with chloretone and pho-
tographed about six months after hatching.

267. Year-old larva captured and fixed in formalin August 27, 1909; a few weeks
later it was transferred to alcohol, in which it remained nearly a year before being
photographed.

KMBRYOLOGY OF CRYPTOBRANCHUS PLATE 9
BERTRAM G. SMITH

263
264
Z2oes

266

2o7

577

PLATE 10

EXPLANATION OF FIGURES

(Necturus maculosus)

The figures are intended to show in particular the late history of the blastopore.
Figures 268 to 276 inclusive were drawn from the living eggs by Prof. Bashford
Dean; figures 277 to 279 are from preserved material, and were drawn under the
direction of the author by Miss Mabel L. Hedge.

578

EMBRYOLOGY OF CRYPTOBRANCHUS PLATE 10
BERTRAM G. SMITH

Neclurus maculosus

268 269 270

2. ="
, J
‘ ts
‘
mr
t
i
«
¢
4
ry
;
on \
« Fr 5 ‘2
.
'
f
= >
oF n
‘y ’
a f
a) ;
‘ af ‘ ay '
. ? _ nf
7 ‘ 4 4 ;
P ne
; '
oun ?
J r
e J ‘ t ee r i
, e'Y i %
* 4 :
" i ee he

hae

. ia Ae : melts ie
“e

> Ae

an oe m4 j '

:
i
| ers
ee
aw
a
=_¢
>
ix
»
*
i

= at eg
Pes

>

THE STRUCTURE AND METAMORPHOSIS OF THE
FORE-GUT OF CORYDALIS CORNUTUS L.

ROBERT MATHESON

From the Entomological Laboratory, Cornell University

TWENTY-NINE FIGURES

Owing to the large amount of work that has been done and is
still being done on the metamorphosis of the more specialized
insects it has seemed advisable to make a careful survey of some
generalized form. It was with this object in view that the pres-
ent work, though dealing with a restricted region, was undertaken.
The writer hopes to continue these studies as time permits.

The present subject was suggested by Prof. W. A. Riley and
the work was done under his immediate supervision. ‘To Pro-
fessors W. A. Riley and J. H. Comstock I wish to extend my
thanks for advice and criticism.

STRUCTURE

The fore-gut of Corydalis extends as practically a straight,
narrow tube from the mouth to the middle of the second abdomi-
nal segment where it joins the mid-intestine (fig. 1). There are
but two slight enlargements, one at the anterior end, the pharynx,
and a more considerable one near the posterior end, the gizzard
(fig. 1, ph.andk). The caudal end projects into the mid-intestine
to form a characteristic oesophageal valve. Morphologically it
may be divided into five well marked regions: pharynx, oesopha-
gus, gizzard, the region between the gizzard and oesophageal
valve, and the oesophageal valve. Along each side of the fore-
intestine runs a branch of the visceral nervous system (fig. 1, 7)

while various branches of the tracheal system furnish the neces-
sary air supply.
581

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4
DECEMBER 1912

582 ROBERT MATHESON

The tracheal supply of the fore-gut

The distribution of the tracheae to the alimentary canal may be
readily seen in figure 1. From the second and third abdominal
tracheal gills there arise, just previous to their union with the longi-
tudinal trunks, two comparatively large branches which go directly
to the fore-gut. The branch (6, fig. 1) from the second tracheal
gill first divides into two to five branches which, with their many
ramifications, supply the anterior portion of the gizzard and
oesophagus. However, there is much variation in the distribution
of the tracheal supply in different individuals. In some (fig. 1, 6)
the first branch breaks up into two branches each of which finally
divides into a large number of smaller tracheae distributed over
the anterior portion of the gizzard while one small branch (tr,
fig. 2) runs forward along the side of the oesophagus giving off to
it numerous small tracheae. This seems to be the more common
method of distribution of this trunk. If, however, figure 2 be
examined, it is easily seen how greatly modified this may be.
Here the first tracheal trunk, b, divides into five branches, the
posterior one of which is distributed over the caudal portion of
the gizzard while the remainder furnish the usual tracheal supply
to the oesophagus and gizzard. Even in the same individual
‘the branching and distribution of this trunk varies considerably
on the opposite sides.

The second tracheal trunk (c) arises from the third abdominal
gill (fig. 1, 3 ab). It divides just before reaching the intestine
into two branches. The posterior branch supplies the ventral
caecum as seen on the right side of figure 1. The anterior branch
passing forward on the ventral side of the caeca breaks up into
five small branches, each of which in turn divides into numerous
smaller branches. ‘These furnish the tracheal supply of the pos-
terior portion of the gizzard and that portion of the alimentary
canal between it and the mid-gut. There is considerable varia-
tion in. the branching and distribution of this trunk even on the
opposite sides of the same individual. On the left side of figure
1 it may be seen that the large trunk coming from the tracheal
gill gives off two branches instead of one to the ventral caecum.

FORE-GUT OF CORYDALIS CORNUTUS L. 583

The main trunk passing forward first gives off another small
branch which also goes to the caecum. Then it divides into
two branches. The anterior one, breaking up into three small
branches, supplies the caudal portion of the gizzard. The pos-
terior branch divides into three branches which send numerous
small tracheae to the caudal part of the gizzard, the esophageal
valve and the ventral caecum.

The third tracheal trunk (fig. 1, d) arises at the base of the
transverse dorsal trachea of the third abdominal segment. It
proceeds forward and divides into two branches, one of which, pass-
ing in between the caeca, supplies them with numerous branches
and also send tracheae to the fore-gut between the gizzard
and mid-gut. The other branch passes to the dorsal side of the
dorsal caecum and furnishes it with a tracheal supply. It also
sends numerous small branches to the anterior portion of the mid-
gut and the region of the fore-gut behind the gizzard.

The fourth tracheal trunk (fig. 1, e) arises in a way similar to
the second—from the base of the fourth tracheal gill. It passes
forwards and divides into two branches. The posterior one sup-
plies the ventral and lateral surfaces of the mid-gut. The
anterior one divides into many small branches which supply the
anterior and lateral portions of the mid-gut as well as the proxi-
mal ends of the caeca and caudal part of the fore-gut.

The fifth tracheal trunk (fig. 1, g) arises similarly to the third.
Directly at the point of origin or very near it, there is given off
_a large branch (in some cases two) which goes to the testis. This
branching frequently varies on the opposite sides of the same indi-
vidual. The main branch proceeds forwards and divides just
before reaching the digestive tract. The two branches supply
the anterior dorsal and lateral portions of the mid-gut as well as
the proximal ends of the dorsal caeca and posterior end of the
fore-gut.

A sixth trunk (fig. 1, f) arises from the fifth tracheal gill. It
divides into two branches which supply the posterior end of the
mid-gut.

From the photograph (fig. 1) and the above description it will
be seen that the gizzard and mid-gut are well supplied with a

584 ROBERT MATHESON

tracheal system which is in direct connection with the rich in-
coming air supply. The oesophagus requires only a very small
amount of oxygen and consequently has a poorly developed tra-
cheal system.

The pharynx and its muscles are supplied by a small tracheal
branch (fig. 1, a) which arises from the exterior longitudinal
trachea at its point of entrance into the head. Several other
small tracheae from the anterior branches of the main longitu-
dinal trunks supply the cephalic end of the pharynx and mouth-
parts.

The pharynx

That portion of the fore-gut, extending from the mouth to
near the posterior margin of the head, may be designated the
pharynx (fig. 1, ph). It is provided with special muscles and,
when expanded, is somewhat larger than the oesophagus. On
the latero-ventral surface on each side, just behind: the sub-
oesophageal ganglion, are situated two series of small muscles.
These extend in a latero-caudad direction to their points of origin
on the posterior margin of the head (fig. 16). The number of
these muscles varies, the usual number on each side being from
16 to 19. They lie directly ventrad of the large tracheae which
penetrate the head. Directly in front of these lateral-inferior _
muscles and on the ventral surface there is a series of three to
four small muscles on each side of the median line, which pass
almost directly ventrad to a short chitinous apodeme near the
labial region.

On the dorsal side of the Siete directly above the lateral
series of muscles, are found two series of small muscles on each
side of the median line (fig. 15). The inner series includes
five, the outer four, muscles. They are attached on each side of
the middle line of the caudal portion of the vertex (fig. 15).

The function of the pharyngeal muscles seems to be that of
supporting and enlarging the pharynx during feeding. On open-
ing a freshly etherized specimen it is easy to watch their action.
The pharynx in this region is seen gradually to expand, then a
wave-like contraction begins at the mouth and passes down the

FORE-GUT OF CORYDALIS CORNUTUS L. 585

‘oesophagus to the gizzard. This contraction is brought about by
the circular and longitudinal muscles. On cutting the lateral
muscles of the pharynx the expansion preceding the wave-like
contraction ceases, although the latter movement continues.

The oesophagus

The oesophagus, except for a gradual increase in size, extends
as a straight, cylindrical tube from the pharynx to the middle
of the second thoracic segmertt where it passes into the gizzard
(fig. 1, k). Viewed externally it presents no important structural
characters. The arrangement of the circular and longitudinal
muscles, to be discussed more in detail later, can readily be ob-
served in the freshly opened larva (fig. 3). On each side of the
oesophagus runs a large nerve (fig. 3, ), a branch of the so-called
visceral nervous system. Close to each nerve a tracheal branch
runs forward as previously described in discussing the respiratory
system.

In histological structure the oesophagus, as in all insects so
far studied, presents the usual three layers: (1) the epithelium
with its intima; (2) the basement membrane lying directly
beneath the epithelium; (3) the muscular layer consisting of cir-
cular and longitudinal fibers.

The epithelium with its intima and the basement membrane
are thrown into numerous irregular folds (fig. 8). At times these
folds almost completely close the lumen. The folding allows
considerable contraction or distension of the oesophagus and does
not, I think, function in the reduction of the food. The resiliency
of the intima aids materially in the distension of the oesophagus
for the reception of food.

The epithelium consists of a thin layer of flat cells with boun-
daries very indistinct or lacking. The nuclei are oval in outline,
with their long axes usually parallel to the basement membrane
(fig. 17). The cytoplasm presents a granular appearance. This
part of the fore-gut may secrete a digestive fluid, as Plateau
(74) found in several species of insects examined by him but I
found no evidence of it. Its chief function seems to be the for-

586 ROBERT MATHESON

mation of a well defined and somewhat extensive intima which’
lines the interior of the oesophagus and so forms an excellent
resilient tube down which the food may pass. A basement mem-
brane is always present.

The muscles, as shown in figure 8, present an arrangement well
fitted for the work they perform—the transfer of the food from
the pharynx to the gizzard. Next the basement membrane are
the longitudinal muscles, which are so arranged that the fibers
all run obliquely, crossing each other at a rather acute angle and
by this means giving them the effect of two oblique cylinders
one woven into the other. These fibers in the region of the giz-
zard are gathered up into six bundles which pass under the large
longitudinal folds (fig. 3) to their points of insertion on the bases
of the anterior teeth. Lying directly external to the longitudinal
muscles are the circular muscles, small and few, yet with an appar-
ently definite arrangement (figs. 3 and 8, c.m.). In the region in
front of the gizzard the circular muscles increase rapidly in size
and number while from the middle of the gizzard to its caudal end
they appear as a dense, thick mass.

From the arrangement of the muscles their action may be read-
ily assumed without actual experimental observation on the living
animal. On the reception of food the gradual contraction would
force it onward and their relaxation, aided by the resiliency of the
intima, would distend the lumen for the intake of more food.
Thus there would be a somewhat peristaltic action throughout
the length of the oesophagus. This peristaltic action has been
observed in the living specimen and described in discussing the
pharynx.

The gizzard

The cesophagus passes into the gizzard without any marked
constriction. Viewed externally the gizzard presents abarrel-
shaped appearance. The only outward means of locating its
beginning is by a slight enlargement and an increase in the thick-
ness of the circular muscles. The circular muscles are most
greatly increased directly in front of the middle of the gizzard.
This increase occurs at the point where the teeth of the gizzard

FORE-GUT OF CORYDALIS CORNUTUS L. 587

are most heavily chitinized and meet, almost completely closing
the lumen. The lumen can readily be closed by the concerted
action of the circular muscles and the longitudinal muscles
directly attached to the teeth.

Internally the gizzard presentsaremarkable appearance (fig. 4).
It is provided with six large, inward-projecting, chitinous ridges
and alternating with these are six smaller ridges (fig. 4, I.r., s.r.).
Each large ridge consists of two prominent teeth, an anterior
(a.t., figs. 5 and 6) and a posterior one (p.t.), separated by a deep
constriction. The anterior tooth projects caudad, the proximal
part being somewhat narrow and sharp along its inner edge. It
rapidly becomes broader, more strongly chitinized and fits closely
over the base of the posterior tooth. Its entire surface is clothed
with setae which are longer and more numerous along the poste-
rior border.

Though scarcely distinguishable in surface view the posterior
teeth are of two types (compare figs. 5 and 6). They alternate
with each other, three of each in a gizzard. The one shown in
figure 5 appears in side view somewhat like a pistol, the internal,
posterior projecting process answering to the hammer. Viewed
posteriorly it presents a broad, somewhat hollowed surface, nar-
rowed on its anterior margin into a sharp tooth projecting caudad.
The other type is shown in figure 6 and differs from the first in
the lack of the anterior process.

The small longitudinal ridges (figs. 4 and 7, s.r.) project
between the larger ridges, almost completely filling the interven-
ing spaces. They are not divided into teeth, though at their
posterior ends each one presents a broader, setiferous portion
projecting caudad.

The epithelium of the gizzard consists of a layer of flattened
cells with fairly well defined cell walls. In the tips and sides
of the longitudinal ridges the cells are crowded and narrowly
elongate. The nuclei are oval to spherical with densely staining
granules. The cytoplasm is granular and more or less strongly
vacuolate in the cells lining the ridges. The vacuoles are located
at the bases of the cells. The longitudinal ridges are filled with a
non-cellular tissue (fig. 9, st) which is usually designated as a

588 ROBERT MATHESON

sort of connective tissue. It is penetrated with tracheae, and
tracheal end cells with their nuclei are usually found distributed
through it. It does not show any of the characteristics of verte-
brate connective tissue when stained by the differential stains
used by histologists. It may be analogous to the tissue formed in
the developing vitreous body of the vertebrate eye. Here recent
researches indicate that it is derived from the prolongations of the
cells of the retina, therefore ectodermal in origin. In the case of
Corydalis this so-called connective tissue stains in the same way
as the basement membrane and one can readily trace out into it
prolongations of the epithelial cells. The vacuolate condition in
the cells near the basement membrane indicates a secretion. If
the basement membrane be considered as a product of the epi-
thelial cells I think there is no evidence to show that this supposed
connective tissue is not also a secretion of the epithelium. This
tissue is widely distributed throughout the intestine and may
be referred to under the non-committal term of supporting tissue.

The fore-gut between the gizzard and the oesophageal valve

At its posterior end the gizzard gradually passes into a short,
cylindrical tube. The interior of this tube presents internally a
thick, chitinous lining without teeth-like processes. Near its pos-
terior end are found four large, chitinous ridges projecting into
the lumen (fig. 12). These large ridges extend into the anterior
end of the mid-intestine and form part of the oesophageal valve.

Arrangement and action of the muscles

The longitudinal muscles, running in an oblique manner and
forming a net-like cylindrical ring around the oesophagus, group
themselves into six bundles at the beginning of the gizzard (fig. 3).
These six bundles pass respectively under the six large longitu-
dinal folds and attach themselves to the bases of the anterior
teeth. At a point just cephalad of that where these muscles are
attached, there arise six other longitudinal muscles which run
caudad and are inserted on the anterior elongated bases of the
posterior teeth. The fibers of these two sets of muscles form a

FORE-GUT OF CORYDALIS CORNUTUS L. 589

mesh-like arrangement as they pass to their respective points of
insertion and origin on the bases of the anterior teeth.

Behind the anterior bases of the posterior teeth there are no
longitudinal muscles till near the end of the gizzard. The longi-
tudinal muscles found caudad of this point are the terminating
portions of those extending forward from the mid-gut and lie
externally with relation to the circular muscles. When found
underlying the circular muscles they are passing to their points of
insertion.

The circular muscles of the gizzard are much increased in the
posterior half where the powerful teeth are located. Behind the
gizzard they are small and few in number till we reach the oesopha-
geal-valve region. This occurs at the beginning of the four longi-
tudinal ridges which almost completely close the lumen leading
into the mid-gut. Also at this point the longitudinal muscles
from the mid-gut pass in large numbers between the caeca and
are inserted on the chitinous ridges, the ridges alternating with the
caeca.

From the arrangement and position of the muscles of the giz-
zard their action seems apparent. By the contraction of the
circular muscles the lumen can be completely closed, thus exercis-
ing a crushing and straining action on the enclosed food. The
longitudinal muscles, contracting and relaxing at the same time
as the circular muscles, would give to the large teeth a grinding
action on each other effecting a vigorous trituration of the food.
The relaxation of all the muscles combined with the resiliency of
the intima and the onward motion of any food in the oesophagus
would open the gizzard allowing the entrance of more food. Thus
there would be a fairly rhythmic action of the gizzard—contrac-
tion and gradual distension following each other.

The action of the muscles of the oesophageal valve is clear.
The circular muscles by their contraction can readily close the
lumen, preventing the passage of unbroken food particles or, in
connection with the oesophageal valve, regurgitation from the
mid-gut; the longitudinal muscles in turn, aided by the resili-
ency of the intima, open the lumen thus allowing the onward
passage of food from the gizzard.

590 ROBERT MATHESON

The oesophageal valve

In Corydalis the oesophageal valve is not so marked as that
found in many other insects. It is formed by the invagination
of the fore- into the mid-gut. This invagination is short and is
lined internally by a thick intima which is folded into four large,
tooth-like ridges (fig. 12). Externally to the basement mem-
brane lie the circular muscles. Outside of these are found longi-
tudinal muscles, continuations forward of the mid-gut fibers.
They group themselves as they pass between the caeca and most
of them are attached to the four chitinous ridges. Others con-
tinue forward and are attached at various points behind the giz-
zard. ‘These longitudinal muscles, besides their function of
opening the lumen, undoubtedly serve to hold the valve in place
and prevent its evagination by the pushing back of any con-
gested food in the mid-gut. The epithelium lining the valve is
of the ordinary type found in the fore-gut except for a group of
cells located at the point of union between the fore- and mid-gut
epithelium (fig. 10,g.e.). These cells clearly present a glandular
appearance but stand in marked contrast to the glandular epithe-
lium of the mid-gut (fig. 10, ep.). What the particular function
of this group of cells is would be difficult to conjecture as I do
not find any peritrophic membrane and they do not show any of
the characteristics of an imaginal ring as described for so many
insects.

METAMORPHOSIS OF THE FORE-GUT

Pupation

The mature larvae leave the water in the latter part of May or
June and pupate in cavities under stones near the stream. Davis
(03) found that the pupal life lasted from seven to fourteen days,
the average being nine days. The larvae from which my material
was obtained were received about the 23d of May, 1910, and were
placed in a dish with a small amount of water. The dish was then
placed in a box of ‘moist earth over the surface of which were
placed several flat stones. The larvae immediately crawled out

FORE-GUT OF CORYDALIS CORNUTUS L. 591

of the water and began burrowing under the stones. In a short
time many of them had completed their pupal chambers.

The time spent in the pupal chamber varies from one day to
two weeks (Davis ’03). In my specimens the first larva pupated
on June 5, twelve days after the formation of the pupal chamber.
The remainder pupated at various intervals up to the 12th or
13thof June. Specimens were taken at various intervals, from the
time the pupal chambers were formed until the emergence of
the adults. They were killed with chloroform and the alimentary
canal removed under normal salt solution. Various fixing agents
were used, all with good results.

Shortly after the formation of their burrows the larvae become
sluggish. The changes in the intestine which may be observed by
simple dissection are worthy of mention. At first the greater part
of the fore-gut is filled with a dense, blackish fluid, evidently
undigested food material, as this condition is found in the larvae.
while feeding and growing. Very early the fore-gut becomes
emptied and remains empty throughout the entire prepupal and
pupal period except at the posterior end where some of the con-
gested material, formed by the breaking down of the epithelium
of the mid-gut is found. Figure 11 shows the condition of the
fore- and mid-guts at the time of pupation. In this case the intima
of the gizzard was not shed at: the time of the molt, the only
example of this found. It remained lodged at the anterior end of
the oesophagus.

The histological changes undergone by the various systems of
organs during metamorphosis have not been fully investigated
except in a few scatterred species of insects. Even here the
changes observed are so complex that nearly every worker has
offered a different interpretation of the phenomena. Owing to
this wide divergence in the interpretation of the histological
changes it will first be necessary to give a hasty review of the work
already done before taking up in detail my observations on Cory-
dalis. In this review I shall confine myself to the work that has
been done on the fore-gut.

592 ROBERT MATHESON

The epithelium of the fore-gut

Weismann (’64), the pioneer worker on the post-embryonic
development of insects, thought that the entire fore-gut in the
Muscidae was broken down during metamorphosis by a sort of
fatty degeneration, the degenerated epithelium falling into the
lumen of the intestine and forming the yellow body. Weismann
did not solve the method of regeneration of the fore-gut, though
he thought the entire intestine developed from the cellular mass
formed by the broken down larval tissues. This, he believed,
took place by a sort of free-cell formation; the particles isolating
themselves, forming enveloping membranes, losing their fatty
granulations, and each acquiring a nucleus. To these formations
Weismann gave the name of ‘Kornchenkiigeln.’ From such
cells the imaginal intestine developed. Later Kowalevsky (87)
showed that the ‘Kornchenkigeln’ are only phagocytic leucocytes
‘which have been swollen by the engulfment of large quantities
of larval tissues and they take no part in the formation of the
imaginal tissues.

In Corethra according to Weismann (’66), the digestive epithe-
lium passes directly over to the imago without undergoing any
important modifications.

Ganin (’76) states that:

The fore-gut of the larva together with all its appendages are com-
pletely broken down (Anthomyia, Musca, Scatophaga, Eristalis, Stra-
tiomyia, Formica, Myrmica, Lithocolletis, Chrysomela, and Tenebrio).
The food for the formation of the imaginal epithelium is furnished by
the broken down larval cells. The products of the broken down epithe-
lium float freely in a thick fluid between the intima and the basement
membrane. If we observe closely the histological structure of the fore-
gut (at the end of the so-called proventriculus—oesophageal valve) we
find in each mature larva, at the point where the mid-gut epithelium
ends, a narrow strip of peculiar tissue which is sharply differentiated
from the rest. ‘This narrow strip consists of very small, round, trans-
parent cells which possess all the characteristics of young, embryonic
tissues. The later developmental stages demonstrate that from this

small group of cells is built up the entire imaginal epithelium although I
was not able to trace the re-formation to the mouth.

Kowalevsky (’87) is the first worker to give a complete and
detailed account of the metamorphosis of the fore-gut. In the

FORE-GUT OF CORYDALIS CORNUTUS L. 593

blow-fly (Musca vomitoria) he found an imaginal ring situated
at the point of union of the fore- and mid-gut epithelium. This
peculiar ring of cells was present in the youngest larva examined.
During metamorphosis the fore-gut becomes shortened and the
larval epithelium is replaced by the active proliferation of the
cells forming the imaginal ring. These proliferating cells form
the entire imaginal epithelium except the most cephalic part
which he considered as consisting of transformed larval epithe-
lum. The cast-off larval epithelium is destroyed by phagocytes.

Van Rees (’89) in studying the same insect as Kowalevsky,
comes to essentially the same conclusions and verified in a remark-
able way the results of the Russian naturalist.

Rengel (’96) in his studies on Tenebrio molitor dismisses the
fore-gut with the statement that it retains its relative length and
shape through the life of the insect.

Mobuscz (’97) comes to the conclusion that not only the mid-
gut but also the fore- and hind-guts undergo deep-seated changes
during the larval molts and at pupation.

Karawaiew (’98) finds no imaginal ring present in the fore-
gut of Lasius flavus. During pupation the lumen of the caudal
portion of the fore-gut becomes greatly reduced. The epithelium,
which near the middle of the oesophagus consists of a single layer
of cells, becomes several layers thick at the caudal portion. He
was unable to explain this condition. From this region the crop
and gizzard develop through growth and differentiation. In
general he found but slight changes in the epithelium except at
the caudal end where the transformations are not fully explained.

Vesron (’98) as the result of his studies on the metamorphosis
of the silk-worm adopts a new interpretation of the function of
the imaginal ring. He regards it simply asa center of growth
throughout the life of the insect. It originates in the ectoderm
from the germinal band and forms the point of invagination of
the stomodeum. During the larval life this imaginal ring pro-
liferates at each successive molt, thereby increasing the intestinal
epithelium. At the time of pupation it again proliferates in
exactly the same way as at thelarval molts. The cells thus arising
from the imaginal ring unite with those previously formed, with-

594 ROBERT MATHESON

out destroying or replacing them in any way whatever. Thus
he considers the epithelium to pass over to the imago, though some-
what modified to adapt it to the changed conditions. As the
epithelium is sharply differentiated from the hypodermis he ex-
cludes any possibility of regeneration from imaginal buceal dises.

Karawaiew (99) in his studies on Anobium paniceum (Coleop-
tera) finds but slight changes in the fore-gut during metamor-
phosis. The imaginal epithelium develops directly from the
larval without any apparent changes. In this respect this beetle
exhibits much similarity to the ants.

Anglas (00) finds in the bee and the wasp a well-marked zone
of growth between the fore- and mid-guts. He does not regard
this as an imaginal ring but more like the condition found by
Karawaiew in Lasius flavus. By the proliferation of the cells of
this growth area the greater part of the imaginal epithelium is
formed. The proliferating cells invade and engulf the larval
cells. The anterior part of the larval oesophagus becomes that
of the adult. He finds it difficult to distinguish the point of
union of the part formed by the cells from the growth zone and
that formed by the transformed larval epithelium. One passes
insensibly into the other.

Deegener (’00) finds an imaginal ring present in the fore-gut
of Hydrophilus. During the prepupal period this imaginal ring
rapidly increases by mitotic divisions of the cells, while the larval
epithelium is discharged into the lumen where is found a consider-
able number of ‘Kornchenkiigeln.’ During the pupal period the
imaginal epithelium is formed into a cylindrical tube within which
are found remnants of the larval cells and Kornchenkiigeln. Later
Deegener (’04) finds he was in error regarding the presence of
Kornchenkiigeln within the lumen and he states that phagocytes
do not take any part in the breaking down of the larval epithe-
lium.

Perez (’02) describes in Formica an imaginal ring for the fore-
gut, situated as in the Muscidae. During the prepupal period
the cells of this ring actively divide and the epithelium of the
oesophageal valve is discharged into the lumen. By the contin-
ued proliferation of the imaginal ring the fore-gut is elongated and

FORE-GUT OF CORYDALIS CORNUTUS L. 595

the larval epithelium is forced cephalad to become the imaginal
‘oesophagus.

Vaney (’02), in his studies on dipterous larvae, finds imaginal
rings present in the following;—Chironomus, Anthomyia, Stra-
tiomyia, Tanypus and Gastrophilus. In Stratiomyia he also found
in front of the oesophagus a small buccal disc, while in Gastro-
philus there is one just caudad of the pharynx. Only in Gastro-
philus did he study the metamorphosis of the fore-gut. There
was degeneration of the larval epithelium in situ and its destruc-
tion by phagocytes. The imaginal epithelium arises exclusively
from the imaginal ring and the post-buccal disc.

Deegener (’04), in his studies on the metamorphosis of Cybister
roeselii, records a prominent imaginal ring situated at the point of
union of the fore- and mid-gut epithelium. He divides the meta-
morphosis into two distinct periods;—(1) the shedding of the
larval intima and the formation of the pupal epithelium; (2) the
molting of the pupal intima and the rebuilding of the imaginal
epithelium. The first period is completed about twenty-four
hours after pupation. The epithelium of the anterior portion of
the oesophagus becomes that of the pupa, with but slight changes
in the nuclear chromatin and cellular cytoplasm. The epithe-
lium of the posterior portion breaks down and is forced into the
lumen by the rapid forward growth of the imaginal ring. The
degenerated larval cells form a deeply staining mass lying between
the molted larval intima and the developing pupal epithelium.
The imaginal ring cells divide mitotically with extreme rapidity.
The spindles are always located in the plasma layer next the
lumen, their long axes parallel to that of the intestine. Numerous
degenerating nuclei are observed directly in front of the imaginal
ring. Deegener thinks that the larval cells fail inthe struggle with
the imaginal, become broken down and absorbed by the more
actively growing embryonic cells.

The pupal epithelium is completed about the end of the third
day after pupation. Then immediately begins the formation of
the definitive fore-gut. The posterior part, through growth and
a great increase by the division of the imaginal ring cells, becomes
differentiated into the crop and gizzard of the adult. The ante-

596 ROBERT MATHESON

rior part, with but slight changes in its nuclear and cytoplasmic
contents, becomes that of the imago. Deegener states that the
intima, both pupal and imaginal. is formed by a direct transfor-
mation of the inner plasma layer of the epithelium.

Verson (’05) presents the detailed results of his work on Bom-
byx mori. In respect to the fore-gut he affirms the conclusions
given in his previous paper (1898).

Thompson (’05) finds that in Culex the oesophageal epithelium
undergoes no metamorphosis. It is handed over intact to the
imago with only an increase in the number of its component cells.

Van Leeuwen (’07) in his studies on Isosoma graminicola finds
no imaginal ring present, though he states that the posterior part
consists of cells, differing from the rest and containing many
small nuclei. It is here that the most important changes during
metamorphosis occur. There is a marked lengthening of the
caudal part of the fore-gut during the pupal period by mitotic
cell division. The anterior division degenerates in situ and is
replaced by a hypodermal invagination. In the caudal region are
developed the sucking stomach and gizzard.

Lubben (’07) finds but slight changes in the fore-gut of Trichop-
tera during metamorphosis. There is no imaginal ring present
(in the sense of Kowalevsky) though the entire oesophageal valve
might be considered one, as he states Deegener (’04) does for
the beetle intestine. In this portion there is considerable cell
increase. The main changes are the flattening of the pupal
epithelium, a considerable increase in length and the formation
of the characteristic imaginal gizzard. He doubts whether the
intestine functions in the adult.

Russ (’08) in his studies on Anabolia laevis, one of the Trichop-
tera, finds a well marked imaginal ring present, which consists of
tall, cylindrical cells heaped upon one another in several layers.
The anterior part of the oesophagus, except for slight changes,
becomes that of the imago. Except in the very mouth region he
finds in this part a partial renovation through the divisions of the
larval cells. In the caudal portion the larval epithelium degen-
erates, being absorbed and replaced by the actively growing
embryonic cells. The pupal epithelium of the entire fore-gut

FORE-GUT OF CORYDALIS CORNUTUS L. 597

consists of tall cylindrical cells. These become much flattened
in the adult and there is consequently a great enlargement of
the lumen.

Deegener (’08) investigated the metamorphosis of the intes-
tine of Malacosoma castrensis (Lepidoptera), which has a well
marked imaginal ring for the fore-gut. Asin Cybister he divides
the time of transformation into two distinct periods;—the shed-
ding of the larval intima and the formation of the pupal epithelium;
the shedding of the pupal intima and the formation of the imaginal
epithelium. The first period is completed at the end of the first
pupal day. The, pupal epithelium is identical with that of the
larva except for a slight transformation of the cellular contents.
He does not find any activity on the part of the imaginal ring,
although in its neighborhood there may be seen a few degenerating
nuclei during the pupal period. About the end of the third day
after pupation the pupal intima is molted and the formation of
the imaginal epithelium begins. This is completed between the
fifth and sixth days of the pupal life. The pupal epithelial cells,
except for a slight reduction in size and rearrangement of their
chromatin material, pass over directly into the imaginal epithe-
lium.

It might be further noted here that Deegener considers the
basement membrane as a temporary structure, appearing only
while a well developed intima is lacking or to protect the epithe-
lium from the severe contractions of the surrounding muscles.

Perez (’10) in his extensive studies on the metamorphosis of
the Muscidae confirms and extends the results of Kowalevsky
(’87) and Van Rees (’89) in regard to the post-embryonic develop-
ment of the fore-gut.

THE EPITHELIUM IN CORYDALIS DURING METAMORPHOSIS
The prepupal period

The characteristics of the larval epithelium have been fully
discussed and illustrated in the first part of this paper. We shall
now examine the changes undergone during the prepupal period
up to the time of the last larval molt. The first changes appear

JOURNAL OF MORPHOLOGY, VOL. 23,,NO. 4

598 ROBERT MATHESON

in the oesophageal valve region where the cells become elongated
and narrowed and the cell walls more distinct. The nucleiare
more crowded and more chromatic. The cytoplasm appears
more granular and shows a sharper differential staining reaction.
The intima shows little change except for a slight granular appear-
ance near the epithelial cells. It has not yet become separated
from the epithelium. These general observations hold for the
entire fore-gut at an early prepupal stage. But in the anterior
portion the cells do not elongate so much nor do the cell walls
become so sharply differentiated. There is no imaginal ring
present and the peculiar glandular cells found between tbe fore-
and mid-guts show no activity throughout the entire prepupal and
pupal periods.

The absence of an imaginal ring is in marked contrast to the
condition described by several authors for a number of other
insects. In all the dipterous forms so far carefully studied an
imaginal ring is recorded (Ganin ’76, Kowalevsky ’87, Van Rees
89, Vaney ’02 and Perez ’10). It is recorded as present in the
Hymenoptera by Ganin (’76) for Formica and Myrmica, and
Perez (’02) for Formica. Karawaiew (’98) notes its absence in
Lasius flavus, Anglas (’00) in the bee and wasp, and Van Leeuwen
(07) in Isosoma graminicola. In the Coleoptera Ganin (’76)
finds it in Chrysomela and Tenebrio, Deegener (’00) and (’04)
in Hydrophilus and Cybister. Karawaiew (’99) notes its absence
in Anobium paniceum. Ganin (’76), Verson (’98) and Deegener
(08) record its presence in Lepidoptera (Lithocolletis, Bombyx
mori and Malacosoma ecastrensis). In the Trichoptera Lubben
(67) regards the imaginal ring as absent while Russ (’08) records
a prominent one for Anabolia laevis.

In a slightly older prepupa than that just described the intima
becomes completely separated from the epithelium. The inner
surfaces of the cells show in many places small, protoplasmic
projections as if they had been drawn out of the intima. These
are not shed into the lumen but later become flattened down.
There are no signs of degeneration. The nuclei show their char-
acteristic staining reactions while the cytoplasm remains granular
(fig. 19). The intima, however, shows considerable change.

FORE-GUT OF CORYDALIS CORNUTUS L. 599

In the larva not preparing to pupate (fig. 17) the narrow, inner,
chitinized layer stains densely black with iron-haematoxylin
while the remainder usually shows a very characteristic laminate
appearance. In the prepupa, at the time of molting of the larval
intima, the black area is much increased and the inner layer loses
somewhat of its laminate structure. It hasabroken down appear-
ance, often reticulate, the loose intertwining strands filling the
space between the epithelium and intima (fig. 19). The base-
ment membrane is sharply differentiated throughout the entire
prepupal period.

Changes at tume of pupation

The external changes in size, shape, et cetera, may be seen in
figure 11. In this pupa the intima of the gizzard was not dis-
charged at the molt but remained lodged at the anterior end of
the oesophagus. The fore-gut is now a long, cylindrical tube prac-
tically empty, while the mid-gut is greatly expanded by a dense,
yellowish fluid. None of this fluid extends into the fore-gut
except at its posterior end where the nearly evaginated oesopha-
geal valve fails to form a closing bridge. A short distance cepha-
lad, however, the fore-gut is completely closed.

As Corydalis larvae, under similar conditions, vary so much
individually in the time spent in the pupal chamber before pupa-
tion, as wellas the pupae in the time occupied before the emergence
of the adult, I have found it impossible to follow the changes
according to the time element. The changes to be described here
as occurring at the time of pupation cannot be said to always
occur at this period, since my experience shows that the indivi-
dual larvae pupate when different histological changes are in
progress. I shall, therefore, present in as connected and as logi-
cal a way as possible the changes that occur throughout pupation
up to the time of the emergence of the adult, without special
reference to the particular time at which these changes are to be
observed.

With the shedding of the larval intima the oesophageal valve
becomes partially evaginated and appears in cross-section as a

600 ROBERT MATHESON

large, cylindrical tube with a somewhat flattened epithelium in
its anterior portion and distended with fluid from the mid-guv.
In longitudinal section the valve is never seen to form an open
cylindrical tube, but the walls, which in the larva project into the -
mid-gut, now form a wide ring projecting at right angles to the
long axis of the alimentary canal and almost closing the lumen.
The epithelium of the posterior’ portion consists of rather tall,
narrow, cylindrical cells with well defined walls. The nuclei
are large and chromatic, are situated near the bases of the cells,
and do not show any indication of degeneration. The cytoplasm
is vacuolate and granular, the vacuoles being located near the
tips of the cells (fig. 21, v). At this stage no epithelial cells were
found undergoing division. A well defined basement membrane
is present and a pupal intima is formed though it stains but slightly
in eosin and not at all with iron-haematoxylin.

The tall, glandular-appearing cells, situated in the larva at the
juncture of the fore- and mid-gut epithelium, have become
greatly reduced in size and show no signs of secretory activity.
They have become narrow, cylindrical cells like the rest of the
fore-gut epithelium.

Throughout the entire fore-gut the epithelium is thrown into
longitudinal folds, but the folds are much reduced as compared
with those of the larval oesophagus. The lumen is closed a short
distance in front of the oesophageal valve, but appears again
slightly cephalad and continues as a small tube almost closed by
the longitudinal folds. The circular muscles are in a state of
active contraction.

Changes from the time of pupation to the shedding of the pupal intima

The next stage which requires special attention is that from
the time of pupation to the shedding of the pupalintima. The
time this occupies is dependent largely upon the individuals
examined, but requires two to three days. Some time after pupa-
tion in the region formerly occupied by the larval gizzard, the
epithelium, which consists of greatly crowded cells, shows many
nuclei apparently undergoing chromotolysis (figs. 13 and 14, 2).

FORE-GUT OF CORYDALIS CORNUTUS L. 601

The chromolytic drops are located near the basement membrane,
and it is difficult at times to decide whether one is observing a
-degenerating nucleus or one that is massing its chromatin pre-
paratory to division. But, judging from the position of these
deeply staining masses and the work of previous authors, it would
seem conclusive that these are actually degenerating nuclei.
None of the degenerating nuclei or cellular contents are dis-
charged into the lumen but are undoubtedly absorbed by the
neighboring cells. In no case have I observed any part of a
cell or its contents discharged into the lumen or forced out through
the basement membrane. Cephalad or caudad of this region
very few degenerating nuclei are found.

The epithelial cells show decided secretory activity. The
nuclei are prominent, chromatic and located near the bases of the
cells (fig. 21). The cytoplasm is granular and almost filled with
large vacuoles (figs. 13 and 21). At this period also many nuclei
in the gizzard region are undergoing mitotic division. The cell
preparing to divide becomes large, migrates to the inner surface,
retaining only a very narrow connection with the basement mem-
brane, and at the same time possessing a well marked cell wall
(figs. 27 and 28). Very few mitotic figures are found either in the
anterior oesophageal region or near the oesophageal valve. The
epithelium is still thrown into many longitudinal folds which at
this stage practically close the lumen. There is also a consider-
able reduction in the diameter of the canal. Thechanges described
above are usually found in pupae from two to three days old.

At a slightly later period further changes may be observed.
There is a constant decrease in the size of the intestine and a
marked reduction of the longitudinal folds. The epithelial cells
are gradually being reduced in size, due to the active secretion of
a very large quantity of pupal intima. The vacuolate condition
is also being reduced. Many nuclei throughout nearly the entire
fore-gut are undergoing division and many are degenerating
The cellular contents possess the same characteristic appearance
and give the same staining reactions. Here and there the pupal
intima is separated from the underlying epithelium (fig. 29)

602 ROBERT MATHESON

The shedding of the pupal intima

This occurs three to four days after pupation, and the changes
that have been going on up to this time become more intensive.
The epithelial cells retain their characteristic appearance. The
number of degenerating cells is rapidly decreasing, while there is
a marked increase in the number undergoing division. Inany
cross section one can count from four to seven nuclei undergoing
mitosis. The dividing nuclei are not so numerous in the anterior
or posterior portions. The pupal intima now lies free within the
lumen. Between the cast intima and the epithelium one finds a
somewhat granular secretion filling up this area and so closing the
lumen (fig. 26, g.m.). Owing to their great secretory activity
the cells are being rapidly reduced in size; the vacuoles, now nearly
all at the tips of the cells, are smaller or frequently lacking; the
longitudinal folds are lacking except near the posterior end where
numerous small folds are being formed; the intestine is somewhat
larger. The oesophageal valve is now reformed as it will appear
in the imago except for slight cellular changes.

The beginnings of the imaginal intima now show their early
appearance. In figure 22, one observes near the inner surface of
the cells a row of very distinct, sharply staining, black dots,
underlying a granular layer which does not show very clearly in
the photograph. This is shown much better in figure 29, r, from
a pupa only forty-eight hours old, though it is not commonly seen
so early. This condition agrees with what Deegener (’08) con-
siders as the origin of the intima in the developing fore-gut of
Malacosoma castrensis. As the imaginal intima appears in this
place one seems forced to conclude that it is formed not only by
secretion but also by a direct transformation of the cells them-
selves.

Formation of the imaginal epithelium

The changes that occur after the shedding of the pupal intima
are not very considerable. The oesophageal valve remains practi-
cally the same as in the four-day-old pupa. In the region directly
cephalad of the valve the formation of longitudinal ridges con-
*, tinues to a marked degree. This dccupies only a short distance,

FORE-GUT OF CORYDALIS CORNUTUS L. 603

as the major part of the fore-gut consists of a smooth cylindrical
tube. Throughout the entire epithelium there is a gradual reduc-
tion in the height of the cells and an increase in the diameter of
the intestine. From the time of the shedding of the pupal intima
till about the ninth day there occurs a constant increase of cells
by mitotic division, and degenerating nuclei are found in gradually
reducing numbers. With the definitive formation of the imaginal
intima one does not find any more dividing or degenerating nuclei.

Finally, in a pupa ten days old the epithelium appears in practi-
cally the same form as in the adult. It is arranged in numerous,
small, longitudinal folds throughout the greater part of the
oesophagus. In the posterior portion this folding is very marked.
It consists of flattened cells with granular cytoplasm and large
nuclei. The vacuolate condition, which was so marked in the
early stages, has now disappeared. The great reduction in the
size of the cells is due both to the secretion of a large quantity of
granular substance which lies between the shed pupal and the
forming imaginal intima, and to the increase in the diameter of
the canal. The cellular walls are very indistinct or lacking.
The imaginal intima appears as a narrow, light area and is now
well formed. A basement membrane is present throughout the
entire pupal period. At this stage no degenerating nuclei or
any undergoing division are found.

Shortly after the loosening of the pupal intima there occurs a
slight outpocketing of the walls of the oesophagus a short distance
in front of the oesophageal.valve. This has been described and
figured by Leidy (48) who designated it the pupal crop and notes
its resemblance to the sucking stomach of Lepidoptera. During
its formation the epithelial cells in this region undergo rapid and
extensive mitotic divisions. The epithelium lining this outpocket-
ing is identical with that of the oesophagus. What its function
may be is doubtful. It passes over to the imago as a small
appendage of the posterior portion of the oesophagus.

Cell division

The phenomenon of cell division in the epithelium of the fore-
gut is very characteristic. The cell, preparatory to division,

604 ROBERT MATHESON

enlarges and the greater portion migrates towards the distal sur-
face (figs. 27 and 28, m.). Each cell, however, retains a narrow
protoplasmic strand connecting it with the basement membrane
(fig. 28, st). A cell wall is always present and well defined. The
nuclear spindle is nearly always parallel to the long axis of the
intestine and is placed at the inner side of a large vacuole. The
presence of a vacuole in every dividing cell, together with its
staining reaction, clearly prove that these cells belong to the ordi-
nary larval.epithelium and are not tracheal cells which have mi-
grated through the basement membrane, as Anglas (’04) thinks
is the origin of the replacement cells in the Hymenoptera. The
only method of cellular increase in the fore-gut epithelium is by
indirect divisions. Shortly after pupation one finds scattered
cells undergoing division in the larval gizzard region. The num-
ber gradually increases, reaching its maximum shortly after the
shedding of the pupal intima. ‘Then follows a gradual reduction,
till division ceases just before the definitive formation of the imag-
inalintima. The cellular increase is greatest in the larval gizzard
region.

HISTOLYSIS AND HISTOGENESIS OF THE INTESTINAL MUSCLES
Histolysis

Before attempting a discussion of the histolysis of the muscles
in Corydalis it may be well to present as briefly as possible the
current interpretations of so complex aphenomenon. At present
there are several theories, each apparently well founded upon
observed facts:

1. That of the leucocytic phagocytosis of Kowalevsky, Van
Rees, and Mercier. According to these workers the muscles are
attacked, broken down and engulfed by leucocytes before any
chemical changes visible under the microscope have taken place
within the muscle substance. Digestion of the débris occurs
within the leucocytes.

2. That of the modified leucocytic phagocytosis of Loos, Batail-
lon, and Mercier for Amphibia; Deegener, Vaney, Verson, and
others for insects. According to this theory there is, first, a

a

FORE-GUT OF CORYDALIS CORNUTUS L. 605

chemical and physical change in the muscle substance before leuco-
cytic intervention. The activity of the leucocytes may be great
(Mercier, Vaney, etc.) or very unimportant (Deegener, Karawaiew,
etc.). The preceding chemical changes may be slight or very
marked. t

3. The lyocytic hypothesis of Anglas.(’00). According to
this view there is no engulfing of sarcolytes but an extracellular
digestion by means of diastases excreted by the leucocytes. The
tissues thus dissolved are absorbed by the leucocytes. This
process Anglas designates ‘lyocytosis.’

4. The auto-phagocytic theory of Metschnikoff for Amphibia,
De Bruyne and Russ for insects. According to this theory some
of the muscle nuclei surround themselves by protoplasm and
constitute true phagocytes (sarcoclastes or myoclastes of De
Bruyne). De Bruyne found this condition of thedegenerating
muscles of Musca vomitoria while Metschnikoff regards it as the
normal process in the absorption of the tailin Anura. Russ (’08)
adopts this interpretation for the degeneration of the muscles in
Anabolia laevis, one of the Trichoptera.

5. The purely chemical and physical hypothesis of Korotneff.
In a Tineid larva he finds the muscles are entirely broken down by
a chemical liquefying process without any intervention of phago-
cytes. Karawaiew adopts this view for the ants (Lasius niger),
Terre (’99) for the bee and Deegener (’04) for Cybister. Korot-
neff admits a true phagocytic activity where the metamorphosis
is extremely rapid, as in so many Diptera. He advocates the
hypothesis that where metamorphosis continues for a consider-
able time there is no phagocytic intervention, but on the other
hand phagocytes play the all-important role where the changes are
rapid.

6. The theory of Berlese. According to this worker the first
change to appear is the separation of the sarcolemma and the
nucleus from the fibrillar part. This he calls myolysis. The
fibers are chemically dissolved whereas the nucleus with its plasma
survives—it is the living part of the cell capable of regenerating
new contractile fibers. The dissolution of the fibers (fibriolysis)
is probably caused by the extravasation of the intestinal contents

606 ROBERT MATHESON

at the end of the larval life. The partially broken down fibers
(sarcolytes) are taken up by means of the protoplasmic expansions
of the amoebocytes (leucocytes of other authors) and carried by
them to all parts of the body where regeneration is in progress,
furnishing the necessary nutrition. These are the so-called ‘ Kor-
chenkiigeln,’ by Berlese designated sarcolytocytes. The sar-
colytocytes are found only in the more highly specialized insects.
Within the sarcolytocytes no digestion takes place, they simply
act as carriers of the débris. The muscle nuclei, with their cyto-
plasm, become independent cells. ‘The nuclei fragment, each
forming a mass within its own enveloping membrane (sarcocyte).
These sarcocytes play an important réle in the histogenesis of the
imaginal muscles.

The metamorphosis of the intestinal muscles of insects has been
investigated by several workers. In nearly all the dipterous
forms they are destroyed by phagocytes without a previous chem-
ical change visible under the microscope (Kowalevsky, Van Rees
and Perez). Vaney (’02) thinks there is a chemical change within
the muscle before the leucocytes become active (in Gastrophilus).
Thompson (’05) finds no phagocytosis in Culex. In all the Coleop-
tera so far studied there has not been found any phagocytic activ-
ity, the muscles apparently liquefying in place (Karawaiew ’99,
Breed ’03 and Deegener 704). In the Hymenoptera Karawaiew
(98) and Van Leeuwen (’07) hold to the chemical and physical
hypothesis, while Perez (’02) finds phagocytosis in Formica. In
the Lepidoptera Korotneff denies any phagocytosis while Verson
(05) and Deegener (’08) find phagocytic activity following a chem-
ical dissolution. In the Trichoptera-Russ (’08) thinks there is
present a form of auto-phagocytosis though he is doubtful of his
interpretation of the process in the species (Anabolia laevis)
studied.

In Corydalis the fore-gut possesses a well developed muscular
system. The fibers are all cross striated, the striations standing
out very clearly. During the prepupal period there are no visible
changes until about the time of the loosening of the larval intima.
At this time the muscles become strongly contracted, as is shown
by the frequent crenulate condition of the sarcolemma and the

FORE-GUT OF CORYDALIS CORNUTUS L. 607

obscuration of the finer discs, Krause’s membrane and Hensen’s
disc. The nuclei are more chromatic, large and more prominent.
This strikes one forcibly when comparing sections of the muscles
of the larva and the prepupa (figs. 17 and 19). The sarcoplasm is
also more prominent, though undoubtedly this is due to the great
contractions of the fibers. I do not find any marked differences in
the staining reactions of the fibers. Scattered among the muscles
are found a considerable number of leucocytes, which, however,
show no signs of special activity. They are more numerous in the
posterior region, particularly near the oesophageal valve. During
the larval life it is rare to find any leucolytes in the neighborhood
of the fore-gut.

At the time of pupation the changes above outlined become more
marked. The cross striations are beginning to disappear, while
the outer portions of the circular muscles show in many places a
granulose structure. The cross striations are affected earliest
in the neighborhood of the nuclei and show first near the anterior
end of the oesophagus. As to why this degeneration should first
appear around the nuclei seems difficult to explain except on the
basis of the location of the greatest cytoplasmic activity near the
nuclei. This does not indicate auto-phagocytosis in any sense
of that word. It simply indicates that the chemical activity is
greatest where the cytoplasm is greatest and the liquefying
process proceeds outward from the nuclei as centers. This, I
think, is borne out by the facts shown in the degeneration in
the muscles of Corydalis.

The sarcoplasm is constantly increasing in amount while the
nuclei are large and prominent (fig. 20). In cross-sections the
areas of Cohnheim do not stand out so sharply. As yet there is
no noticeable change in the fibrillar structure except near the
“nuclei. Leucocytes are distributed somewhat sparingly around
and among the muscles, but they do not show any activity in the
destruction of them.

At a slightly later period the cross striations have almost
completely disappeared except near the posterior end of the fore-
gut, while here and there scattered strands show a faint cross
striate condition. The greater part of the circular muscles have

608 ROBERT MATHESON

not only lost their cross striations but their fibrillar part as well
(fig. 21). The densely staining portion in figure 21 is the rem-
nants of the fibrillae undergoing degeneration. The muscular
layer now appears as a granular, cellular mass, separated by the
undissolved sarcolemma and remnants of the supporting tissue.
The nuclei are not so deeply staining nor so prominent as at the
time of pupation. Surrounding the fore-gut are found numerous
leucocytes which are undoubtedly taking up some of the broken
down tissues, as many are seen to contain deeply staining granules.
The above described changes are completed about twenty-four
to forty-eight hours after pupation.

The process of liquefaction of the fibrillae continues till about
the time of the shedding of the pupal intima. At this time the
muscular layers have the appearance shown in figure 22. Here
and there are found nuclei undergoing chromotolysis while some
are dividing mitotically. The fibrillar structure is practically
lacking, while the nuclei and surrounding cytoplasm appear some-
what rejuvenated. Many leucocytes are found surrounding the
muscular layers and are absorbing portions of the chromolytic
drops as may be seen from their contents (fig. 22, 1.). We have
now reached the end of the histolytic process. The muscular
tissue is not destroyed by phagocytic leucocytes but seems to
liquefy in place, the process beginning about the nuclei as centers.
Undoubtedly, as previously pointed out, some of the débris is
engulfed by the leucocytes. This maybe seen from their contents,
while several have been found actively surrounding small parti-
cles. Also in the larval gizzard region the muscle layer is reduced,
and the liquefied muscular tissue is taken up by leucocytes which
are most abundant in this region. However, the leucocytes do
not seem to play an active part in the destruction of the muscles.

Histogenesis

Nearly all the methods of muscle regeneration so far described
for the fore-gut may be assigned to the following types or modifi-
cations of these types:

1. The regeneration of the destroyed muscles from mesodermic
cells, myocytes, which are found in the coelom near the inner sur-

FORE-GUT OF CORYDALIS CORNUTUS L. 609

face of the imaginal discs. This type of histogenesis is described
for the fore-gut muscles of Musea vomitoria (Kowalevsky and
Perez), Gastrophilus (Vaney), Malacosoma castrensis (Deegener
08), Isosoma graminicola (Van Leeuwen ’07), and some others.

2. The regeneration of the imaginal muscles from the larval’
muscle nuclei and cytoplasm. The fibrillar part is destroyed
while the nuclei persist, divide and form myoblasts. These form
new fibrillae. This view, or a modification of it, is held by Korot-
neff (92), Anglas (’00) for some muscles, Breed (’03), Deegener
(04) and Berlese.

3. The regeneration of the imaginal muscles from small ima-
ginal nuclei present in the muscles. These nuclei become active
at the time of pupation, while most or all of the larval nuclei
degenerate. This type of regeneration is a complicated one and
the investigators who advance this theory always admit that at a
certain period it is impossible to tell whether one is dealing with
rejuvenated larval nuclei or the so-called imaginal nuclei. This
view has been put forward by Karawaiew (’98) for Lasius niger
and Russ (’08) for Anabolia laevis (Trichoptera).

In discussing the histolysis of the fore-gut muscles of Corydalis
I stated that this process ended about four days after the time of
pupation. As the changes in the epithelium are not governed by
the time element so much as by the individual pupal conditions
so also is the muscle histolysis and histogenesis.

By examining figure 22 one notes that the larval-nuclei clearly
show a rejuvenated condition. They are arranging themselves
in rows, while in the cross sections of the longitudinal muscles
one can see faint indications of forming fibrillae. Leucocytes are
quite numerous and are filled with densely staining granules. In
a slightly more advanced condition (fig. 23) the longitudinal
muscles clearly show the cut ends of forming fibrillae. The
circular muscles do not show well in the figure. It is probable
that figure 23 shows about the same stage of muscle regeneration
as figure 22 although figure 23 is from a pupa nearly eight days
old. However, further cephalad in the latter the fibrillee show
clearly in the circular muscles. During this time one finds nuclei

610 ROBERT MATHESON

undergoing division scattered throughout the muscles. However,
the number that divide is very small.

Figure 24 shows a more advanced stage in muscle regeneration.
The longitudinal muscles show not only the cut ends of fibrillae
but have assumed their definitive arrangement, side -by side
immediately beneath the epithelium. ‘The fibrillae show dis-
tinetly in the circular muscles. Figure 25 shows a somewhat
older condition while figure 26 shows the beginnings of the cross
striations in the muscle fibers. Ina pupa ten days old the muscles
have assumed their definitive form, though the cross striations
do not show as clearly as they will become in the adult.

The type of histogenesis of the intestinal muscles in Corydalis
is that described under (2). We have first a chemical liquefying
of the muscles in place, beginning around the nuclei as centers.
The leucocytes, which surround and penetrate between the muscle
layers, undoubtedly take up and digest some of the broken down
tissues. After the chemical liquefaction the larval nuclei, though
some of them degenerate, become rejuvenated, and from them and
the surrounding cytoplasm are built up the imaginal muscles.

CONCLUSIONS

1. The fore-gut of the larva of Corydalis may be divided into
five well marked regions: pharynx, oesophagus, gizzard, portion
between gizzard and oesophageal valve, and oesophageal valve.

2. The pharynx is provided with a series of dilator muscles
attached to the walls of the head. The oesophagus presents no
features of peculiar interest except perhaps the large number of
longitudinal folds which permit a very great distension of its
lumen.

3. The gizzard is well developed. Although Plateau (’74) and
a number of workers following him have assigned to the gizzard
only a straining function, it is evident that in Corydalis it exer-
cises a grinding and crushing action as well. This is evidenced
by the development and arrangement of muscles, the powerful
. chitinous teeth and their method of interlocking. Bordas (’05)

=

FORE-GUT OF CORYDALIS CORNUTUS L. 611

observed the rhythmic contraction and expansion of the gizzard
in living Vespidae (Vespa crabo).

4. The gizzard does not pass directly into the oesophageal
valve but is joined to it by a short, more or less smooth, tube.

5. The oesophageal valve is short and is lined with four strongly
chitinized ridges which alternate with the caeca. At the point
of union of the epithelium of mid- and fore-guts there is located
a peculiar group of glandular cells whose function is doubtful.

6. The metamorphosis of the fore-gut is of a generalized type.
The larval epithelium becomes partially broken down and the
cells destroyed are replaced by the division of rejuvenated larval
cells. The nuclei always divide mitotically and every spindle is
located at the side of a vacuole. The dividing cell migrates
towards the inner surface, though it retains a connection with the
basement membrane.

7. The histolysis and histogenesis of the muscular coats are
also generalized processes. The muscles liquefy in place. The
greater number of the larval nuclei becomerejuvenated and around
them as centers the new fibrillar structures are developed.

8. The role of the leucocytes is a comparatively unimportant
one. The leucocytes are present throughout the pupal life and
are seen to engulf small particles of the broken down tissues.
They do not take any active part in the destruction of the larval
muscles or epithelium.

612 ROBERT MATHESON
BIBLIOGRAPHY

Anauas, J. 1898 Sur Vhistolyse et ’histogenese du tube digestif des Hyménop-
téres pendant la métamorphose. C. R. Soc. Biol., pp. 1167-1170.

1899 Sur Vhistolyse et ’histogenese des muscles des Hyménoptéres
pendant le métamorphose. C. R. Soe. Biol., pp. 931-933.

1900 a Note preliminaire sur la métamorphose internes de la guepe
et de l’abeille. Le lyocytose. C.R. Soc. Biol., T. 52, pp. 94-97.

1900 b Sur la significance des termes ‘phagocytose’ et ‘lyocytose.’
C. R. Soc. Biol., T. 52, pp. 219-221.

1901 a Observations sur les métamorphoses internes de la guepe et de
Vabeille. Bull. Sci. France et Belgique, T. 34, pp. 363-473, plates 19-23.

1901 b Quelques caractéres essentiels de l’histolyse pendant la méta-
morphose. Bull. Soc. Ent. France, pp. 301-304.

1902 a Le phénoménes des métamorphose internes. Scientia, Series
Biol., T. 17, Paris, 84 pages.
a

1902 b Nouvelles observations sur les métamorphoses internes. Arch.
d’Anat. Micr., T. 5, pp. 78-121.

1904 a De l’origin des cellules de remplacement de l’intestine chez les
Hyménoptéres. C. R. Soc. Biol., T. 56, pp. 173-174.

1904 b Du réle des trachées dans la métamorphose desinsectes. C. R.
Soc. Biol., T. 56, pp. 175-176.

BATAILLON, E. 1900a Le probléme des métamorphoses. C. R. Soc. Biol., T. 52,
pp. 244-247.

1900 b Le théorie des métamorphoses de M. Ch. Perez. Bull. Soc.
Ent. France, pp. 58-62.

BrerRiesE, A. 1898 Osservazioni sopra particolari fenomeni che avvengono nella
ninfosi des Muscidi. Riv. Patol.’ veg., T. 6, pp. 269-272.

1899 Osservazioni su fenomeni che avvengono la ninfosi degli insetti
metabolici. Parte 1. Tissuto adiposi (Trofifitti). Riv. Patol. veg.,
I. 7, pp. 1-155.

1900 Considerazioni sulla fagotitosis negli insetti metabolici. Zool.
Anz., Bd. 23, pp. 441-449.

*
1901 Vorgiinge welche wihrend der Metamorphose der metabolischen
Insecten vorkommen. Zool. Anz., Bd. 24, pp. 515-521.

1902 a Osservazioni su fenomeni che avvengono durante la ninfosi
degli insetti metabolici. Parte 1°. Tissuto adiposa. Memoria sec-
onda. (Lepidotteri, Imenotteri, Neurotteri, Coleotteri). Riv. Patol.
veg., T. 9., pp 177-344.

1902 b Parte II. Tissuto muscolare (Miociti). Riv. Patol. veg., T.
10, pp. 1-120.

FORE-GUT OF CORYDALIS CORNUTUS L. 613

Boas, J. E. W. 1899 Einige Bemerkungen iiber die Metamorphose der Insecten.
Zool. Jahrb., Syst., Bd. 12, pp. 385-402.

Breep, R.S. 1903 The changes which occur in the muscles of a beetle, Thymalus
marginicollis Chevr., during metamorphosis. Bull. Mus. Comp. Zool.,
Harvard College, vol. 40, pp. 317-882.

BruneEvuI, G. 1902 Sul significato della metamorfosi negli insetti. Riv. Ital.
Sci. Nat., T. 22, pp. 100-106.

Bruyne, C. dE. 1898 Recherches au subjet Vintervention de la phagocytose
dans le développement des invertebrés. Arch. de Biologie, T. 15, pp.
181-800.

Curnot, L. 1899 Les prétendus organes phagocytaires décrits par Koulvetch
chez la Blatta. Arch. Zool. Exp., Notes, pp. 1-2.

Derecener, P. 1900 Entwicklung der Mundwerkzeuge und des Darmkanals
von Hydrophilus. Zeitschr. f. wiss. Zool., Bd. 68, pp. 113-168.

1903 Zur postembryonalen Entwicklung des Insectendarms. Zool.
Anz., 26, pp. 547-550.

1904 Die Entwicklung des Darmkanals der Insecten wihrend der
Metamorphose. Zool. Jahrb. Anat., 20, pp. 499-676.

1908 Die Entwicklung des Darmkanals der Insecten wahrend Meta-
morphose. II Theil. Malacosomacastrensis. Zool. Jahrb. Anat., Bd.
26, pp. 45-182. -

1909 Die Metamorphose der Insecten. Berlin. 56 pages.

Dewitz, J. 1902 a Untersuchungen iiber die Verwandlung der Insektenlarven.
Arch. Physiol. (Engelmann), pp. 327-340.

1902 b Weitere Mittheilungen zu meinen Untersuchungen iiber die
Verwandlung der Insectenlarven. Arch. Physiol. (Engelmann), pp.
425. 442.

Farkas, K. 1903 Beitriige zu Energetik der Ontogenese. Dritte Mittheilung.
Ueber Energieumsatz des Seidenspinnens wihrend der Entwicklung
im Ei and wihrend Metamorphose. Arch. ges. Physiol., Bd. 98, pp.
490-546.

FRENZEL, J. 1885 WHiniges tiber den Mitteldarm der Insekten iiber epithelial-
regeneration. Arch. mikr. Anat., Bd. 26, pp. 229-306.

GaNIN, M. 1876 Zur postembryonalen Entwicklung der Musciden. Warschau.
Abdruck aus den Arbeiten der 5 Versammlung russischer Naturf. und
Arzte im Warschau.

Heymons, R. 1907 Die verschiedenen Formen der Insectenmetamorphose und
ihre Bedeutung im vergleich zur Metamorphose andere Arthropoden.
Ergeb. Zool., Jena, Bd. 1, pp. 137-188.

JuspascHsanz, 8. 1910 Zur Kentniss der nachembryonalen Entwicklung de-
Stratiomyden. Jenaische Zeitschr. f. Naturwiss., Bd. 46, pp. 681-736:

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

614 ROBERT MATHESON

Lussen, H. 1907 Die innere Metamorphose der Trichopteren. Zool. Jahrb.
Abth. Anat., Bd. 24, pp. 71-128.

Karawaiew, W. 1898 Die nachembryonalen Entwicklung von Lasius flavus.
Zeitschr. f. wiss. Zool., Bd. 64, pp. 385-478.

1899 Ueber Anatomie und Metamorphose des Darmkanals der larve
von Anobium paniceum. Biol. Centralblatt, Bd. 19, pp. 122-130; 161-
ale

Kowa.evsky, A. 1885 Beitrige zur nachembryonalen Entwicklung der Musci-
den. Zool. Anz., Bd. 8, pp. 98-99.

1887 Beitrige zur Kenntniss der nachembryonalen Entwicklung der
Musciden. I Theil. Zeit. f. wiss. Zool., Bd. 45, pp. 542-594.

KUNCKEL D’Hercutats, J. 1875 Recherches sur l’organisation et le développe-
ment des Volucelles. Paris.

Mercier, M. 1906 Les processus phagocytaires pendant le métamorphose des
Batraciens Anoures et des insectes. Arch. Zool. Exp. et Gen., T. 5 (4),
pp. 1-151.

MetscuHnikorr, EK. 1883 Untersuchungen iiber die intracellulare Verdauung
bei wirbellosen Thieren. Arb. Zool. Inst., Wien., 5 Bd.

1892 Atrophie des muscles pendant le transformation des Batraciens.
Ann. Inst. Pasteur, Bd. 6.
e

Meratnikorr, 8S. 1907 Zur Verwandlung der Insecten. Biol. Centralblatt, Bd.
27, pp. 396-405.

Mosuscz, A. 1897 Ueber den Darmkanal der Anthrenus Larve, nebst bemerkun-
gen zur Kpithelialregeneration. Archiv f. Naturgesch., Bd. 62, pp.
89-128.

Nassonow, N. 1886 Zur postembryonalen Entwicklung der Ameise, Lasius
flavus. Vorlaufige Mittheilung (russisch). Sitz. d. zool. Abth. d.
Gesellsch. d. Freunde d. Naturw.

NerepHaMm, J. G. 1900 Some general features of the metamorphoses of the flag
weevil, Mononychus vulpeculus Fabr. Biol. Bull., vol. 1, pp. 179-191.

Perez, Cu. 1899 Sur la métamorphose des insectes. Bull. Soc. Ent. France,
T. 68, pp. 398-402.

1900 Sur V’histolyse musculaire chez les insectes. C. R. Soc. Biol.,
T. 52, pp. 7-8.

1910 Métamorphose de Muscides (Calliphora erythrocephala Mg.)
Arch. de Zool. Exp., T. 4 (5), pp. 1-274.

RENGEL, C. 1896 Ueber die Veranderungen des Darmepithels bei Tenebrio
molitor wihrend Metamorphose. Zeitschr. f. wiss. Zool., Bd. 62, pp.
1-60.

Roveet, C. 1900 La phagocytose et les leucocytes hematophages.* C. R. Soe.
Biol., T. 52, pp. 307-310.

FORE-GUT OF CORYDALIS CORNUTUS L. 615

Russ, E. A. L. 1908 Die postembryonale Entwicklung des Darmcanals bei
Trichopteren (Anabolia laevis Zett.) Zool. Jahrb., Abth. Anat., Bd. 25,
pp. 673-770.

ScHInDLER, K. 1903 Die Metamorphose der Insecten. Zeitschr. Naturw., Bd.
75, pp. 341-356.

TrERRE, L. 1900 a Contribution a l’étude de Vhistolyse et de l’histogenese du
tissu musculaire chez Vabeille. C.R. Soe. Biol., T. 52, pp. 91-93.

1900 b Métamorphose et phagocytose. C. R. Soc. Biol., T. 52, pp.
158-159.

Van Lrsuwen, W. 1907 Beitrige zur Kenntniss der Metamorphosen. Die
microscopische Anatomie des Darmkanals und dessen Driisen von
Isosoma graminicola Giraud. Tijdschr. d. Ned. Dierk. Vereen (2).

Vaney, C. 1900 Contribution a l’étude des phénoménes de Métamorphose chez
les Diptéres. C. R. Acad. Sci., T. 131, pp. 758-760.
1902 Contribution a l’étude des larves et des metamorphoses des Dip-
téres. These de Lyon. Ann. de l’Univ. de Lyon (n.s.) T. 1, Fasc. 9,
178 pp., 4 plates.

Van Ress, J. 1889 Beitrige zur Kenntniss der innere Metamorphose von Musca
vomitoria. Zool. Jahrb., Abth. Anat. u. Ontog., Bd. 3, pp. 1-134.

Verson, E. 1898 a La evoluzione del tubo intestinale nel Filugello. Continua-
zione e fine. Atti. Ist. Veneto Sc., T. 56, pp. 1273-1315.

1898 b_ L’evolution du tube intestinal chez le ver a Soie. Arch. Ital.
Biol., T. 30, pp. 360-362.

1905 Zur Entwicklung Verdauungscanals bei Bombyx mori. Zeitschr.
f. wiss. Zool., Bd. 82, pp. 523-600.

VIALLANES, H. 1882 Histologie et développement des Insectes. Ann. de Sci.
Nat., Zool., T. (6), T. 14 (1), pp. 348.

WerIsMANN, A. 1864 Die nachembryonalen Entwicklung der Musciden. Zeit-
schr. f. wiss. Zool., Bd. 14, pp. 264-336.

1866 Die Metamorphose von Corethra plumicornis. Zeitschr. f. wiss.
Zool., Bd. 16, pp. 45.

PLATE 1

EXPLANATION OF FIGURES

1 Ventral view of the intestine of Corydalis, showing tracheal connections;
1th, first thoracic spiracle; 2th, second thoracic spiracle; 1ab, 2ab, 3ab, 4ab, abdomi-
nal gills; k, gizzard; n, nerve. The remaining letters are fully explained in the
text.

2 A more enlarged view of the tracheal supply of a portion of the intestine.
Letters fully explained in text.

3 <A view of the muscular system of the fore-gut, showing the wonderful net-
like arrangement.

4 The gizzard spread open; s.r., small ridges; l.r., large ridges. The posterior
end of the gizzard is at the upper part of the photograph.

5 One of the large ridges of the gizzard in lateral view. a.t., anterior tooth;
p.t., posterior tooth.

6 One of the large ridges of the gizzard in side view. There are three of the
ridges as shown in figure 5, three as in figure 6, and six small ridges as shown in
figure 7 in each gizzard.

7 One of the small ridges of the gizzard.

8 Cross-section of the oesophagus, showing the numerous folds.

9 Cross-section of the fore-gut near the middle of the gizzard.

616

FORE-GUT OF CORYDALIS CORNUTUS L.
ROBERT MATHESON PLATE 1

517

PLATE, 2

EXPLANATION OF FIGURES

10 The point of union between the epithelium of the fore- and mid-guts. ep.,
epithelium of the mid-gut; g.e., glandular appearing epithelium of the fore-gut.

11 Fore- and mid-guts of the pupa just after pupation. In this case the intima
of the gizzard remained lodged in the anterior end of the oesophagus.

12 Cross-section of the proventriculus of the fore-gut showing the four promi-
nent chitinous ridges which project into the mid-gut.

13 Cross-section of the fore-gut of a pupa about twenty-four hours old l.m.,
longitudinal muscles; c.m., circular muscles; fb., remnants of the breaking down
muscle fibrillae; «7, degenerating nuclei. Within the lumen lies the loosened intima.

14 A more highly magnified section of the epithelium of figure 13 to show the
degenerating nuclei (x).

15 A lateral view of the muscles of the pharynx, showing the dorsal muscles in
place.

16 Ventral view of the pharynx showing the lateral muscles.

618

FORE-GUT OF CORYDALIS CORNUTUS L.
ROBERT MATHESON PLATE 2

619

PLATE 3

EXPLANATION OF FIGURES

17 Cross-section of the wall of the larval fore-gut to show the structure of the
epithelium and muscles. The following series of photographs are all taken at
about the same point to show the series of changes in the epithelium and muscles
during metamorphosis.

18 Cross-section of intestinal wall of a prepupa showing the earliest changes
in the epithelium and muscles.

19 Cross-section from the same region as figure 18 from a prepupa in which the
intima has been cast off.

20 Cross-section of the intestine of a young pupa. The nuclei of the muscles
are now much more prominent and the pupal epithelium has secreted a new intima.

21 From a pupa about twenty-four hours old. Note the degeneration of the
muscles, only a densely staining mass of fibrillae remaining. The epithelium has
become much enlarged and vacuolate.

22 From a pupa ninety-six hours after pupation. The pupal intima has just
been shed; x, degenerating nuclei. Note the embryonic appearance of the muscle
nuclei. The histolysis of the muscles is now complete and we have the beginnings
of histogenesis.

620

FORE-GUT OF CORYDALIS CORNUTUS L.

ROBERT MATHESON PLATE 3

PLATE 4

EXPLANATION OF FIGURES

23 From a pupa slightly older than that of figure 22. Histogenesis has pro-
eressed somewhat further. Note the cut ends of fibrillae in the longitudinal mus-
cles (l.m.).

24 A stage yet more advanced in the histogenesis of the imaginal muscles.

25 Note the forming fibrillae in the circular and longitudinal muscles. The
muscles are now arranged in their two definite layers and the epithelium is now
reformed. From an older pupa.

26 Cross striations can now be observed in the newly re-formed circular muscles
(c.m.). These striations are faint and remain so till the emergence of the adult.
g.m., granular mass occupying the space between the loosened intima and the
epithelium. From a pupa 160 hours old.

27 This figure shows the characteristic method of cell division; m, dividing
cell; sf., strand connecting the cell with the basement membrane; b.m., note the
distinct basement membrane between the epithelium and the muscular layer.

28 Shows a condition similar to figure 27.

29 The epithelium from a pupa forty-eight hours old; 7, row of black dots near
the outer ends of the epithelial cells where the imaginal intima will be re-formed.

FORE-GUT OF CORYDALIS CORNUTUS L.

ROBERT MATHESON PLATE 4

es

CHROMODORIS ZEBRA HEILPRIN: A DISTINCT
SPECIES

W. M. SMALLWOOD anv ELIZABETH G. CLARK

Contributions from the Bermuda Biological Station for Research No. 26, and from
the Zoological Laboratory of Syracuse University

SIX FIGURES

The following paper supplements one published by the senior
author (Smallwood, ’10). The two papers present the general
morphology of Chromodoris zebra and point out the characters
which distinguish it from C. villafranca. Bergh (’92) believed
the two to be identical. The first paper gave a description of
the external anatomy of C. zebra—the present one, the internal
anatomy of the species. While our study of the internal anatomy
has shown no marked deviation from the general plan of organi-
zation of the Doridae,—so well described by Eliot (710, pp. 36 to 49)
with Doris tuberculata as example,—it has furnished information
concerning the anatomical details of this species not hitherto
published.

DIGESTIVE SYSTEM

The mouth leads into a short but rather wide atrial chamber
(6 mm. long by 9 mm. in diameter), the inner wall of which is
lamellated. At the posterior end of this chamber a ring-shaped
depression forms the boundary between the lateral wall and the
posterior wall, the latter constituting the labial disc. <A vertical
slit-like opening in this disc is the entrance to a passage extending
posteriorly for 4 mm.; this passage is lined with closely set spines,
which form what is sometimes incorrectly termed a ‘prehensile
collar.’ This spinous passage expands at its posterior end into
the main buccal cavity, 5 mm. in length by 6 mm. in diameter.
The floor of the main buccal cavity is formed by the odontophore,
a round muscular prominence bearing the radula and divided into

625

626 W. M. SMALLWOOD AND ELIZABETH G. CLARK

lateral halves by a longitudinal furrow. The walls of the buccal
cavity are chiefly composed of muscles and are greatly thickened
ventrally. The entire mass is cylindrical (fig. 2, phz.).

The radula, borne upon the rounded prominence forming the
floor of the buccal cavity, is folded in the middle and fits into the
longitudinal groove; it emerges from the curved sac which lies
at the ventro-posterior angle of the buccal mass. This sac is
lined with cells whose special function is the secretion of new teeth.
The total number of rows of teeth is 95 to 100; of these 20 to 30
are in the radula sac, the posterior 4 to 5 being in process of forma-
tion. The median tooth is absent. In each transverse row are
165 to 170 teeth on each side. All of the pleural teeth bear two

Fig. 1 Chromodoris zebra Heilprin. Viewed from the right side; the branchial-
rosette turned toward the observer; 2 natural size. Copied from Smallwood, ’10.

denticles; one long denticle terminating in a sharp slender point,
and a shorter one with a slightly blunter distal end (fig. 3).
Slight variations in size and form of teeth occur, as may be seen
in figure 3, and as they near the margin, we find that the
denticles gradually become more and more blunt, until at the
outermost edge, they are mere rounded knobs supported by the
broader base. In some teeth, the denticles are entirely lacking,
doubtless having been worn off. The posterior margin of the
tooth, that is, the side toward which the denticles point, is slightly
crenulated. All teeth are of nearly the same size, measuring 20
micra long, 9 micra wide, and 3 micra thick. The longest denticle
measured was 6 micra long.

Bergh (’78, pp. 28, 29 and 30) gives the following for Chromo-
doris villafranca:

Forma corporis gracilis. _Pedamentum foliorum branchialium humile.

Color paginae superioris glaucescens, fasciis transversalibus latioribus .
caeruleis dilutis et praesertim lineis fulvis (ut plurimum 7) non semper

CHROMODORIS ZEBRA HEILPRIN 627

ga. ----- fHN
or SOP

| \WON
<anen) ri
wiee ahaa == UN.

-£----ven= prer.dt

Fig. 2 Digestive and renal systems with the intestine displaced; ga., stomach;
gl.sal., salivary gland; iglv., crop; in., intestine; oes., oesophagus; phz., buccal
mass; ren., kidney; ren.-pi’cr.dt., reno-pericardial duct; wrt., ureter.

Fig. 3 Pleural teeth, showing the successive steps in lossof denticles from a to
g, also the arrangement of the teeth in relation to each other.

628 W. M. SMALLWOOD AND ELIZABETH G. CLARK

inter se distinctis et saepe divisis ornatus; margo dorsalis fulvus; eclavus
rhinophoriorum caeruleus, margine posteriore linea fulva ornatus; folia
branchialia rhachide extus linea punctorum fulvorum ornata; latera
corporis caerulescentia lineis fulvis 3-4 longitudinalibus (ut in dorso)
ornata; podarium infra caerulescens vel glaucescens.

Dentes radulae grosse denticulati.

Hab. M. mediterr. (M. adriat., Napoli, Panormi, Genova, Bona).

Die Mundroéhre etwa 4 mm. lang, von schén griin-blauer Farbe, ganz
wie bei der vorigen Art [C. elegans]. Der Schlundkopf etwa 6 mm. lang
bei einer Breite von 3 und einer Hohe bis 3 mm.; die Raspelscheide hin-
ten und unten am Schlundkopfe etwa 1 mm. hervortretend, mit der
schmutzig gras-griinen Raspel stark hindurchschimmernd. Die Lippen-
platten grau-gelb, sonst ganz wie bei der vorigen Art; die Elemente
derselben ganz wie bei jener, vielleicht etwas deutlicher am Grunde des
Hakens gestreift. Die Zunge wie oben; die Raspel zeigte 26 Zahnplat-
tenreihen, von denen die ersten 5 incomplet; weiter gegen hinten fanden
sich noch 82 entwickelte und 5 junge Reihen, die Gesammtzahl dersel-
ben somit 63 betragend. In den hintersten Reihen der Zungefanden
sich (jederseits) 85 Platten, und die Anzahl stieg weiter gegen hinten
bis 90. Die Platten von schwach-gelblicher Farbe, denen der vorigen
Art im Ganzen ziemlich aihnlich; die Hohe der diussersten meistens etwa
0,04-0,07 mm. betragend, die der Platten sich im Ganzen bis 0,14,
mm. erhebend. Die dussersten (Fig. 9a) Platten niedriger und mit
stairkerer Denticulation. Die Platten im Ganzen stirker und alle am
Aussenrande und zwar stark denticulirt (Fig. 5-9), die innerste Platte
mit einem starken Dentikel innen am Grunde des Hakens (Fig. 5aa.).

A comparison of these two descriptions shows the following
differences: The rows of teeth in C. zebra are 95 to 100, in C.
villafranca 63. The number of teeth in a row in C. zebra is 165
to 170, while in C. villafranca it is but 85 to 90. The form of the
teeth as shown in Bergh’s figures indicates clearly that the two
forms belong to distinct species.

The oesophagus

The oesophagus of C. zebra leaves the pharynx at the dorso-
posterior angle of the buccal mass. On each side, in a slightly
dorsal position, the ducts from the single pair of salivary glands
(fig. 2, gl.sal.) enter the oesophagus. These glands are long rib-
bon-like structures, folded along their mid-line and somewhat
irregular on the margins. They are 3 cm. long and 2 mm. wide.
From this point the oesophagus (oes.) continues posteriorly as

CHROMODORIS ZEBRA HEILPRIN 629

a slender tube for 1 cm., passes through the nervous collar, and
then expands into a thin-walled sac (15 mm. long by 5 mm. in
diameter), the crop (iglv.).. From the crop it continues, asa
slightly broader tube (3 mm. in diameter), to the anterior portion
of the liver mass, where it passes directly into the anterior end
of the stomach (fig. 2, ga.). Throughout its entire length, the
thin walls of the oesophagus bear longitudinal lamellae on their
inner surface.

The stomach

The stomach is a large thin-walled sac, 30 mm. long by 15 mm.
in diameter, lying within the anterior two-thirds of the liver
mass. Its walls are slightly lamellated and perforated by innu-
merable large openings, the orifices of the evaginations which col-
lectively constitute the ‘liver’ or digestive gland, so that appar-
ently the cavity which we call the stomach is merely enclosed in a
loose meshwork. Closely applied to the walls of the stomach
are the ramifications of the nephridial organ (fig. 2, ren.).

The intestine

The intestine (fig. 2, in.) arises from the right side of the stom-
ach about two-thirds the length of the stomach from its anterior
end. This tube, at first rather narrow, has its greatest diameter
(7 mm.) near its emergence from the stomach, and it gradually
grows smaller as it nears the anal opening. From the stomach,
the intestine passes anteriorly for about 25 mm. to the left side
of the anterior end of the liver mass, where it turns abruptly to
the right and runs diagonally crosswise and backwards for 10
mm.; then it bends again, continuing posteriorly for 40 mm., until
it reaches the anal opening within the branchial circle. The walls
bear longitudinal lamellae upon the interior, two especially con-
spicuous folds lying side by side upon the median ventral surface.

JOURNAL OF MORPHOLOGY, VOL. 23. No. 4

630 W. M. SMALLWOOD AND ELIZABETH G. CLARK
The liver

The liver occupies the posterior portion of the body cavity.
It is a large pyriform organ, the broad anterior end being slightly
bilobed in front. It is surrounded almost completely by the her-
maphroditic gland, whose ramifications may be seen lying on the
surface of the dark colored liver. In this species, the liver com-
pletely envelopes the stomach and the larger part of the kidney.

RENAL SYSTEM

It is impossible to determine anything very definite about the
extent of the nephridial organ without careful histological study.
However, it is possible to say that it consists of a main trunk with
numerous lateral anastomosing ramifications, which cover the
greater part of the dorsal surface of the stomach and probably
run into the liver. <A short ureter (fig. 2, wrt.) leads from the
kidney to the excretory orifice, a minute opening within the bran-
chial circle close to the anal papilla. With the main trunk, where
it merges into the ureter, is joined another duct (ren.-pi’cr.dt.),
which arises from the pericardium, where a valvular enlarge-
ment allows the passage of fluid from the pericardium to the renal
trunk.

RESPIRATORY AND CIRCULATORY SYSTEMS

These two closely related systems embrace the branchiae, the:

heart, the blood vessels and blood spaces, together with the blood
gland. The heart lies within the pericardium on the upper sur-
face of the posterior portion of the liver mass just anterior to
the branchiae. It consists of an anterior muscular ventricle
(fig. 4, ».) and a posterior thin-walled auricle (aurl.). From the
apex of the ventricle a large blood vessel runs forward, giving off
lateral branches to the liver and other organs. Just anterior to
the liver mass occurs an enlargement, which sends one vessel
(va.sng.gen.) to the genitalia, one to the blood gland (gl.sng.) and
another (va.sng.buc.) which continues anteriorly to surround the
buccal mass. Throughout the integument are numerous blood
spaces, which lead to the posterior end of the body and pour

_—

CHROMODORIS ZEBRA HEILPRIN 631

their contents into the lateral blood vessels which enter the auricle
at its posterior angles. It is probable that, as occurs in other
forms, the entire primary body cavity surrounding the various
organs is filled with the blood fluid and that this is collected and
aerated in the branchiae, from which it passes directly into the
middle portion of the auricle through the posterior median blood
vessel of the auricle.

---- va.sng.buc

SS 1bKob iS) no. SEN.

Fig. 4 Heart, blood gland and some of the more important blood vessels.
aurl., auricle; gl.sng., blood gland; v., ventricle; va.sng.buc., blood vessel to buccal
mass; va.sng.gen., blood vessel to genitalia.

The blood gland (fig. 4, gl.sng.) is an irregular leaf-like mass
made up of very small round cells. It lies immediately above the
central nervous system and has an antero-posterior diameter of
7 mm. and a lateral diameter of 5 mm.

There are from 12 to 14 branchiae in the branchial circle. Each
branchia is provided with an afferent and an efferent blood vessel,
which carry the blood respectively to and from the branchiae.

632 W. M. SMALLWOOD AND ELIZABETH G. CLARK

NERVOUS SYSTEM

The nervous system consists of two main divisions—the
central and the visceral or sympathetic part. The central part
of the nervous system, enclosed within a semi-transparent mem-
branous capsule, forms a nerve collar around the oesophagus,
lying just beneath the blood gland. It is made up of three pairs
of large ganglia,—two cerebral, two pleural and two pedal,—
a small pair of buccal ganglia, and several minor ones.

Fig. 5 Central nervous system seen from above and posteriorly. co’ms.cb.,
cerebral commissure; co’ms.pl.+pd., fused pleural and pedal commissures;gn.buc.,
buccal ganglion; gn. oes., oesophageal ganglion; gn.olf., olfactory ganglion; gn. pd.,
pedal ganglion; gn.pl., pleural ganglia; n.mtl., nerves to mantle; n.olf., olfactory
nerve; n.pd., pedal nerves; n.vsc., nerve to sympathetic system: oc., eye; xx, ry,
nerves to liver and genitalia.

The cerebral ganglia are the most anterior pair and are united
with each other by a short (0.5 mm.) commissure (co’ms.cb.),
as seen in figure 5. From the anterior margin of each projects
the almost sessile olfactory ganglion (gn.olf.), which gradually

CHROMODORIS ZEBRA HEILPRIN 633

tapers off into the large olfactory nerve, that passes into the base
of the rhinophore. Upon the dorsal surface, near the external
margin of the cerebral ganglia, arise the two small optic ganglia,
each of which sends a short nerve to the corresponding eye (oc.).
Each cerebral ganglion gives rise to three pairs of nerves, which
pass forward and innervate respectively the oral tentacles, the
mouth, and the lips. Besides these nerves, each ganglion sends
from its mid-ventral surface a connective which passes around the
oesophagus to the buccal ganglia (gn.buc.); these lie, side by side,
ventral to the oesophagus just behind the buccal mass. <A slender
nerve leads forward from each buccal ganglion to the tongue.
Two lateral nerves from each buccal ganglion innervate the buc-
cal mass. Posteriorly, a small ganglion (gn.oes.) is joined to each
buccal ganglion by an exceedingly short‘nerve. This oesophageal
ganglion sends down the oesophagus a single nerve, which has
three branches that spread over the tube.

Lying more dorsal and slightly posterior to the cerebral ganglia,
to which they are joined by very short connectives, are the equally
large pleural ganglia (gn.pl.). These generally give rise to three
pairs of nerves which innervate the mantle. From one of them
arises on the right side a slender nerve (n.vsc.), which communi-
cates with the visceral part of the nervous system.

The pedal ganglia (fig. 5, gn.pd.), not so closely united to the
others, lie in a more lateral position. From 4 to 5 pairs of nerves
arise from them. Of these, two pairs of the smaller nerves
innervate the body wall, while of the remaining larger ones at
least two pairs pass to the foot as anterior and posterior pedal
nerves.

The pedal and pleural commissures (co’ms.pl.+pd.) unite into
a single broad band, which encircles the oesophagus ventrally,
the nerve fibers arising in each case from the posterior margin of
the ganglion. Two large nerves (az.ry.), arising close to this
commissure, innervate the liver and genitalia. Their precise
origin is difficult to ascertain. ‘The remaining pairs of nerves are
distributed posteriorly to the mantle.

The study of the central nervous system of ane nudibranch
brought out the fact that there is considerable variation in the total

634 W. M. SMALLWOOD AND ELIZABETH G. CLARK

number of nerves, as well as the number that arise from each
ganglion. The apparent origin of the nerves as well as the rela-
tion of the ganglia to each other is variable. In some of the speci-
mens studied the three pairs of ganglia comprising the nerve
collar were so closely crowded together that it was well nigh im-
possible to discover their true relations.

The visceral or sympathetic part of the nervous system, as
stated above, communicates with the cerebral nervous mass by
means of a nerve (n.vsc.) from the right pleural ganglion. This,
the visceral portion of the nervous system, consists of an irregular
system of exceedingly fine nerves spreading entirely over the vis-
cera and containing numerous small ganglia distributed along _
their course.

REPRODUCTIVE SYSTEM

The reproductive system is made up of two main portions; one,
known as the anterior genital mass, is an irregularly rounded
body made up of a complex of tubules, cavities and ducts, lying
at the right anterior end of the body cavity; the other, the pos-
terior genital mass, is the hermaphroditic gland, which spreads
over the liver and is joined to the anterior genital mass by the
hermaphroditic duct. The size and shape of both of these por-
tions varies greatly at different seasons of the year, being enor-
mously enlarged during the breeding season, the time at which
this material was collected.

The posterror genital mass

The hermaphroditic gland spreads over the entire surface of
the liver, from which it may be distinguished by its lighter color.
It is made up of numerous lobules opening into a series of minute
tubules, which finally emerge into two large ducts; these unite to
form the hermaphroditie duct. In this gland both eggs and
sperm have their origin. As the hermaphroditic duct (fig. 6,
dt.her.) leaves the gland, it passes to the anterior genital mass,
which it enters on the median side. It immediately enlarges
into a thickened tube, the ampulla (amp.), which is 2 mm. wide

CHROMODORIS ZEBRA HEILPRIN 635

and about 5mm. long. A little beyond the ampullar enlargement
the duct bifurcates, giving rise to the vas deferens (va.df.) and the
oviduct (o’dt.).

The anterior genital mass

The anterior genital mass is made up of male and female geni-
talia. The male genital apparatus is fairly simple. It consists
of a coiled tube, the vas deferens, which extends from the bifur-
cation of the hermaphroditic duct to the enlargement known as
the penis sheath (ivlr.pe.). From the point of bifurcation it

Sp’ the.

Fig. 6 Diagram of the reproductive glands and ducts. amp., ampulla; atr.gen.,
genital atrium; dt.gl.muc., duct of mucous gland;dé. her., duct of hermaphroditic
gland; gl.albm. albumen gland; gl.muc., mucous gland; o’dt., oviduct; ivir.pe.,
penis sheath; sp’cys., spermatocyst; sp’thc., spermatotheca; va.df., vas deferens;
vg., vagina.

gradually enlarges, the enlarged portion being thrown into numer-
ous folds, which lie on the anterior mucous-gland portion of the
mass. ‘The terminal portion, perhaps a third of the length of the
whole, is again narrowed and is very sinuous; it only gradually

636 W. M. SMALLWOOD AND ELIZABETH G. CLARK

enlarges to become continuous with the sheath of the penis, which,
in turn, communicates with the genital atrium (aér.gen.).

The female genitalia are more complex. They may be roughly
divided into two parts; the uterine or oviducal portion, and the
portion specialized for the reception and storage of sperm, which
is in direct communication with the oviduct. The oviduct
leads from the bifurcation of the hermaphroditic duct directly
to the albumen gland (gl.albm.), a large mass of closely convo-
luted tubules constituting the posterior portion of the anterior
genital mass. From the albumen gland the duct passes forward
into the mucous gland (gl.muc.), and terminates in the vagina
(vg.), which is a single broad tube leading to the atrium. The
mucous gland has a distinct duct (dt.gl.muc.) leading to the genital
atrium. At its deep end the vagina expands into a rather large
globular sac, which seems to be filled with detritus and a few free
spermatozoa; this is the ‘spermatotheca’ of Bergh. About half-
way between the external opening of the vagina and the spermato-
theca a much smaller pear-shaped sac (sp.cys.) with thicker walls
is connected with the vagina by a simple duct. This sac is the
spermatocyst. It contains large numbers of spermatozoa. Just
beneath this spermatocyst, on the ventral side of the vaginal tube,
is the orifice through which the oviduct communicates with the
vaginal portion of the female genitalia.

BIBLIOGRAPHY

Beran, R. 1878 Untersuchung der Chromodoris elegans und villafranca. Mala-
kozool. Blatter fiir 1878, pp. 1-36, Taf. 1 und 2.

1892 System der nudibranchiaten Gasteropoden. In Karl Semper’s
Reisen im Archipel der Philippinen, Theil 2, Bd. 2, Heft 13, pp. 993-1165.
Also separately, Wiesbaden: Kreidel, 173 pp.
Extot, C. 1910 A monograph of the British Nudibranchiate Mollusca, Part viii
(Supplementary). London, Ray Society, 198 pp., 8 pls.

Heiuprin, A. 1889 The Bermuda Islands. Philadelphia [vi] + 231 pp., 17 pls.

Smatuwoop, W.M. 1910 Notes on the Hydroids and Nudibranchs of Bermuda.
Proc. Zool. Soc., London, Pt. I, pp. 187-145. (Contrib. Bermuda
Biol. Sta. No. 18.)

PRIMITIVE REPTILES

A REVIEW

S. W. WILLISTON
From the Paleontological Laboratory, University of Chicago

ONE FIGURE

The evidence seems now conclusive that the extensive Ameri-
can reptilian fauna hitherto called Permian is in part of upper
Pennsylvanian, in part of lower Permian age, and, therefore,
is the oldest known. From other parts of the world there are
only a few known forms supposed to be of equivalent, or approxi-
mately equivalent, age, the chief of which, if not the only ones,
are Stereosternum and Mesosaurus from the Santa Catherina
System of Brazil, the latter genus also from the upper Dwyka of
South Africa. From immediately superjacent beds in South
Africa, doubtless of lower Permian age, but one or two reptiles
are known, Archaeosuchus and Eccasaurus, referred by Broom,
the former at least, to the dinocephalian group of the Therapsida.
From the Beaufort beds, upper and lower, of Africa, numerous
genera of reptiles are known, referred to the Cotylosauria and
Therapsida. Because of the close affinity or identity of some
of these genera with those found in Russia, they doubtless-should
all be considered of upper Permian age, as distinguished from
lower Permian. From the Rothliegende of Germany and France,
of lower Permian age, the following genera of reptiles are known:
Stephanospondylus, Phanerosaurus, Datheosaurus, Stereorhachis,
Kadaliosaurus, Callibrachion, Aphelosaurus, and Paleohatteria
(Haptodus). Leaving Archaeosuchus and Eccasaurus out of
account in the following discussion, the present paper will deal
with the carboniferous and lower Permian reptiles only, including

637

638 S. W. WILLISTON

the genera mentioned above and the numerous ones known
from the American Permio-carboniferous. ‘Two other, disputed
groups of contemporary air-breathers, believed to be reptiles
by some authors of repute, will be discussed in the sequel.

All of the so-called orders represented by these reptiles are
believed to be continuous, not only throughout the Permian
of Europe or Africa, but also throughout more or less of the trias
if not the mesozoic, though in most cases the later forms are
either distinctly modified, or have reached a higher degree of
specialization; and their discussion may be omitted here. Pareia-
saurus and Propappus, for instance, from the upper Permian
of Russia and Africa, have dorsal osseous scutes, a distinct acromi-
al process of the scapula, and coéssified caleaneum and astragalus,
all unknown in the American Cotylosauria; characters, which
in themselves are sufficient to separate the Pareiasauria as a
distinct group. Procolophon, also, with its allies Telerpeton,
Sclerosaurus and Koiloskiosaurus from the trias, have departed
so far from the primitive simplicity of the earlier types of America
in the skull structure and in the pectoral girdle as stated
by Huene, that they may well be separated into a group by
themselves. Procolophon has been supposed for years, to be a
primitive form of the double-arched or ‘Diapsidan’ type, notwith-
standing its occurence in much later deposits than those of Paleo-
hatteria. Huene has recently expressed the opinion, one I have
had for the past five years, that the Procolophonia have nothing
to do genetically with the rhynchocephalian or archaeosaurian
phylum.

Using then the term Permocarboniferous to include the upper
carboniferous and lower Permian, and omitting the imperfectly
known Archaeosuchus and Eccasaurus, we have perhaps thirty
or more genera of undisputed reptiles which it will be of interest
to compare directly.

The first and lowest group of these, the Cotylosauria as usually
accepted, began in the upper Pennsylvanian of North America
and continued into the Procolophonia of the trias of South Africa
and Europe; it has generally been believed to be the order from
which all later Amniota have been derived. The second order,

PRIMITIVE REPTILES 639

the Theromorpha of Cope as limited by me,! in part the Pely-
cosauria of authors, began also in carboniferous times and extend-
ed as an order into triassic times in Europe. The third group,
the Proterosauria, as represented by Paleohatteria, Proterosaurus
and less well known allied forms, began, so far as we know, in
the middle Rothliegende of Germany and continued throughout
the Permian; it is generally supposed to be the ancestral group.
from which all later true reptiles, that is the double-arched forms,
arose, and is often included among the Rhynchocephalia, sens.
lat. The fourth group, the Proganosauria, as originally defined
by Baur and later restricted by Osborn and McGregor, has been
accepted by some, though not by all students as an early group
of the Sauropterygia, which continued to near the close of meso-
zoic times.

From a study of all these I have endeavored to summarize
and correlate those structural characters upon which phylogenies
and classification must depend. Many of the known forms are
as yet represented by fragmentary and incomplete specimens
only, whether those of America or of other lands; of the latter
I have little or no autoptic knowledge, a fact less to beregretted
since most of them, and those the most important ones, have
been described and illustrated by competent students. Of the
American forms I have studied with more or less care nearly
all the known material, with the exception of that preserved in
the museum at Munich, which has been ably discussed by Broili.
Of this material I am familiar with complete or nearly complete
skeletons pertaining to the following genera: Diadectes Cope,
Diasparactus Case, Limnoscelis Will., Seymouria Broili, Cap-
torhinus Cope, Labidosaurus Cope, Varanosaurus Broili, Casea
Will., Ophiacodon Marsh and Dimetrodon Cope. I have also
studied the less complete specimens belonging in the following
genera: Diadectoides Case, Pantylus Cope, Edaphosaurus
(Naosaurus) Cope,? Sphenacodon Marsh, Clepsydrops Cope,
Araeoscelis Williston and Animasaurus C. and W., together with
the more fragmentary material of nearly all the other known gen-

1 American Permian Vertebrates, p. 70.
2 The identity of Naosaurus with Edaphosaurus is now established.

640 S. W. WILLISTON

era from Illinois, Kansas, Oklahoma, Texas, New Mexico and
Colorado. And I need not add that I have diligently studied
all the literature bearing on these American reptiles, especially
that of Case and Broili. I am indebted to Dr. Case for the
kindly criticism of the manuscript of the present article.

It will be understood that I give the characters so far as they
seem to be established by numerous forms, though it is quite
possible that future discoveries of new forms, or of better material
of forms now imperfectly known, may require the removal of
some of the characters included under the constant list to the
inconstant or variable list. But the converse will not be
possible unless I have committed errors, since here the evidence
is positive, not negative. Among so many and diverse forms
it must be evident that most of the characters common to all,
so far as the material goes, are to be considered as primitive
ones, that is, characters possessed by the earlier or earliest rep-
tiles far back in Pennsylvania times. But, by no means are
these the only primitive characters, since it is very evident that
even by the close of the carboniferous not a few had become
variable.

CONSTANT CHARACTERS OF PERMOCARBONIFEROUS REPTILES

Crawling reptiles from about 0.3 to about 23 meters in length.
A parietal foramen, postorbital, postfrontal, lacrimal, nasal and
probably at least one additional membranous cranial bone not
occurring in Sphenodon present in the skull. Quadratojugal,
paroccipital, epipterygoid and septomaxillary probably always
distinct; parietals, frontals and nasals never fused in middle
line; stapes probably always large; a separate prearticular® bone;
splenial entering symphysis; coronoid small; teeth on paired
premaxillae, maxillae, palatines, pterygoids and dentaries; ptery-
goids articulating with vomers (prevomers), basisphenoid, pala-
tines and quadrates; quadrate always fixed, though not always
firmly attached.

’ This term (1903) had become established in paleontological literature before
Gaupp (’08), in ignorance of its use, proposed the name goniale for the same bone.

PRIMITIVE REPTILES 641

Vertebrae notochordal (or at least deeply and conically bicon-
cave); all vertebrae to the second-sixth caudal with intercentra
and all to the eighth or ninth with ribs, those of the lumbar,
sacral and basal caudal united more or less suturally with body
and arch, but in all cases the capitulum more or less intercentral
in position. Capitulum of free ribs articulating intercentrally,
resting more or less on the front end of the centrum, but not
articulating with a parapophysial process. Chevrons present
on all the caudals, save the first two-six and the extreme terminal
ones. Proatlas probably always present. Atlas composed of
a large free odontoid supporting the paired neural arch chiefly,
and an intercentrum not much larger than the one following

Fig. 1 Proatlas, atlas and axis of Ophiacodon

between atlas and axis. Axis with anterior zygapophysial facets
for articulation with atlas.

Interclavicle with expanded anterior part and a long posterior
stem; clavicles large, smooth, articulating with interclavicle,
scapula and sometimes with the cleithrum; scapula and coracoid
suturally united, fused in maturity; anterior coracoid, the so-
called procoracoid (the true coracoid of later reptiles) always
forming a part of the glenoid articulating surface; a supracoracoid
and a supraglenoid foramen always present; no ossified sternum.
Pubis and ischium plate-like, the former pierced by a pubic fora-
men (rarely a notch), and both meeting their mates in a median

642 Ss. W. WILLISTON

symphysis. Acetabulum imperforate, formed by all three bones;
pubis, ischium and ilium more or less fused in adult.

Humerus expanded at its two ends into more or less divergent
planes; an entepicondylar foramen present; olecranon more or
less produced; radiale, intermedium, ulnare, centrale, and four
distalia always ossified, a second centrale and the fifth distale
rarely unossified; pisiform usually (perhaps always) present.
Two bones in proximal row of tarsus (caleaneum and astragalus),
at least one free centrale (rarely unossified) and five distalia.
Manus and pes pentadactylate, the phalangeal formula 2, 3, 4,
5, 3-4; a vacuity for the mesopodial perforating artery in both
carpus and tarsus (Limnoscelis?, Diasparactus’?).

INCONSTANT OR VARIABLE CHARACTERS

Skull smooth or rugose, with two, one, or no temporal perfora-
tions on each side; rarely with spiny protuberances; postparietal,
tabulare and supratemporal inconstant, primarily with all known
membrane bones of the Stegocephalia; lacrimal usually extend-
ing to nares; transpalatine not yet demonstrated, quite surely
absent in some forms; no posterior palatine vacuity hitherto
observed; basisphenoid suturally or loosely articulated with
pterygoids.t| Teeth usually thecodont, rarely acrodont, usually
attached to vomers (prevomers), rarely to splenials. Occipital
condyle convex, flat or concave distally.

From two to six or seven cervical vertebrae; twenty-three
to at least twenty-seven presacral; one to three sacrals; and from
about twenty-five to more than fifty caudals.

Articulation of ribs either continuous from capitulum to tuber-
culum, or separated by a distinct neck; that is, the ribs are
either single-headed or double-headed. Spines of thoracic verte-
brae rudimentary, or short, or moderately elongated, or greatly
elongated. Ventral ribs present or absent.

4 Incidentally I may remark here that Versliiys’ recent identification (Zool.
Jahrb., 1912, 651) of the basisphenoid in the American pterodactyl Pteranodon
is indefensible. The basisphenoid invariably forms part of the brain capsule,
lodging the hypophysis. As Versliiys would identify the bone in Pteranodon it

is far removed from all relations with the brain-case or the cranial bones. Eaton,
I believe, is right in restricting the basioccipital to the extreme posterior part.

PRIMITIVE REPTILES 643

Cleithrum well developed, vestigial or wanting; posterior
coracoid bone (the so-called true coracoid bone) usually ossified,
its suture more persistent than that between the scapula and
anterior coracoid; sometimes unossified (Seymouria, Varanosau-
rus), a coracoid emargination, as in the Lacertilia, rarely present.
Pelvis with or without a median puboischiadic vacuity.

Humerus rarely with an ectepicondylar foramen, usually with
an ectepicondylar groove; terminal phalanges flattened or claw-
like. A second centrale pedis rarely present. Articular extremi-
ties of long bones sometimes imperfectly ossified in the adult.

Herbivorous (Casea); insectivorous (Seymouria); carnivorous
(Dimetrodon); or eaters of crustacea, and other invertebrates
(Diadectes?, Pantylus, Edaphosaurus, et cetera).

Subaquatic, littoral, cursorial or climbing reptiles.

The above characters, both ‘constant’ and ‘inconstant,’ are
drawn exclusively from the American forms. To include foreign
forms will require some additions to the latter list: Aquatic,
piscivorous reptiles, with elongated neck, posterior nares and
other aquatic adaptations (Proganosauria); sclerotic plates in
orbits, supracoracoid foramen a notch (Paleohatteria). ;

The foregoing characters are the chief ones upon which we
must depend for the present at least, for the classification of
the known Permocarboniferous reptiles. Doubtless when we
shall have become more intimately familiar with the skull struc-
ture in more forms not a few others will be added.°

These characters-would seem to be ample for the separation
of the known genera into various orders; and they would be if

> At the present time there is much doubt regarding the intimate structure of
the temporal and posterior cranial region in nearly all the genera here under dis-
cussion. In only the Captorhinidae of the Cotylosauria, is the controversy
nearly at rest. In this family (Captorhinus, Labidosaurus) it is now known that a
postparietal, a so-called tabulare, a squamosal and a quadratojugal are present
only. The skull-structure of no single theromorph is beyond controversy. As
to the real homologies of the various bones of the temporal region of the reptiles
we are no nearer the solution than we were a score of years ago; no one really
knows what the homology of the mammalian squamosal is in reptiles. Eighteen
names—twenty if we take into consideration the different usages of supratemporal
and squamosai—have been applied to the four elements more usually called ‘epi-
otic,’ supratemporal, squamosal and quadratojugal, and, except for unifornity

644 S. W. WILLISTON

we could distribute them at our pleasure. Asa rule in taxonomy
we depend not so much upon unique distinctive characters as
we do upon different assemblages of characters common to various
groups. A half dozen characters, if constantly associated in
any given group would serve for its delimitation, the rank of
the group being dependent upon the relative importance of the
characters themselves. And the taxonomist is suspicious of
any generic, family or ordinal classification based upon single
characters. Rightly interpreted, such single characters, unaccom-
panied by other differences, may be and often are merely muta-
tions or sports, unstable and of little significance. From which
remarks it follows that I have no faith in the ‘Mutation theory’
of the origin of species, nor do I believe that any paleontologist
can defend such a theory. And, frankly, neither do I believe
that any theory of the origin of species, or of evolution even,
can ever get very far when time is left out of account. If, in
any series of phylogenetic forms we find a gradual transformation
of structure, the gradual acquisition of new characters, we do
right in uniting them all in a single group, for the sole end of
all taxonomy is phylogeny.

Let us now attempt to define the two larger groups, the Cotylo-.
sauria and Theromorpha, in order to discover their distinctive
or associated characters. Neither order has any single character
to distinguish it from all other reptiles. The non-perforate tem-
poral roof is shared between the Cotylosauria and Chelonia;
the single temporal vacuity is common to both Casea or Edapho-

sake, one is as good as another. Huene, Case and I call the small bone on the
posterior angle of the skull in the Captorhinidae and Procolophonia the tabulare
(‘epiotic’); others call it the supratemporal, and no one knows which it is. I
at one time suggested the possible identity of the posterior arcade bone in the
lizards with the tabulare, but no one seems to approve of it; and yet the small
bone in the same position, and with like arrangements in the Procolophonia, Huene
accepts as the tabulare. How little we really know of the homologies is evident
enough in the fact that so acute an observer as Huene admits the possibility of
the persistence of either supratemporal or tabulare. For the sake of uniformity
solely, without committing myself to any theory of homology, I accept and use
the terms supratemporal for the upper, squamosal for the lower element, wherever
they occur, as in the Ichthyosauria, Lacertilia, Stegocephalia, Cotylosauria,
etc.

PRIMITIVE REPTILES 645

saurus and the Therapsida; the double vacuities to Ophiacodon
and the Rhynchocephalia, et cetera.

The characters found constantly in all members of either order
are given in italics.

COTYLOSAURIA

Temporal region imperforate. Skull rugose or smooth; rarely
with spiny protuberances. Postparietals and quadratojugals al-
ways present. Lacrimals always entering into the posterior border
of the nares. Teeth probably always on vomers (prevomers),
rarely also on splenials; either thecodont or acrodont. Inter-
temporal present or absent, as also either the supratemporal
or tabulare, possibly both; primarily with all temporal bones
of the Stegocephalia. Quadrate partly uncovered with a supra-
tapedial process, or wholly covered. Posterior coracoid some-
times unossified. Neck short. Neural arches of presacral verte-
brae stout; hyposphenes sometimes present; spines rudimentary
or of moderate length; one or two sacral vertebrae; tail moder-
ately short or long. Ribs single- or double-headed; abdominal
ribs present or absent. Cleithrum well-developed, vestigial, or
absent. No puboischiadic vacuity in pelvis. Ungual phalanges
often obtuse, sometimes claw-like. Legs never elongate; ectepi-
condylar foramen never present. Rarely with supracostal dermal
plates (Diadectes). From twenty-three to twenty-six presacral
vertebrae.

THEROMORPHA

One or two temporal vacuities on each side. Skull rugose
(Casea) or smooth, rarely with spinous protuberances. Post-
parietals rarely (Edaphosaurus?) or never present, the tabulare

‘ The Procolophonia have in no sense a perforated temporal region. I suggested
several years ago (Biol. Bull., 1904, p. 166) that the uncovering of the temporal
region in Procolophon was due solely to the extension backward of the orbit, and
not to the fusion of supratemporal fossa and orbit. This view has been adopted
by Huene (Paleontographica, 1912, p. 101) who would, however, give a distinctive
term, pseudostegocrotaphic, to the condition found in this genus and its allies.
But if this term be used with the Procolophonia, one must consistently use both
terms, stegocrotaphic and pseudostegocrotaphic, for the conditions found in
the living Chelonia, in Chelone and Chelydra, for instance.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

646 S. W. WILLISTON

never (?). Lacrimals usually not entering nares. Teeth thecodont
(or protothecodont?), rarely attached to splenials or vomers
(prevomers). Neck of greater length. Neural arches never stout.
Ribs single- or double-headed, the cervical ribs sometimes di-
lated distally. Vertebral spines rudimentary, of moderate length
or greatly elongated. Abdominal ribs present or absent. Two
or three sacral vertebrae; tail short or of considerable length.
Posterior coracoid bone sometimes unossified. Cleithrum well
developed, vestigial or absent. Usually, but not always, a pubo-
ischiadic vacuity in pelvis. Ungual phalanges rarely obtuse,
usually claw-like. Legs usually less stout, sometimes very much
elongate; ectepicondylar foramen rarely present. No dermal
ossifications. Presacral vertebrae from twenty-four to at least
twenty-seven in number. |
The only characters, it is seen, by which an unknown genus
may be referred with assurance to one or the other of these orders
are found in the perforation or non-perforation of the temporal
roof, and the neural arches of the presacral vertebrae. The stout
neural arches are, however, not peculiar to the Cotylosauria
since they occur in the Proganosauria.’ Nevertheless, the Thero-
morpha are, as a whole, a more advanced group of reptiles than
the Cotylosauria, as observed in the longer neck; longer, more
ambulatory and prehensile legs, more slender bones, and espe-
cially in the constant reduction of the cranial bones. No thero-
morph has all the cranial elements of Seymouria or Pantylus,
and it is probable that none has all the elements found in Cap-
torhinus or Labidosaurus even, perhaps the most advanced types
of American Cotylosauria. And it is precisely in the skull struc-
ture that we may hope eventually to reach the most accurate
comprehension of the interrelations of the two groups.
Ordinarily the differences known to exist between the different
groups of the Cotylosauria—Diadectosauria, Pantylosauria,
Pareiasauria and Procolophonia—might be considered of ordinal
value if occurring among living reptiles. The Captorhinidae
or Procolophonidae, for instance, have in the posterior cranial
and temporal region the postparietal, tabulare, squamosal and

PRIMITIVE REPTILES 647

quadratojugal; they have no cleithrum; ventral ribs are present;
and the thoracic ribs are single-headed.

The Seymouriidae, on the other hand, have all the cranial
bones of the Stegocephalia, according to Broili and Williston,
namely, the postparietal, intertemporal, supratemporal, tabulare
sqamosal and quadratojugal. They have no cleithrum nor
ventral ribs; and the thoracic ribs are distinctly double-headed.

The Pantylidae, known only from the skull, have all the known
stegocephalian roof elements of theskull, except the intertemporal;
the teeth are acrodont, and are, for the most part, attached to
broad palatal and splenial plates.

The other families of American Cotylosauria all possess a
cleithrum, no ventral ribs, and single-headed thoracic ribs.

While these differences may seem to be important, it 1s very
much of a question whether in such primitive reptiles they mean
what they would in later and less plastic types. ‘They do mean,
as has been said by Case, that, even as early as the close of the
carboniferous times, the most primitive reptiles known to us
had already undergone many changes and divergences.

PROTEROSAURIA

Paleohatteria. The relationships of Paleohatteria with the
American Permocarboniferous Theromorpha are, I believe, in-
contestable. Already such relationships have been suggested
by Broili and Case, as well as others. Osborn, under the belief
that Dimetrodon and the true pelycosaurs possess two temporal
vacuities on each side, placed them in his superorder Diapto-
sauria; and, for a long while they were classed under the Rhyn-
chocephalia in our text-books.

For many years, ever since its original description indeed,
Paleohatteria has been accepted as either directly ancestral to
or a member of the Rhynchocephalia. Baur united it with his
Proganosauria, which he considered ancestral to all later rep-
tiles. Nor can these relationships with the Rhynchocephalia,
as first emphasized by Credner, be denied. In my contestation
that the Proterosauria are really Theromorpha I am merely

648 S. W. WILLISTON

going back to the first position held by Cope, as well as that
held by Baur and Osborn. It may truthfully be said that it
would not require a very great extension of the list of inconstant
characters given on the preceding pages to include most of Os-
born’s Diaptosauria. I draw the greater dividing line elsewhere,
that is, between the Proterosauria and Rhynchocephalia rather
than between the Pelycosauria and Cotylosauria or Therapsida.
I am proposing no new scheme of classification in the present
paper; I am only endeavoring to show the futility of some of
the present ones. Perhaps, when we know a great deal more
about the older reptiles, we may be able to recognize a true phy-
letic classification, but I do insist that we have not that requisite
knowledge at the present time, that the most we can do is to
venture guesses, shrewd guesses perhaps, but still guesses, and
that it is better to endure our present ills than to fly to others
we know not of.

A minute comparison of the characters of Paleohatteria with
those given under the preceding lists of the American Theromorpha
would require too much space here, and is unnecessary. Suffice
it to say that the diagnosis given by Credner (op. cit., p. 545)
applies almost word for word, to the American forms with the
exception of sclerotic plates in the orbits, which can be interpreted
only as an aquatic character. Not only do the known characters
agree throughout with those of the Theromorpha as a whole,
but almost all of them will apply to the family Poliosauridae,
and Ophiacodon and Varanosaurus especially; the first two
paragraphs of Credner’s diagnosis will, indeed, apply verbatim
to Ophiacodon; and, so far as the characters of Proterosaurus
from the upper Permian are known, there is little or nothing
discrepant; but I omit their discussion here.

While it is very possible that Paleohatteria has two temporal
arches on each side, the presence of the upper fenestra has never
been positively proven. Credner in his first description merely
says that the presence of the supratemporal fenestra is ‘wahr-
scheinlich.’ But its presence or absence is really unimportant
in consideration of the fact that the two vacuities, both upper
and lateral, are known to occur in at least one genus of American

PRIMITIVE REPTILES 649

Theromorpha; and the relations of some of the American genera,
with and without the second fenestra, are so intimate that their
family separation is very doubtful.

My conclusion then is that Paleohatteria and the Proterosauria
can not be distinguished from the Theromorpha by ordinal char-
acters, and should be united with them. I would provisionally
classify the order as follows:

Order Theromorpha

Suborder Pelycosauria
Family Clepsydropsidae (Sphenacodontidae)
Family Poliosauridae
Family Edaphosauridae

Suborder Proterosauria
Family Paleohatteridae

Suborder Caseasauria
Family Caseidae

Suborder
Family Kadaliosauridae (Araeoscelidae)

I am aware that this classification is what might be called
‘horizontal,’ rather than ‘vertical;’ but I also insist that horizon-
tal classifications are absolutely imperative until such time as
we have more than vague surmises and guesses as to the true
lines of phylogeny. Smith Woodard has pertinently criticised
the prevalent fashion of making phylogenies to suit every pass-
ing hypothesis. And the fashion has found its climax in the
‘phylogenies’ of Steinmann, which I can only understand as
proposed in an exquisite spirit of irony.

PROGANOSAURIA

The order Proganosauria was proposed by Baur in 18777 to
include Stereosternum and Mesosaurus only, founded chiefly
upon the five distalia of the tarsus, a character unknown in mod-
ern forms, but one which we have seen is common to all known
Permocarboniferous reptiles. Later’ he added Paleohatteria,

7On the phylogenetic arrangement of the Sauropsida. Jour. Morph., vol. 1,

p. 103, 1887.
8 Amer. Jour. Sci., vol. 36, p. 311, 1889. -

650 S. W. WILLISTON

to the order, dividing it into the two families Mesosauridae
and Paleohatteridae, which he proposed, with the following
definition: ‘‘Humerus with entepicondylar foramen; five distinet
tarsal bones in second row, one for each metatarsal; pubis and
ischium broad plates; each set of abdominal ossicles consisting
of numerous pieces; condyles of limbs not ossified;” all of which
will be found among the characters given for the American Thero-
morpha on the preceding pages. It was doubtless under this
wider conception of the group, that Broom added to the order,
hesitatingly, the genera Saurosternon, Heleosaurus and Heleo-
philus from the middle or upper Permian of Africa. ‘These are
rejected by Huene,® and I agree with him. In 1892 Seeley pro-
posed the ordinal name Mesosauria toinclude not only Mesosaurus
and Stereosternum, but also some of the Nothosauria.!? Osborn
in 1904," restricted the term to its original limits and his example
was followed by McGregor.

It is very clear that the name Mesosauria, so often used for
the group, has no legitimate standing. Seeley and not a few
others have believed that the Proganosauria are primitive Sauro-
pterygians. Osborn and McGregor, on the other hand, would
locate the group among the Diapsida and Diaptosauria, under
the assumption that there are two temporal arches on each side,
an assumption that has recently been denied by Huene (1. c.), who
believes that the group is an independent offshoot from the
primitive Cotylosauria, and more or less related to the Ichthyo-
sauria. He says definitely that Mesosaurus has the supratem-
poral fenestra only, and that the lower arch is lost; as Jaekel
declares is also the case with the Sauropterygia. There may
have been and doubtless were, different ways in which the tem-
poral fenestrations arose among Reptilia, but, with all due re-
spect to the expressed opinions, I am not yet convinced that the
explanations explain. I can conceive how a ‘lateral’ vacuity
may have been converted into an ‘upper’ one or vice versa, or
how the two vacuities may have fused into one, and I am not

* Paleontographica, vol. 59, p. 100, 1912.
10 Quart. Jour. Geol. Soc., 58, p. 586, 1892.
11 Mem. Amer. Mus. Nat. Hist., 1, p. 481.

PRIMITIVE REPTILES 651

yet ready to accept phylogenies and systems of classifications
based upon such vague distinctions. Neither am I convinced
that the quadrates of the Ichthyosauria and the Proganosauria
are of such a distinct type as to require the phylogenetic union
of the orders.

Case has expressed the opinion that ‘‘Mesosaurus is probably
an aquatic adaptation of the Cotylosaurian type’? and I have
also, independently, urged the relationships of the Proganosauria
with the Theromorpha under the assumption of a single temporal
vacuity.!? Nevertheless, the Proganosauria, in their aquatic
adaptations are a definite group of primitive reptiles, of a distinct
type, as shown in their elongated maxillae, posterior nares,
elongated neck, long and flattened tails as figured by McGregor,
short, fan-like scapula, elongated propodials and shortened epl-
podials, et cetera. The vertebrae appear to be almost typically
cotylosaurian in their thickened arches and low spines. The ribs
are described as single-headed, attached below the diapophyses,
but I have given reasons for a different interpretation elsewhere.

Aside from the distinctive aquatic adaptations, Mesosaurus
and Stereosternum show primitive cotylosaurian or theromor1ph
characters in the presence of teeth on the vomers (prevomers)
deeply biconcave vertebrae, pectoral and pelvic girdles, entepi-
condylar foramen and ectepicondylar groove, carpus, tarsus,
digits, et cetera. The aquatic adaptations of the limbs scarcely
differ more from the terrestrial theromorph structure than do those
of the dolichosaurs from the true lacertilians, and the aquatic
modifications of the skull are of but little greater moment than
those of the mosasaurs.

However, I am willing to consider the Proganosauria as a
distinet order of reptiles until more shall be known about them.

REPTILES OF THE LOWER PERMIAN OF EUROPE

As an appendix to the foregoing discussion of the better known
lower Permian and upper carboniferous reptiles, the following

22 Revision of the Cotylosauria, p. 118, 1911.
13 American Permian vertebrates, p. 111, 1911.

652 S. W. WILLISTON

brief notice of other less well known European genera will be
necessary.

Phanerosaurus naumanit H. V. Meyer. The type and known
remains of this genus and species consist of a series of four dor-
sal and a sacral vertebrae. They can not be distinguished
from corresponding ones of Diadectes, and are doubtless truly
cotylosaurian.

Stephanospondylus Stappenbeck. This genus was based upon
the species pugnax Geinitz and Deichmuller, referred by its
authors to Phanerosaurus. So far as the skull and vertebrae are
concerned they seem to be genuinely cotylosaurian, but the
pectoral and pelvic girdles, if correctly recognized and figured by
Stappenbeck, are very aberrant, not only from the Cotylosauria
but from all early reptiles. The clavicular girdle is unlike any-
thing known among American reptiles, and seems more like that
of a temnospondyl. The coracoid and pubis also are unlike
anything known among early (or late) reptiles. I suspect that
the so-called coracoid is in reality the pubis, and that the so-
called pubis is something else, possibly a sacral rib. If it be
really a pubis it indicates a pelvis of a radically distinct type,
one with a large puboischiadic vacuity, as in modern reptiles.

Stereorhachis dominans Gaudry. The relationships of this
genus with the American Pelycosauria in the narrow sense have
long been recognized. Thevenin describes and figures the
typical specimens and refers them to the Pelycosauria. None
of its characters seems discrepant, so far as they are known.
Stereorhachis comes from the lower part of the Autunian of
France, possibly of contemporaneous age with the Clear Fork
beds of Texas. It is associated stratigraphically with Actinodon,
a genus closely allied to Eryops.

Callibrachion Gaudryi Boule. This genus was referred by
Case;“ to the Pelycosauri notwithstanding the reputed opistho-
coely of the cervical vertebrae, and the high coronoid of the
mandible. Huene has since showed" that both these characters
were erroneously ascribed to the genus, though he objects to its

14 Revision of the Pelycosauria, 1907.
15 Centralbl. fiir Mineralogie, p. 534, 1905.

PRIMITIVE REPTILES 653

location among the Pelycosauria, rather than the Proterosauria.
A detailed review of its characters is unnecessary here; the reader
will find them in the cited work of Case or in the original paper.
So far as I can discover, the genus presents no aberrant charac-
ters to distinguish it from the Theromorpha. The temporal
region of the skull is unknown; both upper and lateral vacuities
have been assigned to it under its assumed relations to Paleo-
hatteria and Sphenodon. Possibly Baur’s remarks may be
applicable here, as so often elsewhere:!* ‘‘Dieser Fall zeigt sehr
deutlich, wie leicht man sich taiischen lassen kann, wenn man
durch eine allgemeine giiltige Anschauungsweise beeinflusst
wird.” I doubt very much whether Callibrachion had both
upper and lateral temporal fenéstrae.

The remains of this genus, together with those of Haptodus
and Sauravus cambryi come from the Gargenne schists of the
upper part of the Autunian, whose position in the lower Permian
I supposed was unquestioned. Huene, however, refers both
Callibrachion and Haptodus, as also Aphelosaurus, to the ‘oberen
Perm,’ but gives no reason therefor.

Aphelosaurus lutevensis Gervais. This genus and species are
known from a single specimen from the lower Permian of France.
The specimen lacks the head, neck and tail. No intercentra
have been detected, though they are doubtless present. Slender
abdominal ribs are present. The observed phalangeal formula of
the hand is 2, 3,4, ?,3. So far as the published characters indi-
cate, there is nothing in the genus to distinguish it from the
American Theromorpha.

Haptodus baylei Gaudry. 'Thevenin, from a study of the known
material, reaches the conclusion that the genus is identical with
Paleohatteria Credner published several years later, but is inclined
to think that the species are distinct.!7 Haptodus, like Paleohat-
teria is known only from specimens lacking the ossific ends of the
limb bones, which, taking into consideration the number of speci-
mens studied by Credner, would seem to be sufficient proof that
the character is an adult one. And such is doubtless the condi-

16 Zool. Anzeiger, Bd. 12, p. 239, 1889.
17 Annales de paleontologie,

654 S. W. WILLISTON

tion in the American genus Clepsydrops. These two specimens
of Haptodus add nothing to our knowledge of Paleohatteria, if
they be the same, except that one is a half larger than those
studied by Credner.

Kadaliosaurus priscus Credner. This genus, so far as I am
aware, is represented by a single known specimen, from the mid-
dle Rothliegende of Germany, described and figured by Credner.
The specimen, unfortunately, lacks the skull, the pectoral girdle
and most of the tail, and has only stray bones of the feet. So
far as it goes, as described and figured by Credner, it has a
remarkable resemblance to Araeoscelis Williston, from the Clear
Fork beds of Texas. It differs apparently in the possession of
a well defined ventral armor, of which there is little or no evi-
dence in Araeoscelis; perhaps also in having a short tail, though
I should think it more probable that the tail is long, as in Arae-
oseelis. The vertebral spines in both are very short, the centra
are notochordal, and the body is very slender. The resemblance
of the limb bones in the two genera is astonishing; so great
indeed that Credner’s figures would serve equally well for Arae-
oseelis. Not only are the front and hind limbs of equal length,
but the epipodials are as long as the propodials, and all are very
slender. The femora of both forms are identical in their slen-
derness and sigmoid curvature. What is still more remarkable,
the humeri, apparently agreeing quite in shape, have both entepi-
condylar and ectepicondylar foramina, a character unknown in
any other reptile of lower Permian age. There is no puboischiadie
vacuity in the pelvis of either genus. It would seem impossible
that such extraordinary resemblances should be solely the result
of convergent evolution. I believe that they are related and
that both should be placed, for the present at least, in the same
family. The family differs so much from others of the Permo-
carboniferous that its relationships may well remain a matter of
doubt. Araeoscelis has a single temporal vacuity of large size,
and was a very slender, and slender legged, long tailed lizard-lhke
reptile, almost certainly of climbing habits; its toes have acumi-
nate claws.

PRIMITIVE REPTILES 655

Datheosaurus macrourus Schroeder. This genus and species
from the lower Rothliegende of Germany, is unfortunately known
only from a single specimen, not in the best preservation. It is
a very slender reptile with a very long tail, with no known char-
acters at variance with the preceding lists. Doubtless the
interclavicle has an elongated stem and the humerus an entepi-
econdylar foramen. Whether the temporal region is perforated or
not is not known. In other characters, especially the structure
of the vertebrae, it is sharply excluded from the Cotylosauria.
Its nearest relationship seems to be with Kadaliosaurus, Arae-
oscelis and Paleohatteria.

MICROSAURIA

The name Microsauria was first proposed by Dawson in 1863!8
based upon Hylonomus, a genus described by him from the
carboniferous of Nova Scotia. Nearly thirty years later, Credner
described! more fully and in detail another species, H. geinitzii,
from the middle Rothliegende of Germany, which he referred
to the same genus, together with an allied new genus and species,
Petrobates truncatus from the same horizon. In 1897, Baur?°
expressed the opinion that both forms were, in reality reptiles as
Credner had suspected:

Die Frage ob Hylonomus und namentlich ob Petrobates den Stego-
cephalen oder aber den Rhynchocephalen zuzurechnen seien, lisst sich
nicht durch ein kurzes Wort entscheiden. . . . . Wenn man bel
Petrobates vom Schidel absieht, welche nicht genau genug bekannt ist,
so kénnte man diesen Vierfiissler fiir einen kleinen Rhynchocephalen
aus der Familie der Proterosaurien halten, wenn dem nicht das Vor-
handensein von nur einen Sacral-wirbel entgegenstiinde.”!

I scarcely doubt but that had Credner known of stegocrotaphic
reptiles with a single sacral vertebra, such as Diadectoides and

18 “My first impression is that they constitute a separate family or order, to
which I would give the name Microsauria, and which may be regarded as allied,
on the one hand, to certain of the humbler lizards, as the Gecko or Agama, and
on the other to the tailed Batrachians.’”’ Air breathers of the coal period, p. 47.

19 Zeitschrift der deutschen geologischen Gesellschaft, Bd. 42, p. 241, 1890.

20 Anat. Anz., Bd. 14,

21 Credner, op. cit., p. 257.

656 S. W. WILLISTON

Seymouria, he would have accepted both Hylonomus and Petro-
bates as reptiles. However, Baur’s more decided views as the
reptilian nature of the Microsauria were accepted by not a few
authors, and their reptilian affinities, or at least those of some
of the forms called Microsauria, have been recognized by all.
Indeed, by some writers the Microsauria are considered the an-
cestral type from which some if not all reptiles arose. Never-
theless, they have been generally retained among the Stego-
cephalia, either under the ordinal or subordinal term Lepospon-
dyli, proposed by Zittel, or the earlier Microsauria of Dawson.
That the group Microsauria, as usually defined, is a heterogeneous
one is now admitted, but sufficient evidence to serve as a secure
foundation for its disintegration and the redistribution of its
various forms on a scientific basis has not been forthcoming;
and I doubt whether any proposed scheme will stand the test
of time, or be worth serious consideration. The very recent
scheme of classification proposed by Jaekel, in which he gives
to the group a class value, dividing it into various orders, betrays
such a regrettable lack of knowledge that it may be disregarded.

A few years ago M. Thevenin described a peculiar reptile-like
form from the Stephanian or uppermost carboniferous of France
as Sauravus costei, referring it to the Rhynchocephalia.22 In
the following year I redescribed” an allied form which previously
had been described and figured by Cope*4 under the name Iso-
dectes punctulatus, from the Linton beds of Ohio, of middle
Pennsylvanian age, referring it, together with Sauravus costei
provisionally to the Cotylosauria. As a comparison of the
specimen upon which Cope’s remarks were based with the real
type of the species showed specific differences at least, I gave
to his plesiotype the name Isodectes copei, later changed to
Eosauravus copei2 when it became evident that the species
could not be referred to the Texas genus Isodectes, of much

22 Annales de paleontologie, 3, p. 19, 1907.
28 Journal of Geology, 16, p. 395, 1908; Moodie, Proc. U. 8. Nat. Mus., 37, p.

11, 1909.
24 American Naturalist, 1896; Proc. Amer, Phil. Soc., p. 88, 1877.
25 Bull. Geological Soc. Amer., vol. 21, p. 272, 1910; Case, A revision of the
Cotylosauria of North America, p. 31, 1910.

PRIMITIVE REPTILES 657

later age. In the same and a later paper?® I expressed the con-
viction that:

At least two distinct groups have been associated among what are
called Microsauria, and that one of them, with single-headed inter-
central ribs and intercentral chevrons, represented by Hylonomus,
Petrobates, Sauravus and Eosauravus must be dissociated into a group
more nearly allied to, possibly identical with, the Reptilia in a wide
sense, while the other, of which Urocordylus may be taken as a type,
may remain with the Amphibia.

For the former group I accepted the name Microsauria; for the
latter, tentatively, Lepospondyli. In the same year Thevenin,?’
in a further discussion of Sauravus, with the description of an
additional species, evidently reached a like conclusion, referring
the genus to the Microsauria, taking Hylonomus Dawson as
the type of the order, and placing it among the Reptilia. How-
ever, Dr. Moodie informs me that Hylonomus Credner, upon
which most of our knowledgeof the genus is based, is by no means
certainly proven to be identical with Hylonomus Dawson, the
real type of the Microsauria; that indeed the latter may not
belong in the same group with the former, a conclusion not
surprising, considering the different geological ages of the two.
If such be the case, and it will not be possible to answer the ques-
tion satisfactorily until more material referable with certainty
to Hylonomus lyelli Dawson has been attentively studied, the
terms, as I use them, are merely tentative.

Unfortunately the differential classification of the two groups
is still chiefly dependent upon the skeletal structure aside from
the characters of the skull, and, until the latter is better known
than it is at present among the Microsauria (in the above sense)
we lack a very important means of discrimination. For the
present therefore I would distinguish the two groups, whether
the Microsauria be reptiles or not, as follows (for the present
I leave out of account the Diplocaulidae and Aistopodidae):

Lepospondyli (as typified by Urocordylus). Vertebrae holo-
spondylous, notochordal; ribs elongate, the head not attached

26 Journal of Geology, vol. 18, p. 599, Oct., 1910.
27 Annales de paleontologie, 5, p. 43, 1910.

658 S. W. WILLISTON

intercentrally; chevrons exogenous processes from body of ver-
tebra; no intercentra; a single sacral vertebra; carpus and tarsus
unossified; or, if ossified, a separate intermedium tarsi present;
skull with two oceipital condyles; parasphenoid well developed.

Microsauria (as typified by Eosauravus, Sauravus, Hylonomus
Credner and Petrobates). Vertebrae holospondylous, notochor-
dal; ribs long, attached intercentrally, usually without distinct
separation into head and tubercle; chevrons intercentral, not ex-
ogenous processes from the centrum. One (?) or two sacral verte-
brae; carpus and tarsus ossified; no separate intermedium pedis;
phalangeal formula 2, 3, 4, 5, 4 in foot (Kosauravus, Sauravus).
Interclavicle with long posterior stem. Skull Gmperfectly known
in Hylonomus Credner and Petrobates Credner) evidently stego-
crotaphic; palate with small teeth; parasphenoid slender (in front
at least). Small animals.

Intercentra and dorsal ossifications have been discovered in
none of these genera while ventral ribs are known in all save
the single known specimen of Eosauravus copel, in which they
may have been lost in fossilization. Ossified mesopodials are
known in all the genera, and apparently all have but the two
bones, astragalus and caleaneum in the proximal row of the
tarsus. Hylonomus Credner and Petrobates are from the middle
Rothliegende of Germany, Sauravus Thevenin typically is from
the uppermost carboniferous of France, and EKosauravus Williston
is from about the middle of the Pennsylvanian of Ohio.

Taking these genera into consideration as a whole, it is seen
that there is nothing yet known to differentiate them from the
other primitive genera included under the definitions of the fore-
going pages (unless it be the absence of intercentra); while, on
the other hand, there are a number of characters, and those of con-
siderable importance, that distinguish them from the Amphibia.
It was some of these characters which induced Moodie to give
an. ordinal position to the Diplocaulidae, in which he was quite
right, under the assumption that the Microsauria as a whole are
a homogeneous group. But it is evident that, as usually grouped
the Microsauria are not homogeneous; that some of the genera
hitherto included among them are genuine amphibians, in all

PRIMITIVE REPTILES 659

probability allied to Diplocaulus, while others must be removed
from the order.

But I care not whichever name is applied to the group I here
call Microsauria, since in all probability it will be repeatedly
changed before a fairly definite conclusion is reached; and that
conclusion will not be reached until we know vastly more about
the carboniferous vertebrate fauna than we do at the present
time.

If the Microsauria, as typified by the above mentioned genera,
are true, or real reptiles, what position in the system should
they occupy? The vertebrae are purely holospondylous, and
it is the universal belief that rhachitomous vertebrae are primi-
tive; to assume the reverse is to rewrite the morphology of the
vertebrae, and especially of the atlas of the Amniota; nor has
there been a shadow of an argument so far produced to disprove
the theory. The cleithrum has never been observed in any
microsaur, and its presence in the Cotylosauria can only be
explained as a direct inheritance from the Amphibia. Since
neither the intercentra nor cleithrum could have reappeared
after their one-time loss the probability that any known micro-
saur could have been ancestrally related to the later reptiles is
very slight. If these microsaurs are real reptiles or even indi-
rectly ancestral to any modern reptiles, we must of course seek in
early rocks their ancestral Amphibia. But we have so far no
evidence whatever that Eosauravus and Sauravus may not
actually belong among the Cotylosauria, where some autbors
would place them. It is true the vertebral arches of Eosauravus
and Sauravus do not seem to partake of the peculiar structure
of the true Cotylosauria, but, inasmuch as the postsacral verte-
brae of Seymouria are of the ordinary slender type, the character,
though a useful one, cannot be profound.

I believe that the cited genera at least are real. reptiles—there
is no known reason for including them among the Amphibia—
but, until more is known about them than at present, a definite
position in the system can not be given them. One may con-
tinue to call them, following Baur and Thevenin, true Microsauria
as distinguished from the Lepospondyli; or he may, after Broili

660 S. W. WILLISTON

and Case, locate them among the Cotylosauria, and still use
the name Microsauria in its usually accepted sense. I can not
believe that the time has come for a new system of classification
and new names.

LYSOROPHUS

The history of the genus Lysorophus will be found in the vari-
ous papers of the appended list?® and need not be recounted here.
Suffice it to say that Broili and others urge that it is a reptile;
while Broom and I believe it to be an amphibian closely allied
to, if not a member of the Urodela; Moodie thinks that it is a gym-
nophionian; and Case is confident that it is an amphibian, though
doubtful of its urodelan affinities. The question here at issue is:
Is Lysorophus a reptile? And, if so, to what group of the known
reptiles is it most nearly allied? Lysorophus is known to occur
in the Permocarboniferous of Illinois, Oklahoma and Texas.
My knowledge of the genus is based upon abundant material,
coming chiefly from the Clear Fork beds of northern Texas.
I would characterize the genus and group as follows:

Lysorophus. Slender, Amphiuma-like in form; of from 8
or 10 inches to possibly 18 inches in length; with small limbs.
No pineal foramen; no lacrimal, postfrontal, postorbital, jugal
or quadratojugal bones; an unpaired supraoccipital bone, on
either side of which are two distinct bones, variously identifiable
as the squamosal and tabulare, or supratemporal; prefrontal
(lacrimal?) articulating with nasal, frontal and parietal; orbits
confluent with the open temporal region; quadrate ossified,
suturally united with squamosal, projecting downward and
forward; no temporal bars; palate with large subcranial plate,
probably the parasphenoid, or parasphenoid and _basioccipital;
vomer with rows of teeth; two occipital condyles, their articu-
lar faces directed inward and backward; one pair of ossified

28 Cope, Proc. Amer. Phil. Soc., p. 187, 1877; Trans. Amer. Phil. Soc., 16, p. 287, :
1886. Case, Jour. Geol. 8, p. 714, pl. II, ff. 12a, b, ¢; Bull. Amer. Museum Nat.
Hist., 24, 531; Revision of the Amphibia of the Permianof North America, 68,
141, 1911. Broili, Paleontographica, 51, p. 94, 1904; Anat. Anzeiger, 25, p. 586,

1904; ibid., 33, p. 291, 1908; Zittel, Grundziige der Paleontologie, 2d ed., p. 218,
1911; Williston, Biological Bulletin, 15, 229, 1908; Jour. Geology, p. 600, 1910.

PRIMITIVE REPTILES 661

ceratobranchials, and four pairs of ossified epibranchials. Atlas
short, cylindrical, concave behind, its body composed of a single
bone, that is not divided into odontoid and intercentrum, sup-
porting the paired neural arch; closely united to skull without
intervening cartilage. Vertebrae notochordal, the neural arches
each composed of paired bones meeting in the middle line above,
but not united, and without spines; with a prominent diapophy-
sis on each side of the arch for the support of the (for the most
part single-headed) ribs; no intercentra and no free chevrons.
Limb bones small, without ossified articular ends, the meso-
podials evidently unossified, and the girdles not yet recognized.
No dermal ossifications.

If the foregoing characters be correct, and I feel quite sure
they are, it will necessarily follow that, if Lysorophus be a real
reptile, it must occupy a place all by itself as a separate subclass,
without known descendants or antecedents. It is an established
principle in paleontology that an organ or bone once lost never
reappears in any descendants. We know of no reptile in which
the lacrimal, postfrontal, postorbital, jugal and quadratojugal
are all wanting; we know of but one order, the Sauropterygia,
in which single-headed ribs have become purely neuracentral
in attachment, and surely no one will include Lysorophus in that
order! Nor could it be genetically allied to the Lacertilia, because
the Squamata have single-headed ribs attached to the centrum
only, a more primitive reptilian condition to which the single
headed neuracentral ribs of Lysorophus could not have returned.
And all Lacertilia have bones in the skull which are wanting in
Lysorophus; in none does the prefrontal articulate with the
parietal, excluding the postfrontal. That Lysorophus is widely
remote from all other contemporary reptiles may be seen from
the foregoing pages. The entire loss of the intercentra, not
only from the cervical, dorsal and caudal vertebrae, but also
from the atlas, is a character unknown in any Amniote verte-
brate, but is the universal condition among amphibians. And
so also, is the presence of four (not three) pairs of epibranchials.
If, therefore Lysorophus be really a reptile, it had become so
extremely divergent from all other known reptiles by the close

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

662 S. W. WILLISTON

of carboniferous times that it must represent a group in itself
equivalent to all others then living; and, inasmuch as it could
not have left any known descendants, its claim to subclass rank
is Inoppugnable.

It has been claimed that the skull of Lysorophus is really mono-
condylar, that the condyle is in reality a tripartite one, in which
the middle or basioccipital part was represented in life more
or less by cartilage, or that the atlas articulated in fact with
all three bones—the basioccipital and exoccipitals. I can not
accept this interpretation of the structure. One of the skulls
in the Chicago collection has the ‘atlas’ in close articulation
with the condyles, with no interval to correspond to a cartilagi-
nous continuation of the middle part. The atlas is transversely
truncate in front in the middle, its angles articulating obliquely
with the condyles. A second specimen, in which the atlas has
been cleanly dislodged, has the condyles quite as they are figured
in the cited paper of Case. I will admit that there are some
features here which indicate an ossification of the basioccipital
and its articulation in the middle with the front end of the atlas;
but, under any interpretation of the structure, the mode of artic-
ulation with the atlas is very different from that known in any
reptile. While the single condyle in some Cotylosauria may
be flat or concave, in life it was rounded out by cartilage. In
no reptile is the condylar articulation known to be concave,
as is positively the case in Lysorophus, if it really articulate
in the middle with the atlas.

The general resemblance of Lysorophus to Amphisbaena is
conceded, as is also its resemblance to the Gymnophiona, and
these resemblances have deceived both Broili and Moodie as
to the real relationships of the genus. Both the Amphisbaenia
and the Gymnophiona, as also Lysorophus, are or were burrowing
creatures, either limbless or with small limbs, with a more or
less pointed head, open temporal region, small eyes, and an elon-
gated, ofttimes worm-like body. But these resemblances do not
necessarily indicate genetic relationships of any of them, any
more than do the resemblances between a porpoise and an ichthyo-
saur indicate genetic relationships. It is a curious fact that

PRIMITIVE REPTILES 663
the loss of the lower temporal arcade, and the open condition
of the temporal region, as in these forms, the snakes, lizards
and salamanders, seems to be correlated with a head resting more
or less prone upon the ground.

In my opinion Lysorophus has no direct ancestral relation-
ships with any modern vertebrates; that the Urodela, or even
the Gymnophiona, began in such extremely Amphiuma-like
forms in the carboniferous would be contrary to all our paleon-
tological experience. The whole structure of Lysorophus dem-
- onstrates its aquatic habits, in the chiefly perichondral ossifica-
tion of its bones; in the widely separated sutures of the skull,
as urged by Miss Finney; in the large and doubtless perennially
functional branchiae; and in the mode of its occurrence, in beds
of immense numbers, coiled up and in undisturbed positions.

The ribs of Lysorophus are usually described as single-headed,
attached to a prominent diapophysis; and this is true for nearly
all of them, but apparently not for quite all. In what appear
to be anterior ribs, those probably attached near to the head,
there is a distinction into head and tubercle, separated by a
short neck. Just how the capitular end is articulated has not
been determined, though it appears to be attached to a facet
immediately below the diapophysis; it almost certainly does not
articulate intercentrally, nor has any centrum been found show-
ing a parapophysial facet, such as characterizes the contempor-
ary Crossotelos.

The occurrence of limb bones in Lysorophus seems assured.
To determine if possible whether the first specimens of limbs
found associated with skeletons of Lysorophus were accidental
or not, I some time ago requested Miss Marian Finney, a student
of paleontology in the University of Chicago, to make a careful
search of the abundant material preserved in the collections.
I give the results of her studies, which I can corroborate, in her
own words, as follows:

664 S. W. WILLISTON

THE LIMBS OF LYSOROPHUS
BY MARIAN FINNEY

In 1908 Dr. Williston discovered in’ a nodule containing a
series of Lysorophus vertebrae and ribs limb bones which he
believed belonged with them, but he was not sure that they might
not have been accidental intrusions from some other small, unde-
tected amphibian. To determine, if possible, whether Lysoro-
phus really is a limbless amphibian or not, Dr. Williston re-
quested me to study carefully the material of the collections of the
University of Chicago coming from Texas. As a result of this
study I am able to say with assurance that the evidence proves,
I thing conclusively, that Lysorophus does possess limbs, though
it has been impossible to determine whether the limbs are the
front or the hind, or whether indeed both front and hind.

The limbs in life were evidently loosely attached to the ver-
tebral column, and were easily separated and lost; and the bones
themselves are so minute that, once separated from the skeleton
they would hardly be detected, or if detected correctly recog-
nized. Of the two hundred nodules examined, containing more
or less of the skeleton of Lysorophus I find fifteen containing limb
bones, the most of them isolated; only a few with more than
one bone associated.

The best of these specimens consists of a femur, tibia, fibula,
two metatarsals and two phalanges all related. I assume that
the limb is the hind one because of the shape of the bones only.
The femur is 10.5 mm. in length, hollow throughout, with a
rather slender shaft, and moderately expanded ends; its ends
are mere shells of thin bone, easily crushed. The tibia and fibula
are 6.5 mm. in length, or about three-fifths the length of the
femur and, like it, are hollow. The tibia is thickset, the ends
less expanded than those of the femur. The fibula is slender,
and lies close beside the tibia. Beyond these epipodials are
two bones which doubtless are metatarsals, though not much
shorter than the tibia and fibula; they measure 5 mm. in length.
Beyond these are two small bones which seem to be phalanges,
lying one at the end of the other, each about 1 mm: in length.

PRIMITIVE REPTILES 665

There is no trace of mesopodials, and, in all probability, judging
from the imperfect ossification of the leg bones, the tarsus was
unossified. The limb lies in the matrix beside a long coiled
series of vertebrae, but I can discover no trace of girdle bones.

Another specimen is not so complete and is somewhat larger.
The femur (?) is about one-third longer, and is more massive,
though hollow like the others, and with truncate ends. At or
near the extremity of this bone lie four others, each about half
the length of the femur, which probably are metapodials. This
leg is associated in the matrix with several disconnected Lysoro-
phus vertebrae, but with no trace of girdle bones.

Most of the numerous single bones found seem to be propo-
dials, similar in character to the ones I have called femora, and
of about the size of the larger one I have described. As a rule
only isolated bones have been observed, and always in nodules
containing vertebrae and ribs, and with no indications of other
forms than Lysorophus. With one specimen a skull was found,
but not connected with the vertebrae. In one of the nodules
are two bones, apparently epipodials, in articulation with, a
fragment of a larger bone, a propodial probably; they are 6.5 mm.
in length, or quite the size of the first specimen described. The
farger of the two may be the tibia, the smaller the fibula. Asso-
ciated with these two bones are parts of two lower jaws, measur-
ing 7 and 8 mm. in length; in the larger fragment I count eleven
teeth; they are sharply pointed, conical and recurved; and all
these bones lie associated in the matrix with typical Lysorophus
vertebrae only.

In general appearance Lysorophus must have resembled Am-
phiuma means very much. It had a slender body and slender
tail. In one specimen I count eighteen vertebrae in an appar-
ently complete tail; in another I count twenty in a gradually
tapering series. That Lysorophus had legs there seems to be
no longer doubt, but they evidently were not of much functional
importance, since they are relatively very small. The total
length of the body may perhaps have varied from 10 to 15 inches;
a complete leg could not have been more than 1% inch in length.
Evidently Lysorophus lived in pools or ponds of water, burrowing

666 S. W. WILLISTON

perhaps in the soft mud of the bottom. They are found almost
invariably coiled up, much after the manner of the modern Am-
phiuma. That they were burrowing in habit seems very proba-
ble because of the shape and structure of the skull, resembling
much those of the Amphisbaenia or Gymnophiona, but the very
loose sutures of the skull would hardly be adapted for burrowing
in firm earth, after the habits of the lizards and blind worms,
which have closely united skull bones. Not only this skull struc-
ture, but the mode of occurrence in immense numbers, closely
associated together, as reported by Dr. Williston, and closely
coiled, and especially the well ossified branchiae, seem to prove
that, Lysorophus was a perennial water breather.

University of Chicago, May 5, 1912.

THE PAEDOGAMOUS CONJUGATION OF BLEPHAR-
ISMA UNDULANS ST.

GARY N. CALKINS

From the Department of Zoology, Columbia University

TWENTY-FIVE FIGURES

”

The genus Blepharisma was created by Perty in 1849-1852
with the following diagnostic characters: Body flat, lancet-
shaped, pointed posteriorly and ending anteriorly in a short
snout (Schnabel); the deeply incised region extending from
the anterior end to about the middle of the body is provided
with a row of long straight cilia arranged in parallel lines.
‘Molecular rows’ bearing extremely fine cilia, difficult to see,
are borne by the remainder of the body. ‘Two species were
described as new, B. hyalinum, a colorless form, and B. per-
sicinum, a reddish colored form. The latter he regarded as
possibly the same as Miiller’s Trichoda striata.

Ehrenberg (31) in the meantime described a form under
the name of Bursaria lateritia which he identified as the same
as Miiller’s Trichoda ignita, having a distinct reddish color,
compressed body and with a characteristic shape of a garden
knife.

Stein (’67) described the genus in detail, including as
synonyms Miiller’s Trichoda striata, aurantiaca, and ignita,
Bursaria lateritia of Ehrenberg, and Ypsistoma of Bory, and
characterized two species as follows:

1. B. lateritia: «Body peach color, purple-red or tile-red,
occasionally colorless. The peristome reaches as far as the
middle of the body or beyond this point, with a free-standing
bristle-like undulating membrane attached close in front of
the peristome angle. The nucleus is a single oval body in the -
anterior half of the organism.

. 667

668 GARY N. CALKINS

2. B. undulans St. The color is purple red; the peristome
does not reach to the middle of the body but is usually limited
to the anterior third, with a well-developed undulating mem-
brane attached to the entire posterior half of the left peristome
edge. One nucleus in the anterior half and one in the ‘pos-
terior half, with or without a connecting strand of nuclear
material.

Stein maintained that Perty’s two species were neither new
nor distinct from one another, and interpreted Perty’s ‘mole-
cular rows’ as the body stripings. Stein’s B. lateritia was more
frequently seen, undulans only a few times and then in connec-
tion with numbers of lateritia. Bitschli regards Stein’s undu-
lans as the same as Ehrenberg’s Uroleptus musculus, appar-
ently ignoring the difference in the undulating membrane of
Stein’s two species, but accepting Stein’s characteristics as given
for B. undulans. Stokes, finally, has apparently described the
genus under the name Apgaria.

While there is little difficulty in deciding upon the generic
name Blepharisma, the task is not so easy with the type species.
Stein has shown that Perty’s two species cannot hold, while
the descriptions of Stein’s own species, careful and detailed as
they are, cannot be traced back to any one type of Ehrenberg
or Miiller. Both his lateritia and undulans could be included
in Ehrenberg’s Bursaria lateritia which in turn, might go back
to Miiller’s Trichoda aurantiaca and T. ignita. Neither Miller’s
nor Ehrenberg’s descriptions of these species were sufficiently
accurate to be recognizable today; similarly with all other char-
acterizations save those of Stein, where we have to decide as
to the validity of the two species lateritia and undulans.

The essential differences in structure between the two species
given by Stein are in connection with the undulating membrane
and the macronuclei. In lateritia the undulating membrane
is said to be a mere bristle or cirrus-like structure inserted near
the mouth in the angle of the peristome, while in undulans,
the undulating membrane is much more extensive, extending
half way up the peristome left edge.. The bristle or cirrus was,
probably, only the proximal edge of the undulating membrane.

CONJUGATION IN BLEPHARISMA UNDULANS 669

The nuclear differences are less important and have no specific
value as will be shown in the following pages.

The organism with which we are dealing in the present paper,
belongs to Stein’s second species, Blepharisma undulans, with
the following characters: Size of longest individual observed
220u, shortest, 70u. Color variable, from deep purple violet
to light rose, or perfectly colorless. Form variable according
to the condition of the posterior end, but is usually sufficiently
striking to be easily identified. It is somewhat lancet formed,
thickest in the middle and tapering towards the two ends (fig. 1).
The posterior end is usually broader than the anterior, and, save
in exceptional cases, does not come to a point but ends in a broadly
rounded surface. The anterior end, on the contrary, comes to
a rounded point in which three lines converge, the left margin
of the body, and the right and left borders of the peristome.
The peristome is a deep cleft on the right ventral side of the
body running and deepening slightly in an oblique direction
towards the left (fig. 2). The adoral zone, composed of long
slightly tapering membranelles, begins on the left of the anterior
tip and is continued posteriorly on the left peristome margin to
the mouth where it sinks below the surface with the peristome.
An undulating membrane arises close to the mouth deep in
the peristome and extends along the right margin of the peri-
stome and towards the anterior end, stopping half way between
the mouth and the end. It is about one-half as broad as long
and extends well beyond the surface of the body so that it can
be seen distinctly with the lower powers of the microscope (fig. 3).
A second undulating membrane, much smaller and very diffi-
cult to see, even with the highest powers, extends from the
mouth along the left margin of the peristome inside of the adoral
zone, for a distance equal to about half that of the right mem-
brane. This is probably the same thing as the ‘lip’ described
by Anigstein ‘in B. clarissima. Unlike B. lateritia the body,
as a rule, is only slightly if at all compressed, but at times it is
definitely flattened especially in the oral region. The body
stripings are distinct and from nine to sixteen in number. The
macronucleus is single, double or multiple, and free micro-

670 GARY N. CALKINS

nuclei are absent. The contractile vacuole is terminal and is
usually accompanied by one or more fluid vacuoles. Very
often the entire posterior part is drawn out into a fine vacuole-
filled thread which gives to the organism a striking resemblance
to Spirostomum téres. Not infrequently this entire poste-
rior vacuole-bearing part is pinched off, leaving the remainder
of the body with a decidedly truncate base, recalling Miiller’s
Trichoda aurantiaca (fig. 9).

Habitat and food

Miller found his Trichoda ignita among Lemna in fresh
water, Bory found Ypsistoma on Oscillaria, and Ehrenberg,
Bursaria lateritia on conferva. Stein also speaks of it in con-
nection with filaments of algae. Biitschli speaks of its food
only in a general way as ‘Nahrung fein.’

I found a large number of specimens among decaying algae
in pond water at Woods Hole, Massachusetts, in July, 1911.
One individual was isolated in filtered water from the same
source, and experiments were immediately begun to deter-
mine the best medium for its continued cultivation. Without
going into details in regard to these experiments it will suffice
here to say that the most satisfactory medium was found to
be pond water with about 50 per cent of twenty-four hour hay
infusion. On this medium which is renewed daily, the descend-
ants of the originally isolated specimen have divided 224 times up
to the present day (June 12), or less than one division per day on
the average. All specimens have been colorless for the last
five months. The food, therefore, appears to be bacteria of
ordinary hay infusions, the best results being obtained with a
relatively small number of bacteria, or in other words, with as
little as possible of the bacterial waste matters.

Later in the summer (September) an effort was made to find
other wild specimens from the same source as the original but
without success, so that the observations here recorded were
made on material derived from the original specimen isolated
on the 20th of July. The full history of this culture, to-

CONJUGATION IN BLEPHARISMA UNDULANS 671

gether with experiments on the effects of chemicals of differ-
ent kinds, will be published in another place. In the present
paper I desire only to present the observations on division and
paedogamous conjugation.

Division of Blepharisma undulans

When conditions are fairly uniform in the cultures Blephar-
isma divides about once in twenty-four hours. No regular
rate, however, is characteristic; at one period during the earlier
stages of the culture the organisms divided as often as three
times in one day, but later the division rate fell as low as two
divisions in ten days.

When ready to divide, the cells appear slightly swollen in
the middle and a new peristome appears in the posterior half of
the cell (fig. 4). The individuals become sluggish and, even before
the adoral membranelles are visible, a contractile vacuole ap-
pears immediately posterior to the mouth and in the most
swollen portion. The membranelles of the posterior adoral
zone first appear as cilia slightly more conspicuous than the
cilia on the general surface. At this time there is no trace of
the posterior peristome. A shallow furrow soon appears, how-
ever, and this gradually deepens, while the membranelles grow
in size and vibrate with a rhythmic motion from anterior to pos-
terior (fig. 5). The division furrow begins immediately posterior
to the central vacuoles, below the point of maximum swelling,
and then cuts in from the ventral surface towards the dorsal.
Throughout this period I have been unable to see the conspicu-
ous undulating membrane on either half of the dividing organ-
ism. On the posterior cell it does not appear until the mouth
breaks through in the last five minutes of division. In the
anterior half I have watched its vibrations up to the time of the
* beginning of constriction, when it could not be seen again until
the end stages. Whether this anterior membrane is thrown
off and another formed to replace it I cannot say. Possibly
the mouth being closed at this period of division, the mem-
brane is folded down on the floor of the peristome and is thus

672 GARY N. CALKINS

out of sight. During this middle period of division the organ-
ism remains quiescent or moves but slightly, but as the end
stages come on it moves more freely and with greater rapidity,
both undulating membranes in active motion. Not only are
new membranelles formed in the posterior half but the old
membranelles of the anterior portion are also replaced by new
ones (fig. 6). The new membranelles appear as minute cilia
within the old row and grow quickly, moving at first with a
spasmodic activity quite different from the regular wave-like
motion of the older membranelles. As growth of the new mem-
branelles progresses the regularity of motion increases until two
full sets of membranelles are in play. This condition lasts but
a few minutes, however, as the older set begins to fade away
by absorption until finally replaced by the active new set.

None of the older observers was able to make out a micro-
nucleus in the ordinary vegetative stages, nor during division,
but Biitschli observed two ‘encapsulated’ smaller nuclei during
the process of conjugation of B. lateritia which he regarded as
micronuclei. Balbiani (’60) also is said to have observed the
micronucleus in B. lateritia. Biitschli remarks on the diffi-
culty experienced by himself in demonstrating the micro-
nucleus and states that he was unable to identify it in the ordi-
nary vegetative forms, although he did not doubt its presence
in the ordinary cells, since he found it in multiple number in
conjugating specimens.

The difficulties in demonstrating the presence of micronuclei
have in no wise decreased with the lapse of time since the obser-
vations of these gifted pioneers. Different methods, however,
and especially the use of sections, have enabled us to trace
these elusive bodies and to get some light on their history, al-
though it must be admitted that this history does not corre-
spond with that of any other known ciliate. They are intimately
associated with the macronucleus, and in resting phases of the
cell I have failed to find positive evidence of their presence.
During division, however, they appear as extremely minute,
deeply-staining bodies which apparently divide within the
nuclear membrane of the macronucleus (fig. 8). In total

we
CONJUGATION IN BLEPHARISMA UNDULANS 673

preparations of the dividing cell it is possible to get them in
profile, when they appear associated with the macronucleus
very much as do the micronuclei of ordinary ciliates like Para-
mecium or Colpidium. The relations come out best in sec-
tions and especially sections of late stages in division. In
early stages they are apparently within the nuclear membrane
and are difficult to distinguish from the thick, granular chro-
matin of the macronucleus. Here and there, however, in the
mass of chromatin, larger and more definite deeply staining
bodies may be seen which may be the micronuclei (fig. 3). In
later stages they become much more definite, although still
small and inconspicuous. They lie close against the chroma-
tin of the macronucleus and within the delicate nuclear mem-
brane surrounding this organ. In some cases they are separated
by a shght space, but this condition is exceptional. Ordinarily
they lie in pairs (fig. 8), the highest number that I have seen
being eight, or four pairs. The remarkable significance of this
intranuclear position comes out in connection with the phe-
nomena of conjugation to be described later.

The macronucleus, usually so definite in form and simple in
its changes during vegetative life and division, is extremely
difficult to interpret in the case of Blepharisma undulans. Stein
described B. undulans as having two macronuclei, one in each
half of the cell and with or without connecting chromatin threads.
This condition is fairly common, but so also is the uninucleate
condition and the multinucleate condition, while interme-
diate phases connect these two extremes. The single macro-
nucleus is compact and homogeneous and may or may not
have the characteristic ‘Kernspalt’ common to many of the
ciliates. A common occurrence is the presence of two or three
macronuclei grouped together as shown in figure 1, and in these
cases there is a difference in texture between the different ele-
ments. I am inclined to interpret these as incomplete stages in
the fusion of the portions of the macronucleus after division, for
the nucleus during division always begins as a unit, stretching
out as a single strand and breaking up into numerous portions
at the end of division (figs. 2, 7, 8). The difference in texture

*

674. ‘ GARY N. CALKINS

of the different portions is enigmatical and I have no sugges-
tions to offer in interpretation. While the majority of the
macronuclei are deeply staining, granular and uniform in tex-
ture, this one portion is much less granular, stains lightly and
is more homogeneous than the denser portions. In some cases,
furthermore, there is no such difference between the different
portions, hence it may be only a transient or even a degenerate
condition. Opposed to the latter hypothesis is the fact that
the same texture is noted in the new nuclei formed as aresult
of conjugation. In one case this lighter portion was found at
the extremities of the elongated macronucleus in division, where
it appeared something like the centrosphere of Noctiluca in
division but in other division figures there was no such differen-
tiation.

In division, the central portion of the elongated strand be-
comes thinner until it finally breaks in the middle. The daugh-
ter nuclei do not round out then into definite homogeneous
nuclei, but the reconstruction is accompanied by a fragmen-
tation of the macronuclear material into from two to eight sep-
arated fragments of diverse size. Some of these portions are
extremely small, and when rounded out, have the appearance
of micronuclei, but no two cells are alike in this respect, a fact
which, together with their mode of origin, contradicts any in-
terpretation as to their micronuclear nature. The chromatin
of the macronucleus is finely granular and has the tendency
to collect about the periphery of the elongated strand, leaving
the central portion lighter, thus giving the impression of an
intranuclear canal. In one case the cell was of giant size, with
a macronucleus in the reconstruction stage of division, although
there was no evidence of division of the cell body (fig. 2).

The end stage of division is characterized by the peculiar
twisting movements observed in so many ciliates, the last con-
necting strand of cortical plasm finally giving way and freeing
the two cells, each now with its full complement of cellular
organs, peristome, nuclei, and contractile vacuoles.

Or

CONJUGATION IN BLEPHARISMA UNDULANS 67

Occurrence and phenomena of conjugation

O. F. Miller (1786) described the conjugation of Trichoda
aurantiaca and gave a figure that is easily recognized as a spe-
cies of Blepharisma. Stein described a number of individuals
which he regarded as ex-conjugants, but did not see the con-
jugating cells and drew an erroneous conclusion as to the method
of fusion of the two individuals. Biitschli (76) was the first
to describe the process of conjugation in B. lateritia, making
out the method of fusion of the two individuals and the pres-
ence of the micronucleus. An interesting feature of this work
is that ex-conjugants were unable to live, many individuals
dying on the second day after separation, the remainder on
the third day.

Unfortunately I have been unable to study the exogamous
conjugation of B. undulans, owing to the failure to find other
specimens than those originally isolated. But paedogamous
conjugation occurs constantly and the following account refers
to that alone.

All who have carried on careful cultures of protozoa, isolate
individual cells each day and transfer them to fresh culture
medium, while the remainder are kept together in larger dishes
(I use Syracuse watch glasses) and allowed to accumulate in
large numbers as ‘stock.’ Shortly after beginning observations
on Blepharisma it was noticed that conjugating individuals were
frequent in such stock, amounting in some cases to regular
epidemics. No degree of relationship seems too close to pre-
vent such union, in one case, for example, an individual was
isolated in the 104th generation of the race; it divided during
the night and on the following day the two daughter cells were
firmly united in conjugation. This, I believe, is the closest case
of paedogamy in ciliated protozoa on record.

The externals of conjugation of B. lateritia as described by
Biitschli agree fully with those of B. undulans. The two indi-
viduals unite, first at the anterior tips, as in the case of Para-
mecium caudatum, and then fuse along the somewhat spirally
running central line of the peristome. Stein’s assumption that

676 GARY N. CALKINS

the edges of the peristome fuse is erroneous, while Biitschli
was right in stating that both margins of the peristome are free
(fig. 17). It thus happens that the adoral zone of one individual
overlies the line of fusion on one side, while the adoral zone
of membranelles of the other individual covers the line of fusion
on the opposite side. Fusion, furthermore, does not extend
posteriorly as far as the mouth but involves only about three-
quarters of the peristome, leaving the mouth and a large por-
tion of the undulating membrane free (fig. 14). B. undulans
is never very active but during the process of conjugation it is
even more quiescent than usual. The individuals remain united
for a period of from eighteen to twenty-four hours, rarely longer
than this, and finally separate. The process of separation
begins with the vacuolization of the zone of fusion (fig. 22).
The walls of the central vacuoles break down, leaving the two
cells connected at the anterior tip and at the posterior point of
fusion. The latter connection breaks first and the organisms
finally separate at the anterior tips (fig. 21).

The conjugating population is distinctly smaller than the
non-conjugating, a fact easily noted by the eye. Statistical
confirmation of the size differences has been made by Miss
Watters, whose paper on the subject will shortly appear, and
she permits me to summarize her results to date as follows:
482 conjugating and 1089 non-conjugating individuals were
measured, the average length of the former being 110.08 mi-
crons, and of the latter (including ex-conjugants) 140.37 microns.
The longest conjugating individual was 156.25, non-conjugat-
ing 221.385 microns; the shortest conjugant was 85.93 microns,
non-conjugant 70.31 microns. There can be little doubt
that the shortest non-conjugant was an ex-conjugant, for these
are always minute.

Stein believed that small forms with numerous nuclei were ex-
conjugants, but Biitschli regarded such forms as abnormal
specimens, not necessarily ex-conjugants. Under strictly nor-
mal conditions the nucleus may be multiple in B. undulans,
without indicating any connection with conjugation. The
smaller size of the suspected ex-conjugants and the evidences

CONJUGATION IN BLEPHARISMA UNDULANS 6077

of degeneration, which in Biitschli’s and my experience follows
conjugation, make Stein’s interpretation more plausible.

Bitschl’s observations were more complete and more defi-
nite than Stein’s although the finer details are still unknown in
B. lateritia. A free translation of his deseription of the nuclear
changes runs as follows:

The uniform and finely granular (macro) nucleus as shown by treat-
ment with acetic acid, undergoes no changes during the entire process
of conjugation, and passes on to the separated animal in its original
form. In flattened conjugating animals treated with 1 per cent acetic
acid on the other hand, one notices very distinctly, a number of small
nucleolus (micronucleus)-like bodies. These give the impression of
small cell nuclei which appear as though enclosed within a dark mem-
brane and consist of a dark central or ex-centric granule. Once fine
fibers were seen running from this central granule (karyosome) to
the membrane, and again one finds these vesicles sometimes without
a central granule (karyosome) but composed of fine particles. The num-
ber of such bodies found in the cells of conjugating animals varied
within rather wide limits; once there were only two, again three,
seven, eight, while in one pair the one individual contained eleven the
other one six. After conjugation there is no further evidence of these
bodies; on the other hand in a specimen that had just finished conju-
gation there was one very distinct nucleolus capsule in front and
one behind the nucleus (macronucleus). The capsules were beauti-
fully striated with two dark staining bodies in the centers of each
fiber, which together formed a double granular zone in the equator
of the capsules.

In place of these two undeniable nucleolus capsules one finds in
animals shortly after conjugation two minute pale bodies of oval form
on the side of the nucleus. These, after treatment with acetic acid,
show large dark granules, They grow rapidly into pale round spheres
in which an ex-centric ‘eranule is invariably present. In contrast
to this development the nucleus (old macronucleus) becomes pro-
gressively smaller and denser and shows unmistakable signs of degen-
eration.

On the third day after completed conjugation, animals were found
in which there was no trace of a (macro) nucleus and there is no doubt
that the degenerated nucleus had been thrown out of the cell. The
new light spheres continued to enlarge and appeared to be made up
of small granules while the larger bodies had apparently disappeared
(Pe. pp: 35-316):

All of the stages observed by Biitschli I have been able to
duplicate in the case of B. undulans, and in addition have been
able to trace out the beginnings and to a certain extent the his-

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

678 GARY N. CALKINS

tory of the ‘capsules.’ The first statement, however, in regard
to the unchangeableness of the macronucleus cannot hold for
B. undulans. On the contrary, this is the very center of nuclear
activity, and its early changes initiate the internal processes
of conjugation. These early changes are characterized mor-
phologically by three conditions of the macronucleus, not pres-
ent at other times: (1) all macronuclear material is concen-
trated in one definite oval macronucleus; (2) this oval nucleus
is enclosed in a thick doubly refractive cyst-like membrane;
(3) the cleft or Kernspalt which appears occasionally in the
vegetative nucleus, is here continued into a clear hyaline space
which separates the nuclear material into two distinct parts.

The material of the macronucleus at this stage is condensed
into a firm, massive nucleus with a characteristic appearance
quite different from its appearance at other periods. The
eyst-like membrane has not the least resemblance to nuclear
membranes as usually seen. It has a distinct double contour
and is separated by a clear space from the chromatin contents,
and if seen free from the surrounding cell, would give to the
macronucleus the appearance of an encysted flagellate (fig. 12).

The early changes of the micronuclei have not been seen,
that is, their emergence from the macronucleus. Until fusion
is completed they remain comparatively small and difficult
to see, showing no such enlargement as do the micronuclei of
Paramecium or other ciliates in which conjugation has been
worked out, and staining with the greatest difficulty. In rest-
ing cells the characteristic number appears to be four. In
conjugating cells there is often, but not always, a fifth micro-
nucleus-like body which takes no part in the processes accom-
panying conjugation (figs. 12, 18 and 20). It degenerates and
ultimately disappears. It may be an extra vegetative micro-
nucleus, since the number of these nuclei varies in the vegeta-
tive cells; or it may be a residual degenerating nucleus from a
first maturation division of the four nuclei, an interpretation
which would bring the maturation phenomena of Blepharisma
in accord with the phenomena in other ciliates. On this inter-
pretation there would be three other similar nuclei to account

CONJUGATION IN BLEPHARISMA UNDULANS 679

for, and it seems hardly probable that they, or the spindle fig-
ures giving rise to them, should have been overlooked. On
the other hand they are so minute and so difficult to stain that
this possibility must remain open.

The first evidence that I have found as to maturation divi-
sions is the late stage in mitosis of the four functional micro-
nuclei (fig. 12). They lie close together in the anterior third
of the cell, the daughter nuclei appearing first as homogeneous
chromatin then as vesiculate nuclei. One of the eight result-
ing nuclei divides again, forming the two pronuclei, while the
other seven swell, degenerate and finally disappear (fig. 14).
The two pronuclei are practically the same in size and appear-
ance, although in one instance the two migratory nuclei were
larger than the stationary (fig. 13). At the time of fusion the
pronuclei are elongate and spindle formed, union taking place
first at the ends (figs. 14 and 15). After fusion the synearyon
becomes spherical and the chromatin is uniformly distributed
(fig. 16).

The interchange and fusion take place at the posterior limits
of the zone of coalescence and the fertilization nuclei remain in
this region during their further changes. Each nucleus forms
a spindle and divides by typical mitosis (figs. 18 and 19), the
full spindle figure (fig. 19) being much larger than any of
the maturation nuclei and showing typical fibers and chro-
matin. The two nuclei in each cell resulting from this first
division do not pass immediately into the second spindle stage
but evidently remain for some time in the resting stage, since
they are often found in various positions relative to the old
macronucleus (figs. 21, 22 and 23). Each nucleus then divides
again by mitosis, this being the stage described by Biitschli.

The period of nuclear quiescence after the first division is
characterized, in the majority of cases, by the commencement
and completion of separation of the conjugating cells. This
is accomplished by the progressive vacuolization of the proto-
plasm in the line of fusion. The animals separate, first in the
center of the fused areas by coalescence of the vacuoles, then at
the posterior or oral region, and finally at the anterior tips.

680 GARY N. CALKINS

In the meantime the old, encapsulated macronucleus under-
goes fragmentation, the process showing no time relations in
regard to the changes of the micronuclei (figs. 13 to 23). Sooner
or Jater the fragments show a well-marked granular degenera-
tion (fig. 14) and ultimately disappear, probably by absorption
in the protoplasm.

The four daughter nuclei formed by the second division of
the fertilization nucleus next show characteristic changes lead-
ing to the formation of the new macronuclei. They appear as
deeply staining granules surrounded by a more feebly staining
homogeneous matrix. The latter increases remarkably in bulk,
probably by elimination of chromatin material from the micro-
nucleus, until it attains the size of the macronucleus in vegeta-
tive cells (figs 24 and 25). These are the ‘pale round spheres’
which Biitschli described in 1876. In them, either central or
ex-centric, are the micronuclei which are occasionally found
in the process of division (figs. 24 and 25). These ‘pale round
spheres’ are the new macronuclei. The further history of the
nuclei of the ex-conjugants is unknown. The organisms rapidly
grow smaller and ultimately die without division.

DISCUSSION

The structures ard activities of Blepharisma undulans bring
up again some large questions in protozoology. Of these, two
in particular demand further consideration, viz., the origin
and significance of macro- and micronuclei, and the effects of
conjugation.

There is a strong probability that B. undulans and B. lateritia
are one species, since the nuclear and peristomeal structures
of undulans are so variable. In both types the micronuclei
have hitherto been overlooked in the vegetative stages. Ble-
pharisma clarissima Anigst., on the other hand, appears to be a
distinct species with a characteristic beaded macronucleus
and numerous micronuclei visible in the vegetative stages.
The contractile vacuole, also, with its feeding canal is a specific
characteristic.

ental

CONJUGATION IN BLEPHARISMA UNDULANS 681

The failure of previous observers to find micronuclei in Blepha-
risma undulans in the vegetative stages is explained by their
formation and retention within the macronuclear membrane.
Lebedew (’08) describes micronuclei emerging from the macronu-
clei during conjugation of Trachelocerca phoenicopterus, and
the observations of Neresheimer (’08) show that some analo-
gous process occurs in the parasitic ciliate Ichthyophthirius.
The latter observations seem to be inconclusive, and the possibility
at least is open that the new micronucleus does not actually
penetrate and enter the old macronucleus, but gives rise to a
new macronucleus as in Blepharisma, by secretion. In this
connection Neresheimer says:

The most remarkable thing after encystment is the origin of the
micronucleus (Nebenkern), which up to now has invariably been seen
in young specimens without any inkling as to where it comes from.
It appears suddenly and in a most peculiar manner, when from twenty
to thirty fragments are present in the cysts. In sections of this stage,
on the inside of the now spherical nucleus, there is a small, highly
staining granule which appears like a comet with a tail of plastin and
chromatin streaming out behind, connecting it with the nucleus. Soon
after, this connecting strand disappears and the complete micronucleus
takes a position some distance from the macronucleus.

Neresheimer does not say that it emerges from the macro-
nucleus, but one certainly gets the impression that it is thrown
out of the macronucleus. From now on the micronucleus
divides with a typical mitotic figure at each division of the
cell, while the macronucleus divides by direct division. Ulti-
mately the micronucleus divides twice in succession; three of
the four resulting nuclei degenerate, the fourth divides again
to form male and female pronuclei characteristic of the ciliates.
Conjugation occurs not between two cells, but by union of the
two nuclei thus formed and the synearyon then wanders into
the old macronucleus where it remains for a time as a karyo-
some and finally disappears.

While it is probable that the micronuclei in these young
forms emerge from the nucleus, we need further evidence than
Neresheimer brings forward to support the view that the syn-
caryon penetrates the old macronucleus instead of forming a

682 GARY N. CALKINS

new one by division or by secretion. In Blepharisma the new
macronucleus is entirely different in texture and appearance
from the old and there is no doubt of its new origin.

In Blepharisma then, we find quite a different process of
nuclear metamorphosis from that which occurs in the better
known ciliates like Paramecium or Colpidium, although in
essence they are the same. In Paramecium caudatum four
of the daughter micronuclei form four new macronuclei, losing
their identity as micronuclei with the metamorphosis. In
Blepharisma four, that is, all of the micronuclei, form macro-
nuclei but retain their identity within the organs which they
create, remaining and dividing within the macronuclear mem-
brane until the next following conjugation.

Compared with the rhizopods the relations of idio- and tropho-
chromatin are here reversed. In the former the idiochromatin
is extruded from the trophochromatin or vegetative nucleus,
remaining in the cell body as chromidia. In Blepharisma the
trophochromatin is excreted from the micronucleus resulting
in the ‘purification’ of the idiochromatin which remains per-
manently in the macronucleus as the micronucleus. In both
cases it amounts to the same thing in the end, both idio- and
trophochromatin coming from the syncaryon, or its immediate
derivatives.

It is not improbable that other forms of ciliates in which
micronuclei have not been observed in the vegetative stages,
will be found to have nuclear relations analogous to those in
Blepharisma. Here, also we have a clue to the explanation
of the nuclei of forms like Opalina where dimorphic nuclei
are absent. Such nuclei can be interpreted as amphinuclei
in which idio- and trophochromatin are not separated but re-
main permanently as simple nuclei. In these the idiochro-
matin prevails over the trophochromatin in so far as the vital
reproductive processes are concerned. In the majority of
other ciliates the dimorphic conditions are fully established,
an illuminating stage being seen in Blepharisma undulans where
the two remain together, but, nevertheless, are differentiated.
From such a condition to that in forms like Paramecium is

CONJUGATION IN BLEPHARISMA UNDULANS 683

but a step in advance. The syncaryon of Blepharisma divides
twice before nuclear differentiation. In Paramecium it divides
three times. If the four nuclei of Blepharisma should divide
again, and by the division should separate tropho- from idio-
chromatin, the former then swelling into macronuclei, the con-
ditions would be identical with those of Paramecium.

Such an hypothesis of the origin of the dimorphic nuclei of
ciliates is much more plausible than one based upon hypothet-
ical binucleated ancestral forms. Metcalf (’09) for example,
has advanced an elaborate theory requiring a number of sup-
positions: first, a delay in the process of division, establishing
a temporary binucleated condition; second, a complete sup-
pression of the delayed division, making the binucleated con-
dition permanent; third, a shifting of division planes from
longitudinal to transverse. Not one of these suppositions has
sufficient evidence to warrant it. The facts indicate that the
macronucleus represents the trophochromatin contained in the
original syncaryon from which it is derived either by secretion,
as in Blepharisma, or by differentiating division, as in Paramecium
and its allies. No complicated phylogenetic hypothesis is needed
in the light of these. .

The second problem suggested by the history of Blepha-
risma undulans is the much discussed matter of conjugation and its
significance. The process and its sequences in Blepharisma throw
no new light on the subject but rather add to the perplexity
of the tangle by an additional difficulty, for conjugation in Ble-
pharisma under cultural conditions, is equivalent to a death warrant.
Bitschli (76) noted that ex-conjugants of Blepharisma lateritia
invariably die: ‘‘Linger gelang es mir nun nicht die aus der
Conjugation hervorgegangenen Thiere lebend zu erhalten: schon
am zweiten Tage nach lésung der Syzygie starben viele ab,
der Rest am dritten Tage’ (p. 316). Nevertheless he made
no remark on this apparent exception to his theory of Ver-
jungung in explaining the significance’of conjugation. Maupas,
likewise, observed fruitless conjugations, interpreting them as
due to close relationship. A few other observers have also called
attention to the invariable death of ex-conjugants. Baitsell (’11),

684 GARY N. CALKINS

for example, isolated 132 pairs of closely related conjugating
Stylonychia pustulata: ‘‘ None of these ex-conjugants divided and
none lived forty-eight hours after separating.”

Two hundred pairs of conjugating Blepharisma undulans
have been isolated and followed in culture after conjugation.
The observations were made at different times during the year,
to allow for possible seasonal predilection. The results have
confirmed Biitschli’s observations in every respect. Not a
single ex-conjugant divided and the majority died within ten
days; two, only, dragged out a miserable, weakened existence
for twenty days. They become smaller and more vacuolated,
until finally they bear little resemblance to the original robust
forms from which they came (figs. 10, 11, 23).

If this negative result were common to all ciliates in culture
there might be a more reasonable chance of interpreting con-
jugation. But this is not the case, nor is it common even
amongst paedogamous ex-conjugants. Paramecium, Colpidium,
Gastrostyla, Euplotes, Vorticella, Didinium, and a score of
other ciliates have been cultivated after conjugation. Stylo-
nychia, with Maupus, continued to live. Death after conjugation,
therefore, seems to be rather an exception than a rule.

There is reason to believe that failure to live after conju-
gation is due to the conditions in the laboratory cultures. We
have found, for example, that while ex-conjugants of Parame-
cium caudatum, from natural ponds or other external sources,
continued to live in the proportion of from 70 to 80 per cent,
those derived from prolonged cultures continued to live in the
proportion of only 6 per cent (Calkins ’08, Cull ’07). The
reason for this discrepancy must be sought in the conditions
under which conjugation occurred. So, too, it is probable that
the failure of Blepharisma to live after conjugation is due to
some imperfection in the cultural conditions in the laboratory.
Here the ‘infant mortality’ is 100 per cent, but in Paramecium
caudatum under culture, the infant mortality was 94 per cent,
almost as bad, while under conditions more nearly like those
in nature the rate fell to 30 or 20 per cent. The theoretical
interest centers in the 6 per cent of the culture Paramecium

CONJUGATION IN BLEPHARISMA UNDULANS 6085

that continued to live. Jennings, according to a recent paper
(12), might argue that, since these 94 per cent of ex-conjugants
die, conjugation does not imply rejuvenescence as Biitschli,
Maupas, and others have maintained. The continued life of
the 6 per cent, on the other hand, indicates that these varied
in some way from the others and represented a different mix-
ture of germ plasms. If we go back to the original records of
these 6 per cent we find that in no case, either exogamous or
endogamous, did both individuals of the syzygy continue to
live, one lived, the other died after from one to thirteen divi-
sions. It is evident then, that even the admixture of the germ
plasms from the same cells is not equally potent. The result
may be interpreted either as a fortunate combination of pro-
nuclei, or as a particularly favorable cytoplasmic medium in
which the syncaryon acts, or both together. In any case it
may be granted that conjugation does involve the possibility
of increased variability, but this does not exclude the older view
of rejuvenescence or renewal of vitality through conjugation.
We have repeatedly shown that a change in the chemical com-
position of the medium will counteract a period of depression
in organisms maintained on a constant diet, and renew the
vitality, as indicated by the increased division rate. Engel-
mann (’76) early showed that conjugation results in the ‘reor-
ganization’ of the cell, and it is now a well-known fact that in
this process of reorganization the old macronucleus fragments
and ultimately disappears in the cytoplasm. This disappear-
ance must give rise to a great increase in the nucleo-protein
content of the cell, therefore to a new chemical composition
of the cell as a whole. We have recently shown that, under
certain conditions, nucleo-proteids (especially the purines),
have a markedly stimulating effect on the rate of cell division
(Calkins, Bullock and Rothenburg 712). With this chemical
change in the cell there is a condition which, theoretically,
should work in the same general way as a chemical change in
the surrounding medium and give rise to a renewal of vitality.
The syncaryon in Paramecium caudatum, as we have seen,
divides three times without cell division, and eight nuclei are

686 GARY N. CALKINS

formed. Four of these nuclei become macronuclei, four micro-
nuclei. These are formed during the process of disintegration
and absorption of the old macronucleus. Then follows two
divisions of the cell, resulting in four individuals, each with
the normal nuclear relations. Each cell is now a new individual
with renewed chemical basis and newly combined nuclei. The
power to vary follows from the increased powers of metabolism,
and it seems to us that the phenomena of conjugation are more
reasonably interpreted as processes which have developed for
rejuvenating effects whereby powers of metabolism, irritability,
reproduction, and variation are reinvigorated, rather than as
processes for engendering variations. If it is argued that 100
per cent of Blepharisma die after conjugation and, therefore,
that rejuvenescence is not an effect of conjugation, the same
argument may be returned and the question asked: Where does
variation come in when 100 per cent die? We believe that
unknown, probably cultural, conditions enter into the problem
and make conjugation even more difficult to interpret under
either hypothesis.

.

June, 1912.

LITERATURE CITED

ANIGSTEIN, P. 1912 Beobachtungen_iiber zwei neuen Ciliatenarten. Arch.
f. Prot.

Baitsety, G. A. 1911 Conjugation of closely-related individuals of Stylon-
ychia. Proc. Soc. Exp. Biol. Med., vol. 8, p. 74.

Ba.siANI, E. G. 1860 Recherches sur les phenomenes sexuelles des Infusories.
Jour. de la Physiologie, tom. 4.

Birscuur, O. 1876 Studien tiber d. erst. Entwickelungsvorgan. d. Eizelle.

d. Zelltheilung u. die Conjugation der Infusorien. Abh. d. Sencken-
berg. naturforsch. Gesell. Frankfurt a/M., Bd. 10.

Cauxins, G. N. 1903 Studies on the life history of Protozoa, I. Arch. f. Ent-
wicklungsmechan., Bd. 15, p. 139.

Catxins, G. N., Butuock, F. D. anp Roruensera, G. L. 1912 The effects of
chemicals on the division rate of cells. Jour. Infect. Diseases, May.

Cun, Sara W. 1907 Rejuvenescence as a result of conjugation. Jour. Exp.
Zool., vol. 4.

CONJUGATION IN BLEPHARISMA UNDULANS 687

EHRENBERG, C. G. 1831 Ueber die Entwickelung und Lebensdauer d. Infu-
sionsthiere. Abh. d. Berlin. Akad., p. 1.

ENGELMANN, W. 1876 Ueber Entwickelung und Fortpflanzung von Infusorien.
Morph. Jahrb., Bd. 1,-p. 573.

Jennines, H. §. 1912 Harvey Lecture. Pop. Sci. Mo., May.

LesBepEew, W. 1908 Trachelocerea phoenicopterus. Arch. f. Prot., Bd. 13, p.70.
Mertcaur, M. M. 1909 Opalina. Arch. f. Prot., Bd. 13.

Miuter, O. F. 1786 Animalcula Infusoria, ete. Leipzig.

NERESHEIMER, C. 1907 Fortpflanzung eines parasitischen Infusors (Ichthy-
ophthirius). Sitz. Ber. f. Morph. u Physiol., Miinchen.

Prerty, R. 1849 Mikroskopische Organismen der Alpen u. d. italienischen
Schweiz. Mitt. Naturforsch. Ges. in Bern.

1852 Zur Kenntniss kleinster Lebensformen, etc. Bern.
Stein, F 1867 Der Organismus der Infusionsthiere. Bd. 2. Leipzig.

PLATE 1

EXPLANATION OF FIGURES
Same magnification throughout

1 Normal Blepharisma undulans with macronucleus in typical vegetative
condition after division. From total preparation.

2 Giant specimen with macronucleus in division stage but with no indication
of cell division. Three micronuclei shown. From total preparation.

3 Section showing details of peristomial region and two micronuclei.

4, 5, 6, 7 Successive stages in cell division. Note beginning and develop-
ment of posterior adoral zone. From total preparations.

8 Normal nuclear relations in a dividing cell. Seven of the eight micro-
nuclei shown within the macronuclear membrane. From section.

9 Involution truncated form after casting off the elongated posterior end.
New vacuoles replace those lost. From total preparation.

10 Typical ex-conjugant. Two nuclei from the first division of the syn-
caryon on opposite sides of the old macronucleus. Note the small size of the
cell as contrasted with figure 1. Total preparation from the right side.

11 Degenerating ex-conjugant five days after separation.. From total prep-
aration.

12 Conjugating Blepharisma with macronuclei, enigmatical micronucleus,
and four micronuclei in the late anaphase of mitosis—evidently the second matu-
ration division. From total preparation.

688

1

hi
4

PLAT

CONJUGATION IN BLEPHARISMA UNDULANS

GARY N. CALKINS

fudebid
* om papemaemnnnerny me MTN TS

10

Wi eS
Sad ee
\ ‘ :

PLATE 2

EXPLANATION OF FIGURES

13 Interchange of micronuclei in form of minute spheres. (These may be
fused nuclei and not pronuclei, in which case the two smaller nuclei are ‘reduc-
tion nuclei.’) From section, X 1000.

14 Interchange of micronuclei. Four pronuclei in the region of interchange.
Seven reduction nuclei in individual on right, four on left. From section, X
1000.

15 Union of pronuclei. Three reduction nuclei in same vicinity in each cell.
From section, 1000.

16 Conjugating individuals. The synearyon in each is swollen and the macro-
nuclei are undergoing fragmentation. From section, < 500.

17 Early stage in conjugation, with the first changes of the micronuclei.
Method of fusion and overlapping of the adoral zone distinctly shown. From
total preparation, < 400.

18 Synearyon in each cell preparing for the first division. Breaking up of
the macronucleus within its capsule. From section, * 1000.

19 Full mitosis of the synearyon, first division. From section, X 1000.

20 Second division of the synearyon (on left). A third enigmatical body
in each cell. From section, 1000.

21 Pause after first division of the syncaryon. End stage in the separation
of the conjugants. From total preparation, < 400.

22 Vacuolization of the zone of fusion preparatory to separation. Nuclei
in stage of pause. From total preparation, < 400.

23 Ex-conjugant in same stage as that of figure 21. From total preparation,
x 400.

24 Ex-conjugant showing two new macronuclei with central micronuclei.
The micronucleus in one has just divided as shown by the connecting strand.
From section, * 500.

25 Ex-conjugant showing three new macronuclei (the fourth is in another
section). From section, * 500.

690

2

PLATE

ATION IN BLEPHARISMA UNDULANS

CONJUG

GARY N. CALKINS

’

ATA

4a wobtaiey

THE DEVELOPMENT OF THE SKULL OF EMYS
LUTARIA

B. W. KUNKEL

From the Anatomisches Institut, Freiburg 7. B., and the Sheffield Biological
Laboratory of Yale University, New Haven

THIRTY-ONE FIGURES

CONTENTS
AMG O GUC ET OMA Yes) s1ss0/5 2 chevaichs Gare bk che ces hx loss kd a ee Le Gee 693
Generalerormor, Che: Skulls.) 5 aco. cisites a cto ere cae 695
iamumpoasalerand chords, Gorsalishe: 119251 ae eee eon een eee aca 696
IRCA) CYKCN OIE) hae RP aA Ean Dien een) Ui ee | eee Me 5 eo alas 702
Regio otica... Fee 703
(Cin OySIUIILE), RTC A eee ee ems PRA ines ety Sie ints 5 al sats iod Meet adoloa cals 706
MRO LGINCS DSULA, OLLCR:..--5 6.2.00 ss, 70 TE pw a7ee 8 daa ee alee CA eet el eee 712
PRE CEUTMBMOSTERIUS 3 fas o.a\cies ticd-ape ae cae Gas hake epoca Moreno OPE NCR ec taae ch eee 719
CHISGE ORION CO et ie oe, ein Eh Ftc oru tks Pio So oo 0 oc 719
nelainonsonthemenrves is the regio ovlca. ..cs-eiie ee neki eerie cen aaa 720
VECTOROG MUO PEMD OLALISY, scisierctess. ones os anactosoye 6 See Eck ae ee OR IO oe fe
IRGENO CULTIOHG FINS Aan eae RR eEnT AH dairnais ayaa d mown mia cious chon an arate c 730
Paioquadne tum and mandibles. «...3.-fiycces tose ce aioe ae eee ee ee 740
Ried renner eR TIES 5 siete Sea sud spe Ss ok 319 ake kas A ee a AS ONS 3 A cde 742
ERveronamotvisceral arehes..:. 3. ae. 4x. Se mete eta Oe ae ee ero eas 744
Nema meamem OMCs: haces curse co bela sO er OE ETO ROR eer Te 747
SULIT ATU.) iad os iz. Sitonar cl gis (aSewje lense Bele ee Ge ele RPE Ce ards Meee oe aliens nie: fe afer are 756
Bb liogmap hiya s sii s5 GS aeln tS. sdall Lite chennai Gh eteR a eleio eas atis eiepeia ave ec omeenne 761
INTRODUCTION

Although certain parts of the chondrocranium of the chelonia
have already been studied to some extent, the only previous work
which has aimed to furnish a picture of the skull and its develop-
ment in its entirety has been that of Parker (’80) on the devel-
opment of the skull of Chelone viridis. His results are of great
value, but on account of imperfect methods, many details of
structure and development escaped him and several errors occur.
A preliminary account of the present work has already appeared
(Kunkel 711).

693

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

694 B. W. KUNKEL

Of recent papers which deal with the chelonian chondrocra-
nium especial reference should be made to that of Ogushi (711) on
Trionyx japonicus in which the cartilaginous as well as the bony
elements of the adult skull are carefully described, and to the
extensive paper of Nick (’12) on Dermochelys coriacea.! Neither
of these deals with the embryonic condition of the skull. Several
papers on the development of special regions may be referred to
briefly. Seydel’s (96) conclusions regarding the development
of the nasal capsules are confirmed, but several new facts which
throw light on the significance of some of the characters of this
region are set forth here for the first time. Noack’s (’07) conclu-
sion that the columella arises solely from the otic capsule is
faulty in that my series shows the columella to be in reality made
up of a stapes which arises external to the capsule and an extra-
columella as has been described by Fuchs (’07) and Bender (’11).
My own results accord with those of these investigators regard-
ing the independence of columella and otic capsule. As to the
significance of thé stapes and extracolumella my results confirm
Bender’s as opposed to Fuchs’s. The latter’s conclusions regard-
ing the development of the visceral skeleton I am able to confirm.
Filatoff (06) has called attention to the presence of a processus
ascendens of the palatoquadratum in Emys.

Besides these papers may be mentioned also Filatoff’s (07) on
the metamerism of the head of Emys, Fuchs’s (’07) on the devel-
opment of the roof of the mouth, and Thater’s (10) on the same
subject. Hitherto a complete account of the embryonic chondro-
cranium of the turtle and its development has been lacking so that
comparisons have been possible only on the insecure basis of the
adult condition.

In order to obtain a clear representation of the form relation-
ships of the embryonic skull, a model was made by the Born wax

1In addition to these, should be mentioned a preliminary report by Fuchs
(Ueber einige Ergebnisse meiner Untersuchungen iiber die Entwickelung des
Kopfskelettes von Chelone imbricata. Verhandl. anat. Gesellsch., 26. Versamml.
1912, pp. 81-106.) Unfortunately further reference to it is impossible at this
time. Of special importance is his demonstration of the derivation of the tropi-
basic skull from the platybasic type.

DEVELOPMENT OF THE SKULL OF EMYS 695

plate method of an embryo of Emys lutaria in which the carapace
measured 11 mm. in length. The model represents the cartilagi-
nous skull magnified fifty times with the membrane bones of the
left side removed. At this stage the cartilage is well developed
and ossification has not proceeded far enough to alter materially
the original form of the chondrocranium. The membrane bones are
fairly well developed although the parasphenoideum and quad-
ratojugale as well as the complementare of the lower jaw appear
first in older embryos than that modelled. The present study
undertakes to determine the course of development of the chelo-
nian skull from its early prechondral stage. <A series of embryos
and young with carapace lengths ranging from 4.7 mm. to 28
mm. forms the basis of the work.

It gives me pleasure to make the following acknowledgments:
to Herr Professor Gaupp for the suggestion of the problem in the
first place as well as for his constant valuable criticism and sug-
gestions, and for the generous loan of many series of sections of
Lacerta and Chelone embryos and of many papers from his pri-
vate library; to Herr Professor Keibel who furnished me with a
large number of embryos of Emys lutaria which were collected by
Mehnert; and to Herr Geheime Rat Wiedersheim for the courtesies
and facilities of the Anatomisches Institut of Freiburg 1. B.
where most of the work was done during the college year of 1910-
11. To my father I am under obligation for material aid without
which the present investigation could hardly have been under-
taken.

GENERAL FORM OF THE SKULL

The cartilaginous skull of Emys lutaria resembles rather closely
in its essential features that of Lacerta which has been so fully
described by means of models by Gaupp (00). The Chelonian
skull is less elongated than that of Lacerta, especially in the ante-
rior half, as is shown by the relatively anterior position of the
fenestra hypophyseos. This is due principally to the fact that the
septum interorbitale of the turtle is much shorter than in the
lizards and the olfactory capsule, as will be shown later, is bent
ventrally so that it comes to lie to a greater extent ventral rather
than anterior to the orbital portion of the skull. Besides being

696 B. W. KUNKEL

of more compact form on account of the shortening of the space
between the olfactory and otic capsules, the numerous fenestrae
of the lizard skull are reduced both in number and size. At the
same time also, the slender rods of the lizard’s skull are to a great
extent replaced by broad, more or less continuous, plates. In this
respect the chelonian chondrocranium resembles more nearly
that of Sphenodon as described by Schauinsland (00) and Howes
and Swinnerton (01) and that of the crocodile described by
Parker (’83) than it does that of the lizards (Gaupp ’00) and
snakes (Parker ’78). The greater strength of the jaws of the adult
chelonian and crocodilian seems to be early foreshadowed in the
embryo by the greater solidity of the skull, especially in the por-
tions more intimately associated with the jaws. Notwithstand-
ing, however, this greater solidity of the chelonian chondrocra-
nium, in one respect it seems to be weaker in that it lacks the taenia
marginalis which couples the orbital region with the otic capsule
dorsally. In Emys the orbital and temporal regions are discon-
tinuous except for the trabeculae which lie near the mid-ventral
line, and the temporal and otic regions are discontinuous except
ventrally because of the absence of the taenia marginalis. More
detailed comparisons of the reptilian chondrocrania will be made
below, under the respective parts of the skull.

PLANUM BASALE AND CHORDA DORSALIS

The basal plate forms the entire floor of the parachordal portion
of the chondrocranium (fig. 25) extending forward from the occi-
pital condyle through the occipital and otic regions. It is simple
and continuous as far forward as the anterior half of the otic
capsule where the hexagonal fenestra basicranialis posterior is
situated. In front of this space the floor of the skull is rep-
resented by a heavy, transverse bar, the crista sellaris, which
forms the hinder boundary of the fenestra hypophyseos (f.h.).

There can be but little doubt that the part of the skull here
referred to as basal plate is made up in part, in the otic region,
from the floor of the otic capsule. In a»much younger stage
(carapace length, 4.7 mm.) the boundary between the blastema
of the basal plate and that of the otic capsule is quite distinct.

DEVELOPMENT OF THE SKULL OF EMYS 697

From this embryo it is seen that the blastema of the otic capsule
extends medially slightly beyond the median wall of the mem-
branous labyrinth, so that the basal plate proper is greatly reduced
in width, and its lateral margins are slightly concave antero-pos-
teriorly because of the position of the floor of the otic capsule.
The fusion between the floor of the otic capsule and basal plate
is complete in the stage modelled so that no line of demarcation
is visible.

In general form the basal plate in the stage modelled may be
regarded as hexagonal, exhibiting short anterior and posterior
sides extending transversely, and antero-lateral and postero-
lateral sides. Extending along the mid-ventral line from in front
of the occipital condyle to the fenestra basicranialis posterior is a
low, rounded crest formed by the chorda dorsalis which in Emys
lies in the same plane with and enclosed by the basal plate and at
this stage is so large as to cause the latter to bulge both dorsally
and ventrally in the regions where the basal plate is quite thin.
Immediately behind the fenestra basicranialis posterior, the basal
plate on each side of the middle line exhibits a gentle convexity
on its ventral side which is caused by the extension of the pars
cochlearis as will be described later. The postero-lateral and
antero-lateral margins of the basal plate do not pass directly over
into the lateral portions of the occipital and otic regions, as in
Lacerta, but show a tendency to project freely laterally in the
form of crests. The postero-lateral margin projects laterally
and slightly ventrally as the crista inferior (c.7., fig. 1), whose
posterior end terminates freely at the side of the base of the con-
dylus occipitalis. This crest lies ventral to the anterior end of the
fissura metotica and foramina spino-occipitalia which thereby
come to open into a groove between the crista inferior and the
lateral portion of the occipital region. This groove, lateral to
which the ganglia vagi and spino-occipitalia lie, may be called the
‘suleus supracristularis’ (s.s., fig. 1). Laterally the margins of
the basal plate project beyond the. otic capsule, except in the
posterior portion of the latter, to form a crista substapedialis
(c.st., fig. 4), a horizontal shelf which is continuous anteriorly
with the crista basipterygoidea and on which the foot plate of

698 B. W. KUNKEL

the columella auris rests ventrally, and under which the ramus
communicans n. facialis cum glossopharyngeo lies. The antero-
lateral margin of the basal plate projects as a long ridge which
passes directly into the trabecula cranii in front. This is the
crista basipterygoidea (pr.b., fig. 7) and represents the processus
basipterygoideus of Lacerta, as will be shown later. It is inclined
slightly ventrally so that a shallow groove is formed along the
margin of the ventral surface of the basal plate in which lie the
ramus palatinus n. facialis and arteria carotis interna (fig. 7).
Accordingly it corresponds to the sulcus cavernosus of the adult
pterygoideum (s.c., fig. 25).

The dorsal surface of the basal plate exhibits a longitudinal
ridge in the middle line which becomes less conspicuous poste-
riorly, disappearing completely in the condylar region. It is pro-
duced by the chorda dorsalis whose diameter is greater than the
vertical thickness of the basal plate along the middle line. The
thickness of the basal plate varies, decreasing markedly from the
posterior end.

The fenestra basicranialis posterior (f.b.p., fig. 25) is of a broad,
hexagonal form with anterior and posterior margins extending
transversely and with lateral angles. It lies in the anterior half
of the basal plate between the anterior portions of the two otic
capsules, with its anterior margin formed by the crista sellaris.
At the stage modelled it is closed by a membrane of dense con-
nective tissue beneath which the chorda dorsalis extends.

The crista sellaris (c.s., fig. 25) forms a heavy transverse bar
between the fenestra basicranialis posterior behind and the fen-
estra hypophyseos in front. In the mid-ventral line it exhibits
a pronounced longitudinal ridge, separating two longitudinal
grooves, while its lateral margin projects freely ventro-laterally
continuous with the crista basipterygoidea. Within the groove
thus formed, between the median crest of the crista sellaris and
the projecting lateral crests, lie ramus palatinus n. facialis and
arteria carotis interna.

Between the crista inferior and the pars lateralis of the occipital
region extends the sulcus supracristularis, a wide groove facing
laterally. In front it is limited by the posterior wall of the otic

DEVELOPMENT OF THE SKULL OF EMYS> 699

capsule. Along the dorso-median wall of the sulcus open the three
foramina spino-occipitalia and at its anterior end opens the ductus
perilymphaticus. The ganglia vagi and spino-occipitalia are situ-
ated lateral to the groove.

The foramina spino-occipitalia and facialis lie rather in the
lateral portion of the cranium than in the basal plate, but they
may be conveniently described at this time. There is some
variation in the number and relations of the foramina spino-occi-
pitalia in the series of embryos studied, although they always lie
in two horizontal lines, which converge in front, and are ventral
to the bases of the arcus occipitales and fissura metotica. They
pierce the basal plate in a ventro-lateral direction and open exte-
riorly in the sulcus supracristularis. In the individual modelled
there are three pairs of foramina, of which the anterior is the
smallest and the posterior the largest. The middle foramen is
equally distant from the other two. In the adult, as is well
known, the first and second pairs of nerves leave the cranial cavity
through the same foramen, and only two pairs of foramina spino-
occipitalia are present. In two younger embryos having a cara-
pace length of from 7 to 8 mm. the adult condition was met with.
In one embryo, older than that modelled, the first and second
foramina were united at their external ends but completely sepa-
rated from each other internally. On the other hand in one of the
oldest individuals studied all three pairs of feramina were quite
distinct as in the model.

The foramen facialis (f.f., fig. 6) lies well in front of the anterior
end of the otic capsule and relatively far laterally in the side por-
tion of the otic region. Behind this foramen the basal plate is
continuous with the otic capsule for a long space. In front of
it there is only a slender commissure uniting the basal plate and
the cupula anterior of the capsule.

The foramen abducentis (f.a., figs. 8 and 24) passes hori-
zontally forward at the anterior end of the crista sellaris, as a
moderately long canal opening anteriorly ventral to the pila
prootica and dorsal to the crista basipterygoidea, and poste-
riorly on the dorsal surface of the basal plate immediately behind
the proximal end of the pila prootica.

700 : B. W. KUNKEL

The basal plate is separated from the cupula posterior of the
otic capsule by the triangular, ventral portion of the fissura
metotica; and from the cupula anterior by the fenestra prootica.
Between these two foramina the basal plate is continuous with
the capsule except for the foramen facialis which is so situated as
to leave a slender commissure uniting the basal plate and otic
capsule between itself and the fenestra prootica.

The relations of the chorda dorsalis to the basal plate are of
considerable interest. In the embryo modelled the chorda passes
without interruption from the anterior face of the dens epistro-
phei into the posterior end of the condylus in which region it is
completely imbedded in cartilage. It is surrounded directly by
a sheath which is continuous with the perichondrium of the basal
plate in front and from which, in the region between the dens and
condylus, the ligamentum apicis dentis is derived.

In front of the condyle, where the basal plate becomes gradu-
ally thinner, as far forward as the fenestra basicranialis posterior,
the chorda lies in the same plane with the basal plate and tends
to divide it into two symmetrical halves. The basal plate, accord-
ingly, is parachordal, as has been already described in the skull of
snakes, crocodilians, and Sphenodon; and not hypochordal as in
Lacerta. This condition, however, is probably secondary, since
in younger individuals in which the condyle is not yet developed,
the basal plate immediately in front of the condylar region is
hypochordal to within a short distance of the fenestra basicranialis
posterior where it becomes parachordal.

The tissue immediately dorsal and ventral to the chorda in the
region posterior to the fenestra basicranialis posterior is reduced
to a very thin layer and is not fully chondrified in the stage mod-
elled. In passing through the fenestra basicranialis posterior the
chorda is enclosed in a membranous sheath which is continuous
with that which fills the fenestra, although it comes to lie in a
plane ventral to the basal plate.

The chorda dorsalis is lodged posteriorly on the dorsal surface of
the crista sellaris in a deep groove whose sides gradually close
together anteriorly, converting the groove into a canal which

DEVELOPMENT OF THE SKULL OF EMYS 701

extends forward through the crista and opens on its anterior
surface into the fenestra hypophyseos.

In passing through the crista sellaris the cherda tapers from the
uniform diameter which it exhibits posteriorly to a rounded point.

As might be expected in a part undergoing rapid degeneration,
the anterior end of the chorda exhibits considerable variations in
its relations to the crista sellaris. In several specimens it was
bent in a dorso-ventral direction within the crista so that the
anterior portion lay in a plane dorsal but parallel to the posterior
portion. In most of the embryos studied the crista was com-
pletely perforated by the canal for the chorda, but in one young
embryo the canal terminated within the cartilage. In another
embryo only slightly younger than that modelled, the anterior
end of the chorda projected freely into the fenestra hypophyseos
while in the model, as in most of the specimens studied, the end
of the chorda was flush with the anterior surface of the crista.

At the extreme caudal end of the condylus in the embryo
modelled, the chorda is completely surrounded by cartilage, but
in a very slightly younger embryo the condylus at its extreme
caudal end exhibits a U-shape in cross section quite similar to
that of Lacerta; that is, the chorda was surrounded only ventrally
and on the two sides with cartilage (fig. 15). Later, however,
the chorda is completely surrounded. Immediately surrounding
the chorda the cartilage becomes excavated to form a cup-like
depression, the central cavity, into which the dens epistrophei
fits. The chorda passes from the anterior face of the epistropheus
into the posterior face of the condyle, forming a condylus anularis
characteristic of the Chelonia. The ventral surface of the con-
dyle articulates with the atlas which projects ventrally to it.

In contrast with the condyle of Lacerta, as Gaupp has shown,
that of Emys does not exhibit the two processes, one on each side
of the chorda dorsalis, but rather a single ring-like process around
the chorda dorsalis. In view, then, of the embryonic condition
in Lacerta, the derivation of the mammalian condyles from those
of reptiles is not improbable. In Chelone, according to Gaupp,
the chorda enters the condyle from the ventral side so that,

702 B. W. KUNKEL

even in the adult of this form, the central cavity of the condyle,
in which the ligamentum apicis dentis is inserted, lies near the
ventral margin of the posterior surface of the condyle.

REGIO OCCIPITALIS

The occipital region, like that of Lacerta, may be differentiated
into » basal and two lateral parts. A dorsal region is lacking,
although the tectum posterius, which is continuous with the otic
capsules, and hence properly belongs to the otic region, projects
caudally with its strongly developed processus posterior and closes
in the foramen occipitale magnum dorsally.

The basal portion is represented by the basal plate including
the condylus and cristae inferiores (fig. 1), and the lateral parts
by the arcus occipitales, which are continuous ventrally with the
basal plate and extend freely dorsally, separated from the otic
capsules in front by the fissura metotica and not united distally
with the tectum posterius. Because of the separation distally of
the arcus occipitales from the tectum, it is evident that at this
stage the fissura metotica and foramen occipitale magnum are
not completely separated from one another.

The condylus occipitalis projects posteriorly as a cylindrical
process with convex ventral surface and flat, or even slightly con-
cave, dorsal surface so that a cross section is reniform. Its free
distal surface is flat except for a slight depression which extends
dorsally from the canal in which the chorda lies and gradually
deepens toward the dorsal surface of the condyle. The free end
of the condylus is embraced, except for a short space on the dorsal
side, by the atlas, which has the form of an incomplete ring and
projects forward ventral to the condyle as a stout process.

In later stages the ossification of the condylus is seen to proceed
from three centers, one ventral to the chorda and one on each side
of it; corresponding to the basioccipitale and two pleuroccipitalia
respectively.

The lateral portions of the occipital region are represented by
the arcus occipitales which arise from the dorsal surface of the
basal plate in the region of the foramina spino-occipitalia. The
arcus occipitales are stout, curved, sightly tapering prismatic

DEVELOPMENT OF THE SKULL OF EMYS 703

rods which recall strikingly the neural arches of the vertebrae.
Their medial margins are curved regularly to enclose the foramen
occipitale magnum laterally; their external margins, however,
exhibit a more angular contour because of a marked thickening
midway between their base and apex against which the otic cap-
sule rests posteriorly. The fissura metotica is accordingly greatly
narrowed here but not completely obliterated. The arcus occipi-
tales bound the foramen occipitale magnum laterally and the
fissure, metotica posteriorly and medially.

The foramen occipitale magnum is large and of hexagonal form,
with its dorsal and ventral margins transverse and with lateral
angles. Its plane is vertical and transverse. The condyle and
basal plate form its ventral margin, the curving arcus occipitales
its lateral margins, and the free posterior margin of the processus
posterior of the tectum posterius its dorsal margin.

REGIO OTICA

In contrast to the occipital the otic region is complicated. The
basal plate forms the floor, the otic capsules the lateral walls and
the tectum posterius the dorsal portion. Of the basal plate
there should be mentioned the large, hexagonal fenestra basi-
cranialis posterior in the anterior part of the otic region and
bounded anteriorly by the crista sellaris. The antero-lateral
margins of the basal plate are extended ventro-laterally beyond
the connection with the otic capsule, and form the posterior end of
the crista pterygoidea. The lateral margin of the basal plate
extends laterally, beyond the anterior and middle thirds of the
capsule, as the crista substapedialis (c.st., fig. 4). Posterior to
the level of the fenestra vestibuli the lateral margin of the basal
plate passes into the lateral capsular wall; in front of this fenestra,
however, the lateral extension of the basal plate becomes gradually
more pronounced.

The basal plate and otic capsule are in connection with each
other for a considerable space, extending from the fenestra pro-
otica in front to the foramen jugulare behind and interrupted
only by the foramen facialis which is well in front of the cochlear
portion of the otic capsule and separated from the fenestra pro-

704 B. W. KUNKEL

otica by a short and comparatively slender rod, the commissura
praefacialis which extends dorsally and laterally from the basal
plate to the ventral aspect of the anterior cupula of the otic cap-
sule. Behind the foramen facialis the median wall of the cochlea
passes continuously into the basal plate along a curved line which
is concave on the lateral side.

The prefacial commissure, together with the anterior cupula
of the capsule, forms the posterior margin of the fenestra pro-
otica in which is located, immediately below the capsule, the gan-
glion semilunare of the nervus trigeminus (g.s., fig. 7). Unlike
the condition in Lacerta, the ganglia of the three rami are closely
united so that they appear as a single ganglion wholly within the
fissure, except the anterior extremity which lies external and
ventral to the pila prootica; whereas in Lacerta the ganglion of
the ramus ophthalmicus lies widely separated from the others,
quite far in front of the slender pila prootica.

The fissura metotica (f.m.) is a narrow slit, widening ventrally
to a large triangular space, situated between the otic capsule on
the one hand and the basal plate and arcus occipitalis on the
other. The otic capsule forms the anterior and lateral boundary
of the fissure, and the arcus occipitalis, the posterior and medial
one. In its narrow dorsal portion the capsule comes to lie some-
what external to the occipital arch so that the fissure opens nearly
transversely, while its extensive ventral portion faces laterally.
Dorsally the fissure is continuous with the foramen occipitale
magnum as already described, although its dorsal portion is
closed by a mass of connective tissue,

In the expanded ventral portion of the fissure are situated the
nervus vagus, vena jugularis, and ductus perilymphaticus in their
passage to the exterior of the skull. The ductus perilymphaticus
occupies the extreme ventral corner of the fissure. Dorsal to the
ductus perilymphaticus, but still within the triangular expansion
of the fissure at its ventral end, are situated the jugular vein and
vagus nerve which in the adult pass out of the cranium through
the foramen jugulare externum (f.m., fig. 1).

Beneath the canalis perilymphaticus on the left side of figure
2 there may be seen a narrow groove which becomes wider in a pos-

DEVELOPMENT OF THE SKULL OF EMYS 705

terior direction from the section figured and which becomes con-
verted into a very fine tube anteriorly by the growing together
of the walls of the groove dorsally. In the embryo modelled the
tube extends through only two or three sections and ends blindly
posterior to the cavum cochleae. It apparently is the canalis
hypoperilymphaticus which Nick (12) has described in Dermo-
chelys, Chelydra, and Chelone; like that of the two latter it con-
tains no blood vessels but only a very loose connective tissue.
In an older embryo (carapace length of 13.5 mm.) this canal has
increased considerably in size so that it is one-third as large in
diameter as the canalis perilymphaticus and it communicates
anteriorly with the cavum cochleae at the most ventral and
posterior portion of the latter. While the canalis hypoperilym-
phaticus apparently is not differentiated at a much earlier stage
than that modelled, in an individual recently hatched of 28 mm.
in carapace length, the canal was large.

The posterior wall of the cochlea, which is perforated by the
large, oval fenestra cochleae, bounds the fissura metotica in front.
This fenestra opens anteriorly into the cavum cochleae along the
median wall of the capsule and posteriorly into the extreme ante-
rior end of the sulcus supracristularis. The anterior end of the
sulcus supracristularis corresponds quite closely with the recessus
scalae tympani, as described by Gaupp in Lacerta. This region
communicates with the otic capsule by means of the fenestra
cochleae, with the cranial cavity by the extreme anterior end of
the fissura metotica, and opens widely throughout its extent to
the exterior of the cranium. These communications represent
respectively the fenestra cochleae (s. rotunda), apertura medialis
recessus scalae tympani, and apertura lateralis recessus scalae
tympani.

The principal differences in the relationships of these three
openings in Lacerta and Emys lie in the fact that the plane of the
fenestra cochleae in the latter is vertical, while that of Lacerta
is horizontal, and also in that the front margin of the apertura
lateralis recessus scalae tympani lies slightly behind that of the
apertura medialisin Emys. These slight differences are explicable
by the fact that the cochlea in Emys has developed in a pos-

706 B. W. KUNKEL

terior direction, as is also indicated by the relation of the nervus
glossopharyngeus to the otic capsule. This change in relations
may be imagined to have taken place by supposing that the
posterior extension of the floor of the cochlea has occurred prin-
cipally in front of the fenestra cochleae and has been more rapid
on the antero-ventral aspect of the sac than on the postero-dorsal.
In this event it is apparent that a foramen, for example, which
originally lay in the floor of the capsule will be rotated to occupy
the posterior wall. In the same way also, should the growth in
a posterior direction be more rapid laterally than medially where
the proximity to the basal plate would retard the extension,
the external aperture of the recess would come to be somewhat
posterior to the internal aperture.

CAPSULA OTICA

In general form the otic capsule may be compared to a trian-
gular prism with one long edge situated ventrally and with the
broadest face vertical and medial; accordingly there are an
anterior and a posterior base and also a latero-dorsal and latero-
ventral face, as well as a median one. The two prisms lie with
their axes horizontal and their median faces diverging only
slightly from ventral to dorsal and from posterior to anterior.
In contrast to the condition in Lacerta the lateral walls of the
cranium are more fully represented by the otic capsules and the
brain is confined more completely by them without exhibiting a
tendency to extend laterally dorsal to the capsules. On account
of the unusual extension of the sacculus and lagena in a ventral
direction the height of the capsule is almost as great as the great-
est length of the same. The two bases of the prismatic capsule
project in its long axis as a cupula anterior and posterior, the
latter of which is the more pronounced. The cupula anterior
bounds the dorsal portion of the fenestra prootica posteriorly
and projects freely without connection with other parts of the
skull. The cupula posterior bounds the fissura metotica ante-
riorly and laterally, extending caudally to lie external to the occi-
pital arch as already described. The two capsules are united
dorsally by the tectum posterius whose slender lateral rods arise

DEVELOPMENT OF THE SKULL OF EMYS LOE

gradually from them as triangular plates. The comparatively
simple prismatic form is greatly modified by a number of prom-
inences which for the most part follow the underlying parts of
the membranous labyrinth though not as closely as in Lacerta
because of the greater thickness of the capsular walls.

The capsule may be differentiated into a dorsally situated ves-
_tibular portion and a smaller, ventrally situated cochlear portion
which encroaches upon the basal plate, as already described.
The vestibular portion exhibits prominences corresponding to
the semicircular canals, with their ampullae, and the utriculus;
the cochlear portion which remains simpler in its external form,
exhibits a flattened, oval, pocket-like form.

The prominentia semicircularis anterior (fig. 28) forms the
dorsal and antero-dorsal margins of the otic capsule. Its ventral
end, situated in the cupula anterior, widens to accommodate the
ampulla anterior which lies in a somewhat oblique position so
that it does not produce a marked convexity in the capsular wall.
The plane of the anterior semicircular canal inclines medially
from ventral to dorsal and from anterior to posterior. ‘The prom-
inentia semicircularis posterior is continuous in front with the
anterior prominence and curves along the dorsal and posterior
margins of the capsule to the cupula posterior. The dorso-
median edge of this prominence is continuous for its middle third
with the tectum posterius. The plane of the canal inclines
medially from below and in front, so that its ventral end pro-
jects laterally much as does that of the anterior canal. The
prominentia ampullaris posterior forms a marked convexity on
the lateral wall of the capsule below the ventral end of the pos-
terior semicircular canal and dorsal to the crista parotica (cr.p.,
fig. 28); the prominentia ampullaris posterior accordingly forms
the postero-ventral margin of the capsule and bounds the ante-
rior end of the fissura metotica. On the median wall it forms a
continuous area with the gently bulging prominentia utricularis
which forms an area continuous in front with the prominentia
sinus superioris utriculi, ventrally with that of the sacculus, pos-
teriorly with the prominentia ampullaris posterior, and dorsally
with the prominentia semicircularis posterior. On the median

oP

708 B. W. KUNKEL

capsular wall the prominences of the anterior and posterior semi-
circular canals unite with the dorsal end of the prominentia sinus
superioris utriculi.

The prominentia semicircularis lateralis with that of its am-
pulla forms a horizontal ridge which is situated on the lateral
surface of the capsule between the ventral ends of the anterior
and posterior canals, and which broadens anteriorly to accom-
modate the ampulla. A decided triangular depression marks this
ridge off dorsally from the other two canals. Ventrally it passes
gradually into the prominentia saccularis except at its anterior
end where the prominentia recessus utriculi is situated.

The prominentia recessus utriculi on the lateral aspect of the
otic capsule occupies a circular space ventral to the combined
prominences of the ampullae lateralis and anterior, dorsal to the
foramen facialis, and antero-dorsal to the fenestra vestibuli.
On the median face of the capsule it forms a gentle triangular
convexity with its posterior angle situated immediately dorsal
to the foramen acusticum posterius; its antero-ventral angle
dorsal to the foramen facialis and anterior to the foramen acus-
ticum anterius. ‘The prominentia recessus utriculi is separated
above from that of the anterior semicircular canal by the shallow
fossa subarcuata; it passes imperceptibly into the prominentia
ampullaris anterior in front, and is bounded ventrally by a line
joining the foramina facialis and acusticum posterius; it is con-
tinued posteriorly between the latter and the foramen endolym-
phaticum and forms a marked convexity for the ductus endolym-
phaticus immediately dorsal to its opening into the sacculus, and
ventral to the foramen endolymphaticum.

The prominentia sinus superioris utriculi extends obliquely
antero-dorsally from the foramen endolymphaticum to unite
with the dorsal ends of the prominences of the two vertical semi-
circular canals. In front it forms the posterior boundary of the
fossa subarcuata, posteriorly it is continuous with the combined
prominentiae utricularis and ampullaris posterior.

The prominentia saccularis occupies the entire ventral third
of the capsule. In contrast to the condition met with in most of
the vertebrates, the cochlea of the chelonians extends posteriorly

DEVELOPMENT OF THE SKULL OF EMYS 709

instead of anteriorly so that the unique condition is met with in
that the nervus glossopharyngeus passes through the cavity of
the cochlea in its passage from the cranium to the exterior, as
will be described more fully later. On its lateral aspect the prom-
inentia saccularis is semicircular with its convexity below. It
is situated behind the foramen facialis and the prominerftia
recessus utriculi and beneath the prominentia semicircularis
lateralis, and passes posteriorly beyond the foramen glosso-
pharyngei externum. Its ventral wall is the basal plate. As
already mentioned the prominentia saccularis is apparent from
the ventral side of the skull as a gentle convexity of the basal
plate which lies behind the level of the fenestra basicranialis pos-
terior. On the median aspect of the otic capsule the prominence
extends posteriorly from the region of the foramen acusticum
anterius posteriorly as far as the anterior angle of the fissura
metotica. The convexity of the prominentia saccularis medially
is more pronounced in its anterior portion immediately behind
the slender side piece through which the n. facialis passes than
posteriorly so that the foramen acusticum anterius, which is
situated in this region, comes to face somewhat anteriorly as
well as medially.

The crista parotica forms a stout prismatic structure with its
ventral surface horizontal and extending forward from the ven-
tral limits of the prominentia ampullaris posterior and projecting
freely for a very short distance in front.

The fenestra vestibuli (s. ovalis) (f.v., fig. 28) is situated in the
ventral part of the lateral wall of the otic capsule, in the middle
of the saccular region between the foramen glossopharyngei
externum and foramen facialis, slightly nearer the former than
the latter. It is of equilateral triangular form to accommodate the
foot plate of the columella auris. Its ventral margin is hori-
zontal and is formed by the basal plate which projects laterally
beyond the capsule as a kind of horizontal shelf on which the foot
plate of the columella rests (fig. 5).

Behind the fenestra vestibuli and separated from it by a
slender rod sloping obliquely postero-ventrally, is the external
opening for the n. glossopharyngeus (f.g.e., fig. 28). This is

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

710 B. W. KUNKEL
situated ventral to the prominentia ampullaris posterior and at
the postero-dorsal angle of the prominentia saccularis.

On the median wall of the otic capsule are five or six foramina
arranged in two horizontal rows. In the ventral row are the
foramina acusticum anterius and posterius and the median open-
ing for the glossopharyngeal nerve. The foramen acusticum
anterius is small and is situated above and behind the foramen
facialis and ventral to the prominentia recessus utriculi. Close
behind the anterior foramen is the much larger, oval foramen
acusticum posterius which is situated in the antero-dorsal por-
tion of the prominentia saccularis. Posterior to and separated
by a considerable space from the last is the small foramen glosso-
pharyngei internum. It lies close in front of the fissura meto-
tica and at the upper margin of the prominentia saccularis and
below the prominentia ampullaris posterior. In the upper row
of foramina are the foramen endolymphaticum, a foramen for a
small blood vessel and a third foramen which may or may not be
present and which has apparently no significance, being simply
an unchondrified area. The foramen endolymphaticum is situ-
ated postero-dorsally from the foramen acusticum posterius and
in front of the broad prominentia sinus superioris utriculi. The
small foramen for a blood vessel lies in front of the foramen endo-
lymphaticum; and on the left side only in the individual modelled
between these two foramina is the insignificant unchondrified area
already mentioned.

Viewed from the lateral side the otic capsule (fig. 28) exhibits
the following topography. The anterior margin of the cupula
anterior projects slightly laterally as a rounded ridge, the prom-
inentia semicircularis anterior. Below the ventral end of this, in
the space above the foramen facialis, is a circular convexity,
the prominentia recessus utriculi, which extends behind as far
as the fenestra vestibuli. Extending horizontally from the ven-
tral end of the prominentia semicircularis anterior to the ventral
end of the prominentia semicircularis posterior is the conspicuous
cylindrical ridge comprising the prominentiae ampullaris lateralis
and semicircularis lateralis; ventrally this prominence merges
without a sharp boundary into the prominentia saccularis,

DEVELOPMENT OF THE SKULL OF EMYS 711

dorsally it is separated from the prominences of the vertical
semicircular canals by a deep depression. The prominentia sac-
cularis is interrupted by the large fenestra vestibuli as well as by
the foramen glossopharyngei externum at its extreme upper
and posterior angle. The prominentia semicircularis posterior
rounds off the capsule postero-dorsally and rests ventrally on the
prominentia ampullaris posterior which is separated by a slight
depression from that of the semicircular canal.

The median aspect of the capsule has a much more uniform
surface than the lateral one. The prominences of the two verti-
eal semicircular canals make a less conspicuous ridge framing
the capsule above than they do on the lateral side. Ventral to
the prominentia semicircularis anterior is the conspicuous depres-
sion, the fossa subarcuata, which separates it from the prom-
inentia recessus utriculi. Posterior to the ventral end of the
prominentia semicircularis anterior and ventral to the fossa
subarcuata is the inconspicuous triangular prominentia recessus
utriculi which is dorsal to the foramen facialis and antero-dorsal
from the foramen acusticum anterius. Its posterior angle
extends dorsal to the foramen acusticum posterius and ventral
to the foramen endolymphaticum. Posterior to the foramen
endolymphaticum the prominentia sinus superioris utriculi
extends diagonally forward and upward and continues at its
upper end with the prominences of the two vertical semicircular
canals; posteriorly it merges imperceptibly into the prominentia
utricularis and ampullaris posterior. Ventral to the line connect-
ing the dorsal margins of the two foramina acustica and glosso-
pharyngei internum is the semicircular prominentia saccularis
which bulges suddenly immediately behind the foramen facialis
and extends backward to the fissura metotica.

The posterior wall of the cochlea is perforated near its median
margin by the fenestra cochleae, the anterior end of the canalis
perilymphaticus, as already described. The relation of these
parts is complicated in the stage modelled on account of the
extension posteriorly of the lateral wall of the cochlea so that the
canalis perilymphaticus opens posteriorly into a funnel-like
extension of the anterior end of the sulcus supracristularis.

712 B. W. KUNKEL

In older embryos, as already described, in addition to the canalis
perilymphaticus is also the canalis hypoperilymphaticus opening
into the cochlear cavity from the sulcus supracristularis.

INTERIOR OF CAPSULA OTICA

The interior of the otic capsule (figs. 29 and 31) exhibits a com-
plicated form, but in contrast to that of Lacerta it is somewhat
simpler because of the absence of a septum intervestibulare and
the consequent differentiation of cava vestibuli anterius and pos-
terius. Besides this, the posterior development of the lagena
tends to make the interior more compact. It is possible to dis-
tinguish a cavum vestibuli with a cavum ampullare posterius,
laterale, and anterius, the three canales semicirculares, and
a cavum cochleae which communicates widely with the cavum
vestibuli. Of especial interest in comparison with Lacerta, in
addition to the absence of a septum intervestibulare and the
relation of the cavum cochleae, may be noted the sulcus in the
posterior wall of the cochlea in which the n. glossopharyngeus
is situated in its passage from the interior of the cranium to the
exterior. It is slightly oblique in its course, extending slightly
ventrally from the median to the lateral side on the posterior
side of the cavum cochleae immediately below the recessus ampul-
laris posterior.

The principal part of the total cavity of the capsule is repre-
sented by that of the vestibule. Medially the cavum vestibuli
extends to the median wall of the capsule, but the lateral wall
of the capsule forms only a part of its lateral boundary, the ver-
tically situated septum between the vestibulum and the canalis
semicircularis lateralis (septum semicirculare laterale) being
interposed in the space immediately in front of the recessus
ampullaris posterior and behind that of the lateral ampulla (s.s.l.,
fig. 4). The roof of the cavum vestibuli is incomplete because
of the large foramen pro sinu superiore utriculi. The roof is
formed by the septa semicircularia anterius and posterius which
extend horizontally outwardly from the median wall of the otic
capsule and separate the canalis semicircularis anterior and pos-
terior respectively from the cavum vestibuli. The septum semi-

DEVELOPMENT OF THE SKULL OF EMYS (ks

circulare posterius extends laterally in a horizontal direction to
the septum semicirculare laterale, the septum semicirculare an-
terius to the lateral wall of the capsule. The floor of the cavum
vestibuli is very limited in extent because of the wide connection
ventrally between it and the cavum cochleae which lies below it.
The whole cavity of the capsule including that of the lateral semi-
circular canal is somewhat flattened from side to side.

The cavum vestibuli may be differentiated for clearness in
description into a central portion with posterior, vertical, and ante-
rior limbs. ‘The posterior limb is situated ventral to the septum
semicirculare posterius and encloses the ampulla posterior and
the median limb of the canalis semicircularis lateralis as well
as the posterior portion of the utriculus (fig. 3, left side). The
vertical limb is situated between the anterior end of the septum
semicirculare posterius and the posterior end of the septum semi-
circulare anterius, and encloses the sinus superior utriculi (fig.
4). The anterior limb is situated below the septum semicir-
culare anterius, and encloses the ampullae anterior and lateralis
and the recessus utriculi (fig. 6, left side). The central portion
of the cavity is situated medial to the septum semicirculare later-
ale and encloses the principal portion of the utriculus and the
recessus utriculi.

The posterior limb of the cavum vestibuli is pear-shaped in a
vertical cross section, the narrow end being directed laterally and
dorsally to accommodate the posterior semicircular canal, while
the thickened end is medial and ventral for the accommodation of
the posterior ampulla. Posterior to the septum semicirculare
posterius, the orificium inferius canalis semicircularis posterioris
opens ventrally into the cavum vestibuli on its dorsal aspect.
Immediately ventral to this septum the canalis semicircularis
lateralis opens in a vertical longitudinal plane into the lateral
aspect of the posterior limb of the cavum vestibuli. Dorsally the
cavum vestibuli communicates with the two vertical semicir-
cular canals by means of a common orifice situated between the
septa semicircularia anterius and posterius and near the median
side of the cavity. Opening into the anterior limb of the cavum
vestibuli in a vertical plane facing obliquely forward and laterally

714 B. W. KUNKEL

is the orificium inferius canalis semicircularis anterioris. Behind
this orifice and below the septum semicirculare anterius are the
cavities of the lateral and anterior ampullae which are in the form
of lateral pouches from the cavum vestibuli, that of the lateral
ampulla being somewhat dorsal and posterior to that of the
anterior ampulla. Ventral to the cava ampullaria anterius and
lateralis is that of the recessus utriculi which forms a concavity
in the lateral wall of the cavum vestibuli, so that at this point
the cavum vestibuli exhibits its maximum width from side to
side (fig. 6). The cavum vestibuli has a small bulging on its
median side about midway between its posterior and anterior
ends and at a level with the ventral end of the sinus superior
utriculi for the accommodation of the ductus endolymphaticus as
it proceeds ventrally from its foramen to its opening into the sac-
culus (fig. 4).

The cavum ampullare posterius, occupying the postero-ventral
corner of the cavum vestibuli, is differentiated from the dorsally
situated canalis semicircularis lateralis by a medial thickening of
the lateral capsular wall on the one hand and a ventro-lateral
ridge on the septum semicirculare posterius on the other (fig.
2); anteriorly the cavity of the posterior ampulla opens into the
cavum vestibuli; and dorsally, into the posterior semicircular
canal.

The cavum ampullare laterale lies lateral to the anterior end
of the cavum vestibuli and opens medially into it by a wide
mouth, anteriorly it opens into the cavum ampullare anterius
and ventrally into that of the recessus utriculi; posteriorly it
communicates with the canalis semicircularis lateralis through
the wide orificium anterius canalis semicircularis lateralis. Its
antero-dorsal boundary is formed by the septum semicirculare
anterius, its lateral wall is that of the capsule itself; and medially
it opens by a very wide aperture into the cavum recessus utriculi
and the sinus superior utriculi. At its extreme dorsal end the
cavum ampullare laterale is bounded. medially by the posterior
end of the septum semicirculare anterius.

The cavum ampullare anterius forms a kind of antero-lateral
pocket of the cavum vestibuli, being continuous with the anterior

DEVELOPMENT OF THE SKULL OF EMYS 715

end of the cavum recessus utriculi. The canalis semicircularis
anterior opens into it from above by the orifictum inferius canalis
semicircularis anterioris. The cavum ampullare anterius is differ-
entiated from that of the recessus utriculi which lies behind it
by a broad vertical ridge on the lateral capsular wall. .

The cavum utriculi, which occupies the middle portion of the
cavum vestibuli, is continuous dorsally with that of the sinus
superior utriculi, anteriorly with the recessus utriculi and ampulla
lateralis, posteriorly with the ampulla posterior and canalis semi-
circularis lateralis, and ventrally with the cavum sacculi. Its
median wall is formed throughout by the median capsular wall;
laterally it is bounded by the septum semicirculare laterale and,
ventral to the septum, by the lateral wall of the capsule which
is continuous with it. Anterior to the septum the cavity opens
laterally into the lateral ampulla and the recessus utriculi, and
posterior to the septum it is continuous with the median limb
of the lateral semicircular canal. At its anterior end the cavum
utriculi widens abruptly in a lateral direction to form the cavum
recessus utriculi, which lies rather ventral to the lateral ampulla.

The cavum sinus superioris utriculi extends dorsally from the
cavum utriculi between the septa semicircularia anterius and
posterius. Opening into it from the anterior and lateral aspect
are the cava recessus utriculi and ampullare laterale. Dorsally
it opens into the orificia superioria canalis semicireularis anterioris
and posterioris. Medially and laterally it is limited by the respec-
tive capsular walls.

The cavum recessus utriculi forms a pocket extending ven-
trally from the cavum ampullare laterale and anterior to the
cavum utriculi. Anteriorly it opens into the cavum ampullare
anterius and ventrally into that of the sacculus, but at its extreme
anterior end the ventral capsular wall limits it.

The canalis semicircularis posterior curves laterally and ven-
trally from its union with the sinus superior utriculi, in a poste-
rior direction to the posterior ampulla, into which it opens from
above by means of the orifictum inferius canalis semicircularis
posterioris. ‘The canalis semicircularis lateralis connects with the
posterior portion of the posterior semicircular canal by means of

716 B. W. KUNKEL

the orificium posterius canalis semicircularis lateralis which is
situated immediately ventral to the septum semicirculare poste-
rius and somewhat lateral to the posterior semicircular canal.

The canalis semicircularis anterior curves from its posterior
end, first in an antero-lateral direction, then ventrally, and at its
extreme antero-ventral portion, somewhat medially, so that its
entrance into the cavum ampullare anterius through the orificium
inferius canalis semicircularis anterioris is from above and later-
ally. It is bounded dorsally, medially, anteriorly, and in part
laterally by the corresponding walls of the capsule; ventrally it
is bounded by the septum semicirculare anterius which also lies
somewhat lateral to the canal at its posterior end. The antero-
lateral end of the canal is limited behind by a medial thickening
of the lateral capsular wall which separates the cavities of the
anterior and lateral ampullae.

The canalis semicircularis lateralis has two quite distinct
portions, a median limb and a lateral one. The cavity of the
median limb is confluent with that of the utriculus, the pos-
terior ampulla and the inferior end of the posterior canal. It lies
lateral and dorsal to the utriculus and posterior ampulla and
anterior to the posterior semicircular canal. The cavity of the
median limb is bounded laterally by the septum semicirculare lat-
erale and dorsally by the septum semicirculare posterius. The
lateral limb is separated, except at its posterior end, from the
median limb by the septum semicirculare laterale. The canal is
horizontal in position and is continuous in front with the lateral
ampulla through the large orifictum anterius canalis semicircu-
laris. Posteriorly the lateral limb of the lateral semicircular
canal is continuous with the median limb immediately behind
the septum semicirculare laterale and also with the ventral and
posterior end of the posterior semicircular canal.

The cavum sacculi (c.sac., fig. 29) exhibits several features of
great morphological interest on account of the relations of the
ductus perilymphaticus and the development of the ductus
cochlearis in a posterior direction. It has the general form of a
flattened pocket which is rather shallow in front and increases
quite regularly in depth posteriorly so that the posterior wall is

DEVELOPMENT OF THE SKULL OF EMYS | 717

high and vertical in position. Dorsally the cavum sacculi stands
in wide open connection with that of the vestibulum. Its boun-
daries are the capsular walls laterally, medially, posteriorly, and
anteriorly; and ventrally the basal plate which bulges ventrally
as already noted because of the great extension of the lagena in
a ventral direction. In its ventral portion the cavum sacculi
exhibits an antero-lateral and a postero-medial lobe which stand
in wide open communication with each other; the boundary
between the two extending from the median wall at the level of
the foramen acusticum posterius in a postero-lateral direction.
The antero-lateral lobe extends on the outer aspect posteriorly
beyond the fenestra vestibuli which interrupts its lateral wall; on
its median aspect it extends posteriorly only as far as the foramen
acusticum posterius. In it-is situated the sacculus. The postero-
medial lobe is somewhat deeper than the previous one so that its
ventral end projects slightly below the level of the lateral lobe.
It is somewhat triangular in form with the posterior side in a
vertical plane. This encloses the lagena, the ductus perilympha-
ticus and the n. glossopharyngeus in its course between the exter-
nal and internal foramina of that nerve. Accordingly this cavity
exhibits two small but very important extensions; namely, the
sulcus glossopharyngeus (s.g.) and the canalis perilymphaticus
(c.per.). ‘The sulcus glossopharyngeus is situated on the posterior
cochlear wall ventral to the cavum ampullare posterius and
extends transversely from the foramen glossopharyngei inter-
num in a lateral and ventral direction to the foramen glosso-
pharyngei externum. It is in the form of a cylindrical groove of
uniform size which opens into the cavum sacculi as already indi-
cated. Near its median extremity it intersects the ventral wall
of the cavum ampullare posterius so that it shows communica-
tion with this part as well. It is completely filled by the nerve
which thus comes to lie with its anterior half projecting freely
into the cavum sacculi. The canalis perilymphaticus extends
in an anterior direction from the fenestra cochleae horizontally
through the posterior cochlear wall. It opens into the sacculus
at a point half way between the dorsal and ventral ends of that
part and somewhat medial to the middle plane of the cavity. .

718 B. W. KUNKEL

The canal opens at its anterior end in an oblique plane along the
median wall of the cochlea a short distance in front of its extreme
posterior end (fig. 3, left side). The ductus perilymphaticus
fills the canal completely.

The septum semicirculare posterius is in the form of a horizon-
tal plate which extends laterally from the median to the lateral
capsular wall and septum semicirculare laterale. It passes dor-
sal to the cavum ampullare posterius and ventral to the canalis
semicircularis posterior as has been already noted. The sep-
tum begins at its posterior end as a ridge which extends laterally
from the median capsular wall and differentiates the canalis
semicircularis posterior from its ampulla. It terminates abruptly
anteriorily behind the sinus superior utriculi. Dorsal to the
septum is situated the canalis semicircularis posterior while ven-
tral to it are the ampulla posterior, the utriculus, and the median
limb of the lateral semicircular canal. The two latter are differ-
entiated from each other by a broad ridge on the under side of
the septum.

The septum semicirculare laterale is a vertical plate situated in
a longitudinal plane separating the ampulla posterior, the cavum
vestibuli, and median limb of the lateral semicircular canal on
the one hand from the lateral limb and ampulla on the outer
side. At its posterior end the lateral margin of the septum semi-
circulare posterius and the lateral capsular wall unite.

The septum semcirculare anterius is a plate which extends
between the lateral and median capsular walls in an oblique
position, its anterior end lying somewhat higher than its posterior.
It lies dorsal and medial to the ampullae lateralis and anterior,
and ventro-lateral as well as posterior to the canalis semicircu-
laris anterior. The septum is slightly twisted; at its posterior
end its lateral margin lies at a higher level than its median mar-
gin, while at its anterior end the lateral margin is lower than the
median margin. In. front the septum arises from the median
capsular wall as a ridge which projects laterally and differentiates
the anterior semicircular canal from its ampulla. Posteriorly
the septum arises from the lateral capsular wall as a ridge which
demarcates the cavum sinus superioris utriculi from that of the
lateral ampulla.

DEVELOPMENT OF THE SKULL OF EMYS 719
TECTUM POSTERIUS

The tectum posterius, which is the only part of the chondrocra-
nium roofing in the central nervous system dorsally, is continuous
ventro-laterally with the otic capsules, and so may properly be
described with the otic region notwithstanding the fact that it
extends posteriorly to close in the foramen occipitale magnum
above. The tectum posterius may be differentiated into a me-
dian and two lateral portions. The latter are slender cylindrical
rods which slant medially and dorsally from the upper margin |
of the otic capsule in order to support with their distal median
ends the large median portion which is of rhomboidal form with
its long axis in the sagittal plane. This median portion accord-
ingly exhibits a long, slender triangular processus ascendens which
slants upward in an anterior direction and a stouter processus
posterior which is more nearly semicircular in form and extends
with its postero-ventral extremity to limit the foramen occipitale
magnum above. In cross section the processus ascendens be-
comes crescentic toward its apex, with the convexity directed
dorsally, while that of the processus posterior is more nearly
oval. The tectum is slightly curved in the sagittal plane, thus
exhibiting a postero-dorsal convexity and an antero-ventral
concavity. The saccus endolymphaticus is situated close below
and in front of the slender lateral portion.

CRISTA PAROTICA

The crista parotica (cr.p., fig. 28) forms a short, longitudinal
crest, extending in an anterior direction from the ventro-lateral
aspect of the prominentia ampullaris posterior slightly beyond
the foramen glossopharyngei externum, dorsal to which it ends
freely in a vertical, transverse face (cr.p., fig. 2). It is triangular
in cross section exhibiting a ventral and a median surface, the
latter attached to the otic capsule except at its extreme anterior
end which projects freely. There is accordingly a dorso-lateral
face which makes a broad, shallow groove with the part of the
external wall of the otic capsule lying dorsal to it in which the
postero-ventral portion of the quadratum rests.

720 B. W. KUNKEL

Ventral to the crest runs the vena capitis lateralis which joins
the v. jugularis interna immediately posterior to the crista
parotica a short distance anterior to the fenestra metotica and
continues obliquely upward and forward over the lateral wall
of the capsule ventral to the prominentia semicircularis lateralis.

A separate processus paroticus articulating with the crista
parotica, and which Gaupp has described in Lacerta, is not
present at any stage of Emys in the series studied.

RELATION OF THE NERVES IN THE REGIO OTICA

The relations of the nerves in the otic region of the skull exhibit
several points of considerable interest. As described by Bojanus
the n. glossopharyngeus passes directly through the cavity of
the osseous labyrinth of the ear in Emys. In the stage modelled
both the median and lateral walls of the cochlea are perforated
near their posterior ends and immediately ventral to the ves-
tibular portion of the capsule by the foramina glossopharyngei
internum and externum respectively. The n. glossopharyn-
geus passes from the inner foramen to the outer one in a ventro-
lateral direction. Between the two foramina the posterior wall
of the cochlea is excavated, as has been already described, to
form a deep groove, the sulcus glossopharyngeus, in which the
nerve lies with its anterior surface freely exposed in the cochlear
cavity. The dorsal wall of the groove is incomplete at its median
end because of the intersection of the cavum vestibuli with the
sulcus glossopharyngeus, so that the nerve has its dorsal as well
as its anterior surface for a short space exposed freely in the cavity
of the capsule.

With reference to the nervus glossopharyngeus, the lagena
extends much further ventrally and posteriorly than it does in
Lacerta in which form the n. glossopharyngeus leaves the skull
together with the vagus through the fissura metotica which
extends quite far forwards ventral to the pars cochlearis.

From the foramen glossopharyngei externum the nerve extends
in a ventro-lateral direction to the ganglion petrosum (s. glosso-
pharyngei) which lies ventral to the external fenestra and lateral
to the sulcus supracristularis.

DEVELOPMENT OF THE SKULL OF EMYS 721

The nervus acusticus exhibits a ramus vestibularis and a ramus
cochlearis. ‘The former is the smaller of the two and is situated
slightly dorsal and somewhat anterior to the latter. They enter
the capsule through the foramina acustica anterius and posterius
respectively. The anterior foramen is situated at the extreme
anterior end of the pars cochlearis immediately ventral to the
recessus utriculi; the posterior foramen is situated nearer the
middle of the pars cochlearis, immediately in front of the postero-
medial lobe of the cochlear cavity and ventral to the cavum
utriculi.

The nervus facialis finds its exit from the cranial cavity through
the foramen facialis which is situated a short distance dorsal to
the basal plate and in front of the anterior wall of the cochlea.
The ganglion facialis (s. geniculi) lies external to the chondrocra-
nium, separated from it by a thin layer of connective tissue, in
front of the foramen and dorsal to the lateral extension of the
basal plate already mentioned. The ganglion, accordingly, rests
in a broad groove between the basal plate on the ventral side and
the prefacial commissure, which unites the basal plate and otic
capsule, on the dorsal side. This groove is apparently the homo-
log of the fovea genicularis of the rabbit, as described by Voit
(08, p. 448).

The ramus palatinus nervi facialis arises from the lateral aspect
of the ganglion facialis and passes in a ventro-medial direction
around the free edge of the crista substapedialis and then turns
in an anterior direction in the sulcus palatinus which is medial to
the rudimentary processus basipterygoideus on the ventral side
of the basal plate.

The ramus hyomandibularis arises from the posterior end of
the ganglion geniculi and runs posteriorly in the fovea genicularis
and comes to lie further laterally and dorsally as it proceeds and
passes dorsal to the columella auris. At the level of the latter
the ramus hyomandibularis gives off the chorda tympani which
extends in a lateral direction along the dorsal margin of the colu-
mella and then turns sharply forward and ventrally to reach the
mandible, as Noack (’06) has already described. The: main

T22 B. W. KUNKEL

portion of the hyomandibular branch continues in a posterior
direction in the space betweeri the quadratum and the otic capsule.

At the point at which the palatine branch reaches the suleus
palatinus there is given off in a posterior direction the ramus com-
municans n. facialis cum n. glossopharyngeo which extends pos-
tero-laterally along the ventral surface of the chondrocranium
beneath the pars cochlearis of the otic capsule, the crista sub-
stapedialis, and the planum basale to unite finally with the ante-
rior end of the ganglion glossopharyngel.

REGIO ORBITO-TEMPORALIS

The orbito-temporal region shows the same general configura-
tion as that of Lacerta except that it is of much heavier and more
compact structure, being made up of broader plates instead of
slender rods. The fenestrae are also less numerous and relatively
smaller than in Lacerta, in which respect the chelonian skull
resembles more closely the primitive condition of Sphenodon.
The whole region has the form of a shallow longitudinal trough, .
within which the anterior end of the brain is supported on its
ventral side. Ventral to the trough, the septum interorbitale
(s.7., fig. 9¥ lies in a sagittal plane and increases in height from the
posterior to the anterior end. The sides of the trough are formed
by the pila prootica, the subiculum infundibuli, the pila metop-
tica, and the planum supraseptale. These walls are perforated
by the fenestrae prootica, metoptica, optica, and foramen oph-
thalmicum, besides a small perforation in the subiculum infun-
dibuli, present only on the right side of the embryo modelled
which represents the beginning of the absorption of the cartilage
at this point.

As in all the Sauropsida, the orbito-temporal region is differen-
tiated into a posterior temporal and an anterior orbital portion.
The temporal portion is continuous behind with the otic region
and exhibits a basal and two lateral parts. The orbital region
is characterized by the high septum interorbitale which forms
a continuous wall between the two orbits. This septum is a
thin vertical plate, as will be described later; in the stage modelled
it is thin, but in younger embryos it is relatively thicker and made

DEVELOPMENT OF THE SKULL OF EMYS 723

up of two parallel plates which are united to each other only
along the ventral margin by a high commissure (figs. 22 and 23),
a condition which apparently may be referred to a primitive form
in which the skull was platybasic. Anteriorly the septum inter-
orbitale passes without interruption into the septum nasi, which
in my youngest stages is apparently of unpaired origin.

The temporal region is made up of the trabeculae and pilae
prooticae, both arising from the anterior aspect of the crista
sellaris, the latter from the dorsal side, the former from the ven-
tral side. The bases of these two structures are separated from
each other by the nervus abducens which, as already mentioned,
penetrates the crista sellaris ventral to the base of the pila prootica
QZr fie. 8):

The trabeculae project horizontally forward in the plane of
the basal plate as triangular rods which converge in the median
line and enclose the semicircular fenestra hypophyseos in front
and on the two sides. They continue further forward as the
thickened ventral margin of the septum interorbitale and fuse so
that no trace of the paired nature of the septum remains. Pos-
teriorly the trabeculae are continuous with the antero-lateral
margins of the basal plate which represent the crista basiptery-
goidea, as-already described.

At the level of the posterior margin of the fenestra hypophyseos
a separate cartilago articularis, described by Gaupp in Lacerta,
is represented by a rudimentary, imperfectly chondrified mass of
tissue which is attached to the ventral edge of the crista basi-
pterygoidea (c.a., fig. 8). This cartilage forms a small, roundish
knob, about two-thirds as broad as long, which projects down-
ward from the crest with its free end directed anteriorly and ven-
trally. It extends lateral to the ramus palatinus n. facialis and
its free end is embraced laterally as well as ventrally by the ptery-
goideum. In older embryos the crista basipterygoidea becomes
relatively more prominent and is partly surrounded by the dorso-
median portion of the pterygoideum, and the cartilago articu-
laris disappears as a distinct piece, probably fusing with the
crest (c.pt., fig. 20).

724 B. W. KUNKEL

The pila prootica which forms the lateral portion of the tem-
poral region may be described in general as a broad U-shaped
plate of nearly uniform width, with its convex side anterior and
its concave side posterior, and exhibiting accordingly a median
and a laterallimb. It is attached to the basal plate by the median
limb, while the external limb of the U is somewhat dorsal to the
former and ends freely close to and medial to the anterior cupula
of the otic capsule. The plane of the whole plate is so curved
that while it is nearly horizontal in a transverse direction at its
attachment to the basal plate, the free lateral limb is nearly verti-
eal in its position. ‘The median limb is inclined antero-dorsally
from its proximal end where it is attached to the front surface of
the crista sellaris. The lateral limb of the pila exhibits a small,
anteriorly directed process which approaches a posteriorly directed
process from the postero-lateral angle of the planum supraseptale.
These two processes narrow the fenestra metoptica dorso-later-
ally without closing it in completely, and suggest, in their posi-
tion, an incomplete taenia marginalis. In another embryo,
slightly younger than that modelled, this anteriorly directed
process of the lateral limb is greatly enlarged and is perforated
by a large foramen, the significance of which is still undetermined.

In the embryo modelled, the pila prootica grows very thin
toward its dorso-lateral margin and passes gradually into a thin
layer of dense connective tissue which extends dorsally and medi-
ally and encloses the cranial cavity above.

A peculiarity of the individual modelled, which has not been
met with in any of the other specimens studied, is the presence
of a short cylindrical rod-like process from the dorsal surface of
the medial limb of the pila prootica near its distal end which
extends postero-laterally in a horizontal plane and ends freely in
the cranial cavity medial to the anterior margin of the fenestra
prootica. The nervus oculomotorius, as it passes anteriorly
over the medio-dorsal surface of the pila prootica in order to leave
the cranial cavity through the fenestra metoptica, lies dorsal and
medial to this process, while the n. trochlearis, which is approxi-
mately parallel to the oculomotor, lies lateral and dorsal to this
same process. Beneath the pila prootica, running horizontally

DEVELOPMENT OF THE SKULL OF EMYS 125

forward from the anterior apex of the ganglion semilunare, are the
nervi nasalis and frontalis of the ramus ophthalmicus.

The pila prootica bounds the fenestra prootica in front and on
its two sides and forms the posterior margin of the fenestra
metoptica. Immediately external to the fenestra prootica, and
occluding it almost completely, is the ganglion semilunare whose
anterior apex lies in front of the fenestra.

The fenestra hypophyseos is semicircular, the transverse front
margin of the crista sellaris forming its posterior boundary and
the two trabeculae enclosing it laterally and in front.

The arteria carotis Interna enters the cranial cavity through
the postero-lateral angles of the fenestra hypophyseos which are
not excavated to form incisurae caroticae as in Lacerta. The
ventral side of the crista sellaris, however, as already noted, bears
a broad longitudinal groove on each side of the middle line,
the sulcus palatinus, in which the arteria carotis interna, together
with the ramus palatinus n. facialis, lies; so that the extreme
anterior end of the sulcus where the artery turns dorsally to enter
the fenestra hypophyseos might be looked upon as an indica-
tion of an incisura carotica. Posteriorly the sulcus palatinus
continues as the sulcus cavernosus. i

The anterior three-fourths of the fenestra hypophyseos is com-
pletely occupied by the hypophysis cerebri which has a slightly
oblique position so that its anterior surface rests partly upon the
subiculum infundibuli in front of the fenestra.

The subiculum infundibuli arises in front of the fenestra hypo-
physeos from the dorsal surface of the, trabeculae as a thin plate
made up evidently of two symmetrical halves which are slightly
inclined toward each other in the middle line to form a shallow
longitudinal trough. The infundibulum and antero-dorsal end
of the hypophysis are supported by this plate. The subiculum
infundibuli is triangular, with its apex directed posteriorly and its
base anteriorly forming the hinder margin of the fenestra optica,
and, with its antero-lateral angles prolonged to form the short
pilae metopticae, and with its postero-lateral margins bounding
the ventral half of the fenestra metoptica in front The whole
subiculum infundibuli inclines rather sharply upward in front,

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

726 B. W. KUNKEL

being supported ventrally by the septum interorbitale which
increases rapidly in height from its posterior end. The lateral
margin as it passes over into the pila metoptica is incised by the
broad sinus oculomotorius for the oculomotor nerve as it passes
obliquely to the outside of the cranium through the fenestra
metoptica. Near the anterior margin, the subiculum infundibuli
is perforated by a pair of large oval foramina (foramen ophthal-
micum, f.o., fig. 25) through which the arteria ophthalmica
passes to reach the orbit. In the embryo modelled there is a
small fenestra on the right side only which represents the first
step in the resorption of this part of the chondrocranium. The
subiculum infundibuli is much thicker anteriorly and laterally
than medially and behind, and in older embryos in which the
resorption has been carried farther, the entire posterior part of
the subiculum has disappeared and the foramen ophthalmicum
becomes a sinus, opening freely behind into the fenestra metop-
tica. In older embryos the interorbital septum becomes fenes-
trated by a long, narrow opening which extends both anterior
and posterior to the subiculum infundibuli, its anterior end lying
beneath the fenestra optica, so that, as in Lacerta, a cartilago
hypochiasmatica is differentiated.

The pila metoptica is a short, stout rod extending antero-later-
ally from the antero-lateral angles of the subiculum infundibuli
to the posterior margin of the planum supraseptale. The nervus
oculomotorius passes ventral to the pila metoptica, leaving the
cranial cavity through the sulcus oculomotorius. The nervus
trochlearis passes out of the cranial cavity through the fenestra
metoptica lateral and dorsal to the nervus oculomotorius, lying
parallel to the latter beneath the pila metoptica.

The fenestra prootica is a large, oval opening between the ante-
rior cupula of the otic capsule and the prefacial commissure behind
and the pila prootica in front. On account of the curved form of
the latter the fenestra is enclosed by it dorsally as well; the
closure, however, is not complete on account of the absence of a
taenia marginalis and the failure of the pila prootica to fuse with
the otic capsule, a condition which resembles Sphenodon. The
fenestra is completely filled by the large ganglion semilunare

DEVELOPMENT OF THE SKULL OF EMYS G28

which extends with its anterior apex ventral to the pila prootica
(g.s., fig. 8). In Emys the ganglia of the three branches of the
n. trigeminus are consolidated into one mass unlike the condition
in Lacerta where that of the ophthalmic branch is separated from
that of the other two and lies quite far anteriorly.

The fenestra metoptica, through which the nn. oculomotorius
and trochlearis leave the cranial cavity, is a narrow slitlike open-
ing, having an oblique position, sloping forward and dorsally
from its base. It is bounded posteriorly by the pila prootica,
and anteriorly by the subiculum infundibuli, pila metoptica, and
planum supraseptale. From its anterior side the fenestra is
narrowed by the prolonged postero-lateral angle of the planum
supraseptale and again by a short posteriorly directed process
midway between this process and the lateral end of the pila metop-
tica. On the right side of the embryo modelled this second proc-
ess is reduced in the median part of its length so that only a proxi-
mal stump remains projecting from the posterior margin of the
planum supraseptale and a small isolated rod of cartilage which
lies freely within the fenestra (fig. 24). In another embryo of
approximately the same age, the fenestra metoptica becomes
regularly wider laterally and is narrowed only by the single
posteriorly directed process of the postero-lateral angle of the
planum supraseptale. The fenestra metoptica is not fully closed
dorsally because of the absence of a complete taenia marginalis,
although, as already described, the projection from the front
margin of the pila prootica and that from the hind margin of
the planum supraseptale suggest together an incomplete taenia
marginalis.

The fenestra optica (f.opt., fig. 24) is of irregular triangular
form, somewhat broader than long. ‘The planes of the two fora-
mina are slightly inclined toward each other in accordance with
the obtuse angle made by the two halves of the planum supra-
septale. The two foramina are separated from each other in the
middle line only by the free dorsal margin of the septum inter-
orbitale. Posteriorly they are bounded by the subiculum infun-
dibuli, and for a very short distance postero-laterally, by the pila
metoptica; anteriorly they are bounded by the planum supra-

728 B. W. KUNKEL

septale. The orbital portion of the obito-temporal region is
characterized by the septum interorbitale which increases in
height from behind and carries the planum supraseptale with its
dorsal margin so that the floor of the cranium is raised.

The septum interorbitale is a triangular plate lying in the sagit-
tal plane, arising posteriorly from in front of the fenestra hypo-
physeos and rapidly increasing in height anteriorly almost to
its extreme anterior end where the dorsal margin inclines rapidly
ventrally to pass into the septum nasi. The ventral margin is
thickened throughout its length to a cylindrical rod which is
continuous behind with the trabeculae and tapers anteriorly in
the region of the olfactory capsule to a thin edge. Its dorsal
margin is continuous with the subiculum infundibuli and planum
supraseptale except in the region of the fenestra optica where
it is free. The septum forms a continuous plate throughout with
no fenestrae such as Lacerta exhibits, although at a later stage,
as already mentioned, a long slit in the posterior portion of the
septum ventral to the fenestra optica causes a cartilago hypo-
chiasmatica to be differentiated.

Of especial interest in the stage modelled is the structure of
the extreme anterior dorsal corner of the septum where it passes
over intothe planum supraseptale posteriorly and the commissurae
spheno-ethmoidales anteriorly. At this point the septum is very
clearly paired, the right and left halves being separated from each
other by a shallow groove which is continuous postero-dorsally
with the troughlike planum supraseptale. By a comparison
with younger stages the significance of this peculiarity becomes
clear. In an embryo having a carapace length of 7 mm. the
portion of the septum lying in front of the fenestra optica is
represented by two vertical plates, parallel to each other which
are fused together along their ventral edges (fig. 23). Dorsally
these plates diverge rapidly from each other along a line corre-
sponding approximately to the dorsal margin of the septum in the
older embryo. In the region of the fenestra optica, and for a
short distance in front of the fenestra, the septum is thick so that ©
its ventral cylindrical margin is not differentiated from the rest

DEVELOPMENT OF THE SKULL OF EMYS 729

of the septum. In this young stage, accordingly, the cavum
cranii extends ventrally between the two plates of the septum
in the region between the eyes as a deep, narrow cleft.

The planum supraseptale forms a broad shallow trough in
which the cerebral hemispheres and olfactory lobes rest. Its
plane rises obliquely from behind, on account of the increasing
height of the septum interorbitale, but at its extreme anterior
end its plane falls rapidly ventrally so that it is almost vertical,
a condition associated with the rotation ventralwards of the olfac-
tory capsule as will be described later under the ethmoidal region.
In general, it is broadly oval in form, somewhat wider than long,
with its anterior and lateral margins evenly rounded, but its pos-
terior margin irregularly indented as has been described above.
Medially its posterior margin forms the front boundary of the
fenestrae opticae and further laterally the front boundary of the
fenestra metoptica; between these two fenestrae is the pila me-
toptica. The postero-lateral angle of the planum supraseptale is
produced behind as a slender rod-like process which approaches
the front edge of the pila prootica. The postero-lateral portion
of the planum supraseptale, corresponding somewhat in position
to that of the fenestra epioptica of Lacerta, is perforated by
numerous small foramina which represent the first stages in the
absorption of the cartilage in this part of the cranium.

At its anterior end, the planum supraseptale becomes very
narrow and passes into the anterior end of the septum inter-
orbitale. At the same time the two halves of the planum supra-
septale, which form an obtuse angle with each other throughout
most of their length, come to lie at the extreme anterior portion
with their median surfaces almost parallel and separated from
each other by a narrow cleft.

The commissurae spheno-ethmoidales, diverging from each
other, extend in an antero-ventral direction from the extreme
anterior margin of the planum supraseptale and serve to connect
the walls of the olfactory capsule with the orbital region. They
enclose the fenestrae olfactoriae laterally.

730 B. W. KUNKEL

REGIO ETHMOIDALIS

The ethmoidal region of Emys exhibits greater differences from
that of Lacerta than does any other region of the chondro-
cranium. It has been carefully described by means of models
by Seydel (’96), whose results are in the main confirmed by the
present study, and by Nick (712) in Chelydra serpentina, Chelone
midas, and Dermochelys coriacea. The homologies of certain
parts, however, which were not discussed by Seydel, have been
determined by study of a more complete series of embryos. To
Seydel’s observations there are only a few details to be added.

In striking contrast to the condition in Lacerta, the capsule
is more compact and is made up of continuous plates of cartilage
interrupted only to a limited extent by fenestrae and not modified
by alar processes and a complicated concha; besides this, the whole
capsule has undergone a bending in a ventral direction. The
differences in general form of this region in Lacerta and Emys are
to be correlated with the greater strength of the jaws of chelonians,
which condition requires that the bones against which the lower
jaw impinges (premaxillare, maxillare, vomer, palatinum) have
a more solid foundation to rest against than in such forms as
the snakes and lizards. Gaupp (06, p. 45) has already called
attention to the fact that the form of the olfactory capsule is
controlled as well by the structure of the jaws as by the form of
the olfactory sac itself.

The olfactory capsule is divided into two tea ction halves
by the septum nasi which continues in an anterior direction from
the septum interorbitale. The postero-dorsal wall of the capsule
accordingly forms the anterior boundary of the orbit and the
anterior wall forms the anterior limit of the head. On account
of its position relatively far ventral to the other portions of the
skull, its ventral wall projects below the level of the lower edge
of the septum interorbitale to afford a prominent, and at the
same time solid, support for the upper jaw. Besides the septum
nasi which forms the median wall of each half, the cartilages
making up the walls of the capsule may be differentiated into the
tectum nasi dorsally, the paries nasi laterally, the solum nasi and

DEVELOPMENT OF THE SKULL OP EMYS Gol

the cartilago paraseptalis ventrally, and the planum antorbitale
forming the postero-dorsal wall which separates the olfactory
sac from the orbit. These regions for the most part pass over
into each other by even transitions.

The entire dorsal and posterior portion of the capsule, as
Seydel has already shown, is free from the septum as is also the
case in Lacerta; that is, the planum antorbitale terminates freely
above and behind. Besides the connection between the septum
nasi and the walls of the capsule there is a further connection of
the capsule with the rest of the skull by means of the commissura
spheno-ethmoidalis, a cartilaginous rod extending posteriorly from
the tectum nasi on each side to the anterior extremity of the
planum supraseptale, where the latter passes into the septum
supraseptale, and enclosing the two fenestrae olfactoriae on the
sides, and thus separating the fenestra olfactoria from the fissura
orbitonasalis.

The capsule opens in front to accommodate the apertura nasalis
externa, by the fenestra narina, and posteriorly by the fenestrae
basales for the choanae. There is a longitudinal slit in the floor
of the capsule which Seydel has called the foramen praepalatinum.
The capsule is perforated dorsally by the fenestrae olfactoriae,
ventral to which on each side is the fissura orbitonasalis, through
the most dorsal part of which the ethmoidal branch of the ramus
ophthalmicus n. trigemini passes in its course from the orbit to
the olfactory sac. Besides these openings there is also a small
foramen epiphaniale for the passage of the n. lateralis nasi rami
ophthalmici from the capsule to the exterior. A foramen apicale
is lacking in all stages studied.

In contrast to Lacerta the entire capsule is shifted ventrally
and so rotated that the anterior end is slightly ventral to the
posterior end. This change in position is indicated partly by the
fact that the plane of the fenestra olfactoria is inclined from the
horizontal, as in Lacerta, so that its ventral end is below the pos-
terior. Besides this, the capsule is somewhat compressed later-
ally so that its floor is pressed ventrally and extends well below
the lower margin of the septum nasi. The two halves of the cap-
sule are thus separated on the under side by a groove which

Vou B. W. KUNKEL

becomes shallower in front; and the fenestra basalis is brought
well below the ventral margin of the septum, the upper side of
the fenestra standing at the level of the ventral margin of the
septum (figs. 9 and 27).

The contour of the walls of the capsule follows closely that of
the underlying olfactory sac. Accordingly the capsule may be
differentiated into two distinct regions, a relatively high, some-
what conical pars olfactoria situated dorsally, and a broader
cylindrical pars respiratoria, opening in front by the fenestra
narina and behind by the fenestra basalis. These two regions
are separated by a shallow groove which extends diagonally pos-
teriorly and ventrally and becomes more pronounced at its pos-
terior end. The depression causes the interior wall of the capsule
to be thrown up into a corresponding ridge which supports the
lateral ‘Grenzfalte’ of Seydel.

Above the groove just described is a second depression in the
lateral wall of the capsule. It is somewhat shallower than the
first and appears only along the middle third of the capsule. It
is the external manifestation of a corresponding ridge in the inte-
rior of the capsule. The accompanying prominent fold of the
nasal epithelium supported by the thickening of the lateral wall,
is evidently a rudimentary concha. It extends posteriorly as
far as the recessus ducti naso-pharyngei, which is immediately
ventral to it. In older embryos than that modelled, the concha
becomes much more prominent on the inner side of the paries
nasi (fig. 19).

The planum antorbitale is a thin curved plate making up the
posterior and postero-dorsal wall of the capsule. It slopes rapidly
ventrally from its antero-dorsal end and passes with an even
curvature laterally and anteriorly into the paries nasi. In the
stage modelled it is separated medially from the septum nasi by
the fissura orbitonasalis so that it has a free margin which is
rolled inwardly to project into the cavity of the capsule, parallel
to the septum (fig. 12). The infolded margin is broadest in front;
behind it disappears entirely in the region of the fenestra basalis.

The tectum nasi is short in an antero-posterior direction and
comparatively narrow from side to side. In front it falls sud-

©

DEVELOPMENT OF THE SKULL OF EMYS 733

denly ventrally to form the front wall of the capsule dorsal to the
fenestra narina. Posteriorly it passes into the planum antorbi-
tale and is continuous with the septum nasi and commissura
spheno-ethmoidalis. Between the septum nasi and the anterior
extremities of the commissurae spheno-ethmoidales accordingly,
the tectum bounds the fenestra olfactoria in front. Laterally
the tectum nasi bends smoothly ventrally to form the paries
nasi. Anteriorly it slopes abruptly ventrally and almost verti-
cally to enclose the pars olfactoria in front. On account of the
rounded form of this part, the tectum nasi exhibits in this region
two gentle convexities separated from each other by a groove
marking the position of the front margin of the septum nasi
with which it is in perfect continuity. Below each of these
cupulas, immediately above the fenestra narina, there is a short
blunt, anteriorly directed process which, in later stages, becomes
more conspicuous. To this process may be given the name
processus supranarinus (pr.s., fig. 25).

The paries nasi forms a practically continuous plate of cartilage
of slightly convex form limiting the olfactory capsule laterally.
It is continuous medially and dorsally with the tectum nasi-as
well as with the planum antorbitale; ventrally throughout its
entire length it is continuous with the solum nasi. Its anterior
end passes into the tectum nasi dorsally and ends freely ventrally
as the lateral margin of the fenestra narina. This free margin
is incised to form a very broad shallow sinus (sinus exterfus) for
the accommodation of the duct of the glandula nasi externa which
empties into the entrance passage of the olfactory sac (figs. 14 and
27). Posteriorly the paries nasi continues with the solum nasi
to form the cartilago ectochoanalis (c.e., fig. 9) which bounds the
fenestra basalis ventro-laterally. Above the cartilago ectocho-
analis the posterior margin is deeply incised to form a rounded
sinus in which the recessus ducti nasopharyngei lies and which
I would eall the sinus nasalis posterior. ‘The paries may be differ-
entiated into a dorsally situated pars olfactoria and a ventrally
situated pars respiratoria corresponding to the respective parts
of the olfactory sac lying within. These two regions are sepa-
rated from each other by a broad, shallow groove, which extends

734 B. W. KUNKEL

horizontally backwards from the level of the upper margin of the
fenestra narina to the apex of the sinus nasalis posterior and at
the same time grows shallower. The paries nasi is of uniform
thickness throughout except in the region immediately posterior
to the groove on the exterior, which marks the position of the
concha. Here the wall of the capsule is thickened to form a
more prominent, rounded ridge on the inner surface. In front
of this thickening the paries nasi exhibits on its medial surface
a ridge (corresponding to the groove already mentioned on the
external surface) which supports a fold of the epithelium of the
olfactory sac representing a rudimentary concha (fig. 11).

The floor of the nasal capsule offers the most striking differences
from the conditions met with in other reptiles. The solum nasi
projects in front ventral to the fenestra narina to form the-pro-
cessus infranarinus, which thus limits the sinus externus below
and medially. Behind the fenestra narina the solum nasi is
continuous with the septum for a short space, and laterally bends
upward to unite with the paries nasi with an even curvature.
This portion of the solum nasi represents the lamina transversalis
anterior of Lacerta. Posterior to this portion the solum nasi is
separated from the septum by an elongated slit to which Seydel
has given the name foramen praepalatinum (f.p. fig. 25). Behind
this foramen the floor of the capsule again unites with the sep-
tum. In this region the ventral margin of the septum inclines
strongly dorsally in a posterior direction. From the posterior
margin of the solum there projects posteriorly a short cartilago
paraseptalis which bounds the fenestra basalis medio-ventrally
and is separated from the septum by a narrow sinus paraseptalis.
Laterally the solum nasi at its transition to the paries nasi is
prolonged, as already mentioned, to form a short spout-like
process extending for a short distance posteriorly and supporting
the ductus naso-pharyngeus ventrally and laterally, which may
accordingly be called the cartilago ectochoanalis. The lateral
margin of this limits the sinus posterior below.

The significance of that portion of the solum nasi which is
continuous with the septum between the foramen praepalatinum
in front and the fenestra basalis behind is difficult to understand.

’
DEVELOPMENT OF THE SKULL OF EMYS 135

In a younger embryo than that modelled (fig. 30) the solum nasi
is separated from the septum completely behind a slender lamina
transversalis anterior so that the entire posterior balf of the solum
nasi is free from the septum much as in Lacerta, as Gaupp has
shown. Accordingly the foramen praepalatinum of Emys cor-
responds to the anterior portion of the cleft separating the sep-
tum from the cartilago paraseptalis, and the portion of the solum
nasi which is continuous with the septum behind the foramen
praepalatinum represents a portion of the cartilago paraseptalis
which has become extended medially and fused with the septum.
The solum nasi exhibits in the region lateral to the cartilago para-
septalis in the younger embryos several foramina which in the
embryo modelled have become occluded but which are still mani-
fest as thinner spots in the floor of the capsule.

The cartilago paraseptalis of Emys is relatively much shorter
than that of Lacerta and does not unite in the stage modelled
with the planum antorbitale by its posterior end. At a later
stage there is such a fusion similar to that in Lacerta. But,
unlike the condition in Lacerta, the choanae of Emys lie well pos-
terior to the cartilago paraseptalis instead of anterior to the pos-
terior end of the cartilage.

The relationship of parts of the floor of the capsule of Emys
may be derived from that of Lacerta by a widening in a median
direction of the cartilago ectochoanalis so that the fenestra
basalis is moved posteriorly by an obliteration from in front, and
at the same time by a broadening of the cartilago paraseptalis
in its posterior half so that it fuses with the septum. That these
changes may actually have taken place is indicated by the nasal
capsule of the younger embryos in which, as has already been
noted, the cartilago paraseptalis is separated from the septum for
its entire length, and by the fact that, lateral to the cartilago
paraseptalis, the floor of the capsule is perforated by a series of
fenestrae arranged in a longitudinal row.

Anteriorly, the solum nasi projects forward as a short process
ventral to the apertura nasalis externa; it passes over smoothly
into the paries nasi on the two sides. While the floor of the cap-
sule is nearly horizontal, except for a ventral crest along the lower

736 B. W. KUNKEL

margin of the septum, which Parker (’80) termed the ‘prenasal
cartilage,’ farther posteriorly the capsules extend much farther
ventrally so that there is a fairly deep median groove between the
two bulging capsules. The solum nasi projects posteriorly as
the cartilago ectochoanalis. Throughout the greater part of its
length the solum nasi is continuous with the septum, the foramen
praepalatinum and the slit separating the cartilago paraseptalis
and septum being the only points at which the floor is incomplete.

The septum interorbitale continues into the septum nasi inter-
rupted only by a small circular perforation situated near the
dorsal margin immediately in front of and between the bases of the
commissurae spheno-ethmoidales, so that there is formed a very
high ventral commissure and a slender, rodlike one between the
septum interorbitale and septum nasi.

The septum nasi is a thin vertical plate with a longitudinal
ridge on each side, the crista longitudinalis septi, which extends
from in front at the level of the upper margin of the fenestra
narina posteriorly and ventrally to the point at which the car-
tilago paraseptalis becomes separate from the septum (figs. 13
and 18). The septum is much shorter along its dorsal than along
its ventral margin, and is continuous anteriorly with the tectum
nasi, but posteriorly, in the region of the fenestra olfactoria, it
ends freely dorsally. The upper margin slopes downward in
front. ‘The ventral margin is continuous with the floor of the
capsule except in the region of the foramen praepalatinum and
that of the short cartilago paraseptalis. Like the upper margin it
slopes downward in front, but more rapidly than the upper mar-
gin so that the septum nasi is much higher in front than behind.
Especially in the region of the anterior half of the foramen prae-
palatinum the septum projects below the level of the floor of the
capsule to form a slight ventral crest along the mid-ventral line,
the prenasal cartilage of Parker. The front margin of the septum
is continuous in its dorsal half with the transverse vertical front
wall of the capsule and rounds off gradually into the dorsal mar-
gin; in its ventral half the septum ends freely between the two
fenestrae narinae to separate them completely from each other.
In a somewhat older stage of Emys than that modelled, Seydel

DEVELOPMENT OF THE SKULL OF EMYS (ei

describes the septum as not extending as far forward as I have
found it, so that the two fenestrae narinae are less perfectly sepa-
rated from each other.

Of special interest in connection with the septum nasi is a long
rodlike cartilage’ which, at its posterior end, is continuous with the
septum by the crista longitudinalis septi; it extends anteriorly,
parallel to the crest and separated from it by a narrow space,
and ends freely a short distance behind the anterior margin of
the septum. The glandula nasalis media is medial and ventral
to this rod while lateral to it and supported by it is the median
fold of the olfactory sac which separates the pars olfactoria from
the pars respiratoria. To this rod I would give the name pila
supraglandularis (p.sg., fig. 13).

In an embryo of 13.5 mm. carapace length the pila supraglan-
dularis is continuous at its posterior end with the floor of the cap-
sule immediately in front of the foramen praepalatinum (figs. 16
to 18). From this relation it follows, as Seydel has shown, that
the foramen praepalatinum opens into the olfactory capsule
directly in a dorsal direction behind and in a lateral direction
further forward beneath the pila supraglandularis. Later in
embryonic development (16 mm. carapace length) the pila supra-
glandularis fuses along its whole length with the crista longitu-
dinalis septi and forms a sloping roof extending ventro-laterally
from the septum dorsal to the glandula nasalis media.

The glandula nasalis media les with its posterior end in the
anterior half of the foramen praepalatinum and extends obliquely
forward and dorsally, parallel to the septum, beneath the pila
supraglandularis. The gland opens at its anterior end into the
olfactory sac ventral to the anterior end of the crista longitudinalis
septi.

The cartilago paraseptalis is a short plate forming the ventro-
median boundary of the fenestra basalis. It is continuous in
front with the floor of the capsule which exhibits ventrally in its
posterior part a moderately deep groove between its two halves.
The cartilago paraseptalis accordingly exhibits a ventro-median
and a dorso-lateral face at its anterior end, but toward its poste-
rior free end it is rotated slightly so that the faces are directed

(oo. * B. W. KUNKEL

medially and laterally and the cartilages of the two sides are
nearly parallel. The dorsal margin of the cartilago paraseptalis
lies slightly above the level of the ventral margin of the septum;
further anteriorly, where it is more nearly horizontal, its medio-
dorsal margin is at the level of the lower margin of the septum.
The postero-dorsal angle of the paraseptal cartilage comes to lie
in contact with the planum antorbitale for a very short space so
that the fenestra basalis is almost completely surrounded by car-
tilage. It fails, however, to fuse with the planum antorbitale, as
in Lacerta until a much later stage (c.p., fig. 10).

The cartilago ectochoanalis is a posterior prolongation of the
solum nasi which forms a short process ventral and lateral to the
choana, with its dorsal surface concave and ventral surface con-
vex, forming thus a trough-like projection. It is differentiated
from the cartilago paraseptalis only by a shallow incision at its
posterior end.

The fenestra basalis through which the choana passes poste-
riorly is vertical and transverse in position and is situated below
‘the level of the ventral margin of the septum interorbitale. In
the stage modelled the fenestra is not completely separated by
cartilage from the fissura orbitonasalis because of the failure of
the cartilago paraseptalis and planum antorbitale to fuse dorsally.
The ventro-lateral margin of the fenestra projects posteriorly as
the short cartilago ectochoanalis, above which the lateral margin
of the fenestra is incised to form the broad sinus nasalis posterior.
The medio-ventral margin is likewise incised to form the sinus
paraseptalis. The fenestra is oval with the long axis oblique
from dorso-medial to ventro-lateral, and with its dorsal end more
acute than the ventral. In the Emys embryo modelled the dorsal
end of the fenestra is not closed by cartilage but by a mass of
dense connective tissue between the planum antorbitale and car-
tilago paraseptalis.

The two fenestrae narinae together form a large round foramen
which, as it were, truncates the nasal capsule anteriorly. Unlike
that of Lacerta, which faces ventrally and laterally, that of Emys
is directed anteriorly in a vertical transverse plane. Dorsally

DEVELOPMENT OF THE SKULL OF EMYS 739

and ventrally from each foramen the capsular wall is produced
in an anterior direction into a short process, the dorsal one of
which is slightly stronger than the ventral. Laterally the margin
of the foramen is incised to accommodate the duct of the glan-
dula nasalis externa which passes medially to open into the nasal
sac. The fenestra narina lies in the ventral half of the capsule
so that the entire front wall of the pars respiratoria is lacking.
The septum nasi ends freely in front on a level with the lateral
margins of the fenestra so that the two fenestrae are completely
separated from each other.

A small fenestra epiphaniale in the antero-dorsal portion of the
paries nasi, near the point at which the latter passes over into the
tectum nasi, is the only perforation of the paries nasi in the stage
modelled. Through this fenestra the nervus lateralis nasi of
the ophthalmic branch of the trigeminus passes to the exterior
of the capsule.

The foramen praepalatinum is an elongated slit separating
the solum nasi from the septum nasi for a space posterior to the
lamina transversalis anterior. It is divided by a projection of its
lateral margin into a smaller anterior and larger posterior lobe.
The indentation of its lateral margin is produced by the posterior
end of the pila supraglandularis which is attached at this point
to the solum nasi as already described.

The fenestra olfactoria (f.ol., fig. 24) as already mentioned,
slopes ventrally in front from the antero-ventral margin of the
planum supraseptale to the tectum nasi. It is a long oval,
bounded medially by the free margin of the septum nasi and
laterally by the commissura spheno-ethmoidalis. The planes of
the two fenestrae are slightly inclined toward each other because
the septum does not extend dorsally as far as the level of the com-
missurae spheno-ethmoidales.

The fissura orbitonasalis is continuous ventrally and ante-
riorly with the sinus paraseptalis. Dorsally it is separated from
the fenestra olfactoria by only the slender commissura spheno-
ethmoidalis. The nervus ethmoidalis rami ophthalmici trigemini
gains access to the nasal capsule from the orbit through it.

740 B. W. KUNKEL

PALATOQUADRATUM AND MANDIBLE

The palatoquadratum of Emys is of special interest since it
has a distinet pars palatina, or processus pterygoideus; a condi-
tion which recalls the primitive one of the skull of Sphenodon and
the Anamnia in general. It is entirely disconnected from the
rest of the skull. The pars quadrata lies close alongside of the
lateral wall of the otic capsule and the pars palatina alongside of
the crista basipterygoidea, both parts, however, are separated
from the median portion of the skull by at least a thin layer of
non-cartilaginous tissue. The form of the pars quadrata has
been very aptly compared to that of the external ear of man with
the convex margin directed anteriorly and the concave margin
posteriorly; accordingly the anterior portion of the pars quadrata
is much higher dorso-ventrally than the posterior portion. It is
of about the same size as the otic capsule so that it conceals the
latter somewhat when viewed from the side. Its ventral portion,
the pars articularis, lies well below the level of the basal plate so
that its dorsal margin lies also somewhat below that of the otic
capsule. In its posterior portion, the pars quadrata, fits into the
broad, shallow groove formed on the lateral wall of the otic
capsule by the crista parotica and the portion of the prominentia
canalis semicircularis lateralis which lies dorsal to it. The median
wall of the pars quadrata has a nearly vertical position so that it
is much further removed from the capsule ventrally than dorsally
where the lateral semicircular canal projects. In the compara-
tively narrow space between the quadratum and the otic capsule
are situated the arteria carotis interna and vena capitis lateralis
which extend, parallel to each other, obliquely from postero-ven-
tral to antero-dorsal, as has already been described by Noack.

The pars quadrata may be differentiated into two distinct
portions, a postero-dorsal one which encloses an extension of the
tympanic cavity, regarded by Hasse as the homologue of the
antrum mastoideum of human anatomy (pars mastoidea), and
an antero-ventral portion which lacks the outer wall of the former
and so exhibits an imperfect cup shape (pars articularis). The
pars mastoidea of the quadratum forms a hollow cone, flattened
from side to side, with its apex directed posteriorly and its base

DEVELOPMENT OF THE SKULL OF EMYS 741

directed anteriorly. The lateral wall of this region exhibits an
irregular fenestra which indicates simply the beginning of the ab-
sorption of the cartilage under the influence of the adjacent squa-
mosum. The base of this portion of the quadratum opens freely
into the postero-dorsal portion of the pars articularis dorsal to the
incisura columellae. The pars articularis is continuous with the
pars mastoidea by means of the median wall and the dorsal mar-
gin of the latter which is rolled over laterally and ventrally to a
slight extent to enclose partially the concavity of this region from
above. ‘The anterior and ventral margins of the anterior half
of the quadratum are rolled over laterally to form a rim enclos-
ing the tympanic cavity in front and below. Ventrally this rim
becomes somewhat heavier and higher to form the processus
articularis. The anterior half of the pars quadrata is much higher
than the posterior, so that the processus articularis projects well
below the ventral margin of the posterior part. Between the
latter and the ventrally projecting processus articularis in front
-is a broad triangular sinus (incisura columellae) for the stalk of
the columella.

The processus articularis bears a saddle-shaped articular sur-
face for the articulation of the lower jaw. The axis of the surface
is slightly oblique, sloping ventrally and posteriorly from the
medial to the lateral side.

The pars palatina arises from the anterior end of the median
aspect of the pars quadrata near its ventral angle and extends
anteriorly and ventrally as a gradually tapering rod, its distal
end bending latero-ventrally at a right angle. Of especial inter-
est in relation to the pars palatina is the processus ascendens.
This is a slender, laterally compressed process which arises from
the processus pterygoideus, about midway between its origin from
the pars quadrata and its sharp lateral bend. The processus
ascendens extends lateral to the foramen prooticum, but some-
what removed from it. Its free distal end is close to the ventral
end of the processus descendens of the parietale. Judged by
its relations to the nervus trigeminus and to the palatoquadra-
tum, it is the homologue of the ‘columella’ of the kionocraniate
lizards, as has been pointed out by Filatoff (06). It extends

JOURNAL OF MORPHOLOGY, VOL. 23, No. 4

742 B. W. KUNKEL

medial to the rami maxillaris and mandibularis of the nervus
trigeminus and lateral to the ramus ophthalmicus. As develop-
ment proceeds this process is reduced in size, becoming replaced
by the very variable epipterygoideum of the adult.

The quadratum stands in very close relationship topographi-
cally with the columella auris as has already been shown, the
posterior margin being deeply incised for the accommodation of
the slender stalk of the columella as it passes horizontally from the
otic capsule as far laterally as the outer wall of the quadratum
where it is expanded to form a mushroom like plate which par-
tially fills the cup-like ventral half of the quadratum.

COLUMELLA AURIS

The columella auris in the embryo modelled consists of a single
cartilaginous rod in which only a trace of its origin from two cen-
ters is preserved in the fact that the chondrification of the ex-
ternal end is very distinctly less advanced than of the rest. In
earlier stages, however, the stapes and extracolumella, which
together form the columella auris, are very evidently distinct.
Topographically are to be distinguished the foot plate, stalk, and
insertion piece, of which the first two belong to the stapes and the
last to the extracolumella.

The foot plate is triangular and fits closely into the fenestra
vestibuli, resting with its ventral margin upon the planum basale.
The stalk arises from near the anterior angle of the foot plate,
gradually assuming its slender, cylindrical form and extending
at right angles to it in a lateral direction. It is slightly sigmoid,
with its lateral extremity further forward than its median extrem-
ity. It extends laterally across the cavum tympanicum ventral
to the vena capitis lateralis and nervus facialis and resting in the
incisura columellae of the quadratum as already described. It
passes into the median surface of the insertion piece somewhat
dorsal to the center of the latter. The insertion piece is mush-
room-shaped with its medial surface flat and lateral aspect con-
vex. Its plane is vertical and continuous with the lateral wall
of the quadratum. It aids in closing the cavum tympanicum
laterally.

DEVELOPMENT OF THE SKULL OF EMYS 743

Of especial interest in connection with the extracolumella is
the processus interhyalis described by Bender (’11) in Testudo
and the present writer (Kunkel ’12) in Emys, extending from the
postero-ventral corner medially and ventrally. From the apex
of the processus interhyalis a strand of dense connective tissue
extends ventro-medially towards the lateral aspect of the pars
retroarticularis of Meckel’s cartilage as Fuchs (’07) has described.
This process, in earlier stages, is distinct from the extracolu-
mella as the interhyale.

The origin of the columella auris from two separate centers is
also clearly shown in an embryo of 7 mm. carapace length. As I
have shown in an earlier paper (Kunkel ’12) in this embryo the
insertion piece, corresponding to the extracolumella, is distinct
from the stalk, which, together with the foot plate, represents the
stapes (col., fig. 21). Noack’s conclusion (’07) that the columella
of chelonians is a derivative of the capsular wall is not confirmed
by my observations. In one of my earlier stages (carapace length,
5.2 mm.) the blastema of the columella is distinctly in the pre-
chondrial stage while that of the capsule has not proceeded so
far. In this embryo the stapes extends medially as far as the
lateral wall of the otic sac and is represented by a mass of pre-
chondrium far in advance in development of any in its immediate
neighborhood.

The relation of the nervus facialis to the columella, I find is
essentially as Noack has described. The ramus hyomandibularis
extends caudally in a straight line from the ganglion geniculi,
dorsal to the stem of the columella, to the muscles which it in-
nervates. Almost immediately behind the columella it gives off
laterally the chorda tympani which passes first in a lateral direc-
tion as far as the quadratum and then curves forward, crossing
the columella on its dorsal side, and then turns ventrally in front
of the Eustachian tube eventually to reach the mandible.

Meckel’s cartilage is of strong form, tapering regularly from the
condyle anteriorly. In cross section in front of the condyle it
is elliptical with the long axis of the ellipse oblique from dorso-
lateral to ventro-medial. At the condyle the rami of the mandi-
ble are flattened considerably as if by pressure from above so

744 B. W. KUNKEL

that each one flares out laterally to form a distinct flange. The
upper surface of the condylus mandibularis is saddle-shaped to
fit into the processus articularis of the palatoquadratum. 'The
processus retroarticularis is very poorly developed, extending
only a very short distance behind the condyle as a somewhat
laterally compressed keel. The distal ends of the two rami of the
mandible are united by a strong symphysis in the form of a mod-
erately thin horizontal plate of triangular form which extends
behind the anterior ends of the rami and fills up the angle made
by the two. An independent ‘basimandibular’ element, such as
Parker (’80) describes in Chelone, is not present. In earlier stages
than that modelled, the symphysis is much shorter, although sec-
tions in this region show the rami to be widening medially just
behind their union to fuse finally and increase the strength of
the symphysis.

In the stage modelled, the proximal end of Meckel’s cartilage
has not yet begun to ossify to form an articulare. This ossifica-
tion becomes evident for the first time in an embryo having a
carapace length of, 13.5 mm.

In the earliest stage which I have studied (carapace length,
4.7 mm.) Meckel’s cartilage is present as a cylindrical rod of

prechondrium which tapers gradually as it proceeds from its’ ~

articulation to its distal (anterior) end. At this stage the rami
do not meet in a symphysis. The processus retroarticularis is
relatively longer than at a later stage, possibly because the con-
dylus, in becoming larger, usurps a portion of the original pars
retroarticularis.

HYOID AND VISCERAL ARCHES

The development of the hyoid and visceral arches of Emys
has already been described by Fuchs (’10) to whose account I
can add from my own study of this region only a few points.

The corpus hyale in the stage modelled is simple and of the
pentagonal form usual in the adult chelonians. The apex is
directed anteriorly and from its lateral margins project three
pairs of processes. It is somewhat concave dorsally and convex
ventrally to form a shallow cup with a thickened rim in which the

DEVELOPMENT OF THE SKULL OF EMYS 745

larynx rests. The anterior apex of the corpus is produced ante-
riorly to form a conical processus lingualis which exhibits on its
dorsal surface at its proximal end a shallow longitudinal groove
representing the original space between the pair of processes
of the corpus from which the processus lingualis has developed.
The two antero-lateral angles of the corpus (processus lateralis
anterior) are continuous laterally and posteriorly with the very
short cornua hyalia.. The processus lateralis intermedius is
small and separated from the cornu branchiale primum. The
processus lateralis posterior is continuous at this stage with the
stout cornu branchiale secundum. The cornua hyalia may be
differentiated from the processus lateralis anterior by a deep
cleft which extends anteriorly from behind so that the cornu
itself is represented by a broad, flat, stump which projects pos-
teriorly and is thicker at its distal than at its proximal end.

The cornua branchialia prima are the longest of all the cornua
and are long cylindrical rods, tapering very gradually toward
their distal ends, and curved at first slightly in a postero-lateral
direction and then toward their distal ends in simply a dorsal
one. At their extreme distal ends they bend sharply in a medial
direction and anteriorly to form short hooks whose free ends are
directed forwards. These rods are in close relationship to, but
are not continuous with, the processes of the corpus by their
proximal ends. The cornua branchialia secunda are at this
stage continuous with the processus laterales posteriores. They
are heavy rods of ellipitical cross section, rather thicker than the
cornua branchialia prima. They are parallel in general with the
last mentioned and terminate with their distal ends slightly
upturned some distance ventral and medial to the extremity of
the cornu branchiale primum. The second pair of cornua are
considerably shorter than the first and so do not extend so far
posteriorly.

No trace of an entoglossum occurs in the embryo modelled,
but in an older individual (carapace length, 28 mm.) I find a
thin plate of cartilage just ventral to the anterior end of the cor-
pus hyale. It extends beyond the processus lingualis and is sep-
arated from it by only a thin layer of connective tissue. It is

746 B. W. KUNKEL

somewhat triangular, with the base posterior and apex in front.
This cartilago entoglossalis is imbedded in a much larger mass of
fibrous tissue in which some scattered cartilage cells are pres-
ent. In the adult, this surrounding mass of fibrous tissue forms
a conspicuous heart-shaped plate beneath the anterior end of
the corpus hyale.

In the adult of Emys, as already known, the three pairs of
cornua are distinct from the corpus hyale and are connected with
it by connective tissue. Already in a young individual with a
carapace length of 28 mm. the three pairs of cornua are distinct —
from the corpus although the cornu branchiale primum is the only
part of the hyobranchial arches which exhibits an ossification.
The cornu hyale becomes segmented from the corpus hyale
slightly earlier than does the cornu branchiale secundum; for
example, in an embryo with carapace length of 13.5 mm. the
separation of the cornu hyale is complete while the cornu bran-
chiale secundum is only partially separated.

A separate epibranchiale primum (Siebenrock ’99) is present
as a separate triangular cartilage lying dorsal to the extreme dis-
tal end of the cornu branchiale primum in an embryo having a
carapace 13.5 mm. long. It remains as a distinct cartilaginous
element in the adult.

In the fully grown Emys the cornua and corpus hyale are ossi-.
fied and only the processus lingualis, a small oval foramen in the
anterior part of the corpus, the extreme distal ends of the cornua
branchialia primum and secundum, and the epibranchiale primum
remain chondrified.

The earliest portion of the hyobranchial apparatus to be laid
down in cartilage is the cornu branchiale primum which is already
chondrified in an embryo having a carapace length of 7 mm.
It likewise is the earliest to show signs of ossification. In the
embryo modelled, ossification has already begun in its middle
portion, a short distance behind the posterior margin of the
corpus hyale. The cornu hyale chondrifies later than the cor-
pus hyale and at essentially the same time as the cornu branchiale
secundum; so that in the stage modelled, although the body is
completely chondrified, the cornu hyale is represented by a sepa-

DEVELOPMENT OF THE SKULL OF EMYS 747

rate cartilaginous center at some distance from the corpus and is
surrounded by a mass of chondroblasts which is continuous with
it.

The processus lingualis arises from a pair of short processes
extending forward from the front margin of the corpus hyale at
a short distance from each other and converging slightly toward
their distal ends. In the stage modelled it is present as a single
median rod of cartilage with only a slight indication on its dorsal
surface, in the form of a shallow longitudinal groove, of its paired
origin. Cross sections, however, prove its double nature even
at a much later stage.

MEMBRANE BONES

The membrane bones are all laid down in the stage modelled
with the exception of the parasphenoideum, quadrato-jugale, and
the complementare of the lower jaw.

The squamosum (s., figs. 24, 25 and 26) is a thin plate of bone
of irregular triangular form, exhibiting a convex lateral and a
concave median face. It lies parallel to the outer side of the
quadratum, overlying the posterior extension of the same, and
separated from it by only a narrow layer of connective tissue.
Its posterior margin projects slightly beyond the quadratum
while its dorsal margin shows a tendency to curve medially to
embrace the quadratum: more closely. The ventral angle of the
bone is also bent medially below the posterior end of the quad-
ratum.

As the embryo increases in size the squamosum comes to lie
with its postero-ventral angle resting against the crista parotica
while its dorsal margin extends medially above the quadratum
to come in contact with the lateral wall of the otic capsule dorsal
to the prominentiae canalis semicircularis lateralis and ampul-
laris posterior. The anterior angle of the squamosum also comes
to extend relatively further forward in older embryos.

The first appearance of a squamosum is in an embryo having
a carapace length of 7 mm., at which stage it has the form of
a shallow saucer in contact by its margins with the posterior
extension of the quadratum.

748 B. W. KUNKEL

In its relation to the quadratum, the squamosum agrees closely
with that established by Thyng (’06) as the criterion of that
bone.

The quadrato-jugale is not present in the stage modelled, but
appears first in my next older embryo (carapace length, 13.5
mm.). In this it forms a triangular plate with its ventral angle
greatly prolonged in front of the pars quadratum of the palato-
quadratum. Its postero-dorsal angle lies a short distance ven-
tral to the anterior angle of the squamosum and its postero-
ventral margin follows closely the free anterior margin of the pars
articularis of the palatoquadratum. Its anterior portion comes
to lie ventral to the posterior end of the postfrontale and pos-
terior to the posterior end of the zygomaticum. In older embryos
the posterior end of the quadrato-jugale is overlapped externally
by the squamosum.

It is of special interest in this connection that in such chelonians
as Cistudo ornata, Chelodina longicollis, and Geoemyda spin-
osa, the embryonic character of the absence of a quadrato-jugale
is retained through life.

The zygomaticum (z., fig. 26) in the stage modelled is in the
form of along, slender plate of bone bent so that it exhibits one
limb extending horizontally forward and the other obliquely
postero-dorsally. At the same time the plane of the bone is
somewhat twisted on its own axis so that whereas the postero-
dorsal portion is sagittal, exhibiting a lateral and medial surface,
the anterior portion is more nearly horizontal, exhibiting a dorsal
and a ventral surface. The bone lies freely in the postorbital
region quite far in front of the front margin of the pars quadrata.
Together with the postfrontale, it completes the posterior margin
of the orbit. Its anterior end extends forward dorsal to the
hinder end of the maxillare.

The maxillare (m.) consists of a thin elongated plate of bone
having in general a horizontal position and forming the ventro-
lateral margin of the anterior end of the skull. Its dorsal sur-
face is slightly concave so that it exhibits a shallow longitudinal
groove; ventrally it is strengthened by a vertical rib which grows
higher toward the anterior end. The plane of the horizontal

DEVELOPMENT OF THE SKULL OF EMYS 749

portion of the maxillare becomes somewhat twisted at the ante-
rior end where the bone embraces the olfactory capsule so that
the dorsal surface becomes strongly inclined toward the median
plane of the head. There may be distinguished three definite re-
gions, the processus palatinus, processus alveolaris, and processus
‘ praefrontalis. The processus palatinus is horizontal, exhibiting a
dorsal and a ventral surface; it is of nearly uniform thickness
throughout, but becomes slightly thicker at the anterior than
at the posterior end. Near its posterior end it is perforated
by a small foramen for the ramus maxillaris nervi trigemini,
which passes from above through the foramen to continue
forward on the ventral side of the maxillare, lateral to the pro-
cessus alveolaris. This latter has a vertical position, spring-
ing from the mid-ventral line of the processus palatinus and
increasing in height from posterior to anterior. For the ante-
rior one-third of its length the maxillare is made up entirely of
the processus praefrontalis which is continuous ventrally with
the anterior end of the processus alveolaris and dorsally with
the processus palatinus. At the junction of the processus pala-
tinus and praefrontalis are several irregular foramina (foramina
alveolaria externa superiora.)

The processus praefrontalis has a concave median and convex
lateral surface and embraces the posterior two-thirds of the -
olfactory capsule ventro-laterally. Its antero-median margin is
hollowed slightly in order to accommodate the praemaxillare.
The dorsal margin of the processus praefrontalis lies in* the same
plane with and rather close to the ventral margin of the prae-
frontalis.

The processus palatinus is overlapped dorsally for a very short
distance by the anterior end of the zygomaticum so that the
two bones together complete the margin of the orbit posteri-
orly and ventrally. It lies in the same horizontal plane with
the palatinum which comes to lie in close relation to it by its
anterior end.

The praemaxillare (prm.) is of small irregular form situated
ventral to the olfactory capsule in the region of the foramen prae-
palatinum and medial to the anterior end of the maxillare. Like

750 B. W. KUNKEL

the maxillare it exhibits a processus palatinus and a processus
alveolaris. The former is horizontal, exhibiting a dorsal and a
ventral surface, and is prolonged anteriorly at its median mar-
gin to form a short process which extends nearly as far as the
anterior margin of the solum nasi and which may represent a
very rudimentary processus praenasalis. Even in the adult
Emys, however, the praemaxillare remains wholly ventral to the
olfactory capsule and does not extend anterior to it to separate
the two fenestrae narinae from each other as in Lacerta, for
example. The processus alveolaris is a vertical plate extending
from the ventral surface of the processus palatinus in an oblique
direction from postero-lateral to antero-medial.

The postfrontale (pst.) has the form of a triangular plate with
an apex directed forward and its upper margin horizontal; its
antero-ventral margin is slightly concave and thickened by
a laterally projecting flange which arises near the ventral angle
of the bone and extends to the anterior apex. It lies in the space
between the quadratum posteriorly and the orbit anteriorly,
lateral to the parietale. It approaches the posterior extremity
of the zygomaticum with its ventrally projecting angle, and the
anterior margin of the quadratum in its most lateral portion with
its posterior margin. It has no share whatever in the formation
- of the cranial wall. ;

The parietale (par.) and frontale (f.) have essentially the
same relations to the chondrocranium that they do in Lacerta;
that is, tHeir position is wholly lateral and in the stage of Emys
modelled they do not begin to roof in the cranial cavity. The
parietale is large and triangular in form exhibiting long dorsal
and antero-ventral margins, a short posterior one, and a lateral
and a medial face. Its postero-dorsal angle is prolonged slightly
to form a small processus occipitalis, extending dorsal to the otic
capsule for a short distance. Ventrally the parietale is pro-
longed to form a processus inferior which extends downward
lateral to the pila prootica to come into close relationship with
the free end of the processus ascendens of the palatoquadratum.
The processus inferior thus comes to enclose a space external
to the primordial cranial cavity so that the latter is increased in
size on the sides. This space is the cavum epiptericum.

DEVELOPMENT OF THE SKULL OF EMYS fol

The antero-ventral margin of the parietale lies parallel to the
free margin of the planum supraseptale at a short distance dor-
sal and lateral to it, while its anterior angle comes to lie medial
to the posterior end of the frontale. The posterior margin of
the parietale rests with its projecting processus occipitalis against
the anterior cupola of the otie capsule.

The frontale is a narrow, curved plate that continues the con-
tour of the parietale forward. Its ventral margin, like that of
the parietale, lies parallel to the free margin of the planum
supraseptale and completes the orbit dorsally. Its posterior end
overlaps laterally the anterior end of the parietale and its
extreme anterior end is embraced by the dorsal extremity of
the praefrontale.

In contrast to the condition in Lacerta the frontale in Emys
extends less far posteriorly so that it fails to be overlapped by
the anterior extremity of the postfrontale. The frontale at this
stage of development is relatively small, its place being taken to
a certain extent by the extremely large praefrontale in front and
the unusually large parietale behind.

The praefrontale (pf.) or lacrimale is very large, completing
the orbit anteriorly and embracing the greater part of the olfac-
tory capsule. In form it exhibits a vertical plate which extends
beyond and lateral to the antero-ventral end of the frontale,
and a larger oblique portion lateral to the olfactory capsule and
lying dorsal to the anterior portion (processus praefrontalis) of
the maxillare. The anterior margin of the vertical limb of the
praefrontale becomes very thick, so that there is a distinct ante-
rior, as well as a lateral and a medial face. It may be that
this thickening of the dorsal portion of the praefrontale stands
in connection with the fact that a nasale is lacking.

The pterygoideum (pt.) is strongly developed in the stage of
Emys modelled, and exhibits clearly many of the characteristics
of the adult form. It is a long, slender bone with ventral face
extending horizontally and a longitudinal crest, the crista ptery-
goidea (c.pt., fig. 20), extending dorsally, which diminishes in
height from posterior to anterior and separates the fossa supra-
pterygoidea, lying lateral to it, from the sulcus cavernosus on

752 B. W. KUNKEL

its median side. At its anterior end the crista pterygoidea dis-
appears and the pterygoideum extends forward as a small hori-
zontal process between the posterior end of the platinum laterally
and the fused trabeculae medially. Near the anterior end of
the lateral margin of the pterygoideum a conspicuous triangular
process—the processus ectopterygoideus—projects laterally be-
neath the distal end of the processus palatinus of the palato-
quadratum. Its posterior end les. with the crista pterygoidea
between the base of the processus pterygoideus of the palato-
quadratum laterally and the projecting antero-lateral margin of
the planum basale, coming to le in close relation to the rudi-
mentary cartilago articularis (c.a., fig. 8). The body of the ptery-
goideum lies in a plane ventral to the planum basale, extending
laterally beneath the processus pterygoideus of the palato-
quadratum and medially beneath the processus basipterygoideus
to enclose from the ventral side a space in which run the ramus
palatinus n. facialis and ramus communicans n. facialis cum
glossopharyngeo, as well as the arteria carotica interna, which lie
in the sulcus cavernosus on the median side of the crista ptery-
goidea.

The palatinum (pal.) is a flat, triangular plate with its apex
directed posteriorly and base anteriorly. It lies at the same
level as the processus palatinus of the maxillare and occupies
the space in the roof of the mouth between the maxillare, vomer,
and pterygoideum. Its posterior end lies external to the ante-
rior prolongation of the pterygoideum and in front of the pro-
cessus ectopterygoideus. The plane of the palatinum is inclined
so that its median margin is somewhat dorsal to its lateral
margin.

The parasphenoideum is wanting in the stage modelled, but
in an older embryo (carapace length, 13.5 mm.) it is present as
a small tripartite plate lying ventral to the region of the fenestra
hypophyseos. Its anterior median process extends as far for-
ward as the front margin of the fenestra while its postero-lateral
processes extend scarcely as far laterally as the lateral margins
of the fenestra.

DEVELOPMENT OF THE SKULL OF EMYS 753

The parasphenoideum lies immediately in front of and par-
tially embraces on the two sides the stalk of the hypophysis at
this stage. It is also of interest that a small lamella of bone
extends horizontally forward from the front margin of the crista
sellaris to occlude partially the fenestra hypophyseos. In an
older embryo having a carapace length of 16 mm. this lamella
from the crista sellaris has extended itself further forward and
fused completely with the parasphenoideum so that only a very
small opening is left between the posterior processes of the
parasphenoideum and this lamella for the accommodation of the
hypophysial stalk. In the meanwhile the ossification of the crista
sellaris, which becomes the basisphenoideum, has set in.

The basisphenoideum of the adult Emys must therefore be
composed of two parts, an anterior part not preformed in car-
tilage and a posterior part laid down originally in the crista
sellaris as a replacing bone.

The vomer (v.) in the stage modelled is an unpaired bone hav-
ing the form of a shallow trough which lies with its dorsal concave
face ventral to the anterior end of the septum interorbitale and
separated from it by a thin layer of connective tissue. In form
the vomer tapers slightly in a horizontal plane toward the ante-
rior end and also grows thicker dorso-ventrally.

In an embryo having a carapace length of 8.5 mm. the vomer
shows clearly its paired nature, being represented by a pair of
thin lamellae set at an angle to each other so that they fit about
the ventral edge of the septum interorbitale between the pos-
terior opening of the ductus naso-pharyngeus behind and the
cartilago paraseptalis in front. In an embryo having a cara-
pace length of 13.5 mm. the vomer extends relatively further
posteriorly than in the stage modelled and its lateral margins
lie close to the median margins of the two palatina which, however,
come to lie rather dorsal to the vomer.

The investing bones of the adult lower jaw are all present
in the stage modelled except the complementare which appears
first in an embryo having a carapace length of 13.5 mm. They
may be designated as the dentale, angulare, supra-angulare, and
goniale. An operculare is lacking as in the adult. The arti-

754 B. W. KUNKEL

culare, an ossification of the proximal end of Meckel’s cartilage,
is not apparent in an Emys embryo having a carapace length of
11 mm.

The dentale (d.) has the form of a long, slender plate exhibiting
in general a concave medial and convex lateral surface. It is
provided with flattened flanges extending both laterally and
medially from the dorsal margin. It lies lateral to Meckel’s
cartilage for almost the entire distance from the anterior, distal
end to the fovea articularis, overlapping with its posterior end
the anterior end of the supra-angulare so that the latter is inter-
posed between the dentale and Meckel’s cartilage.

By reason of the medial projection of the dorsal margin of
the dentale, this bone comes to arch in dorsally the suleus pri-
mordialis which accommodates Meckel’s cartilage. Beneath the
lateral flange of the dentale extends the portio alveolaris inferior
of the ramus mandibularis n. trigemini which gains access to this
groove (sulcus alveolaris inferior) from the sulcus primordialis
through the foramen canalis alveolaris inferioris. This foramen
is situated near the posterior end of the dentale immediately
ventral to the thickened dorsal margin. In older embryos the
suleus alveolaris inferior becomes converted into the closed
canalis alveolaris inferior by the further rolling over of the lateral
margin of the dorsal flange.

The supra-angulare (sa.) is a small elongate plate of bone which
lies along the lateral aspect of Meckel’s cartilage in its proximal
portion. It exhibits a slender posterior projection which lies
along the lateral side of the fovea articularis. The anterior third
of the supra-angulare at this stage is overlapped laterally by the
posterior end of the dentale. The ramus recurrens cutaneus
mandibulae from the mandibular branch of the trigeminus
passes obliquely backward along the dorsal margin of the supra-
angulare and somewhat medial to it.

The angulare (a.) is a long slender plate of bone lying on the
ventral side of the posterior end of Meckel’s cartilage in the
region of the supra-angulare. It forms, accordingly, the floor of
the sulcus primordialis in its posterior part. It is much shorter
than the supra-angulare and extends neither as far anteriorly

DEVELOPMENT OF THE SKULL OF EMYS Caa

nor posteriorly as the latter. Its antero-lateral margin lies in
close proximity to the ventral margin of the dentale.

The goniale (g.) is a thin, elongated plate of bone situated along
the median aspect of Meckel’s cartilage, from the posterior end
of the latter for one-third of its length. It exhibits a convex
median and a concave lateral aspect to fit more closely about the
posterior end of Meckel’s cartilage medially. Its postero-dorsal
angle lies along the median aspect of the fovea articularis.

Of especial interest in connection with the goniale is the rela-
tion of the chorda tympani. In the embryo modelled the nerve
enters the goniale at its posterior extremity by means of the cana-
lis gonialis which extends forward for a short distance and opens
anteriorly on the lateral aspect of the goniale as a well marked
suleus so that the chorda tympani reaches the canalis primordialis
and extends further forward in this canal between the goniale
on the median side and Meckel’s cartilage on the lateral side.
In an older embryo (carapace length, 13.5 mm.) a canalis goni-
alis is wanting. Instead of this the chorda tympani is lodged
in a groove on the median side of the goniale. The nerve passes
obliquely forward and ventrally and for a short space les along
the ventral margin of the goniale. The latter then becomes
suddenly wider by a ventrally projecting extension which is
lateral to the nerve. For a considerable distance the chorda
tympani is thus separated from Meckel’s cartilage by the goniale.
A foramen finally allows the nerve to pass into the sulcus pri-
mordialis.

A third relationship of the chorda tympani and goniale was
encountered in an embryo having a carapace length of 16 mm.
In this individual the nerve entered the canalis gonialis from the
lateral aspect of the bone and both before and after entering the
canalis gonialis it lay in a deep groove communicating with the
suleus primordialis.

The complementare is not yet laid down in the embryo mod-
elled. In an embryo having a carapace 13.5 mm. long, however,
it is present as a triangular plate of bone lying dorsal to the goniale

with its anterior end somewhat further forward than that of the
goniale. It lies medially to the ramus mandibularis n. trigemini.

756 B. W. KUNKEL

SUMMARY

The principal points which I would emphasize in conclusion
are as follows:

1. The chondrocranium of Emys resembles closely that of
Dermochelys and Chelone and, in general features, is similar to
that of Lacerta, but is of heavier construction.

2. The planum basale is parachordal in position, surrounding
the chorda dorsalis on all sides.

3. The condylus occipitalis is annular in form, a central con-
cavity being present around the chorda, dorsalis.

4. The arcus occipitales do not reach the tectum posterius
distally so that the foramen occipitale magnum is not separated
completely from the fissura metotica.

5. There are three foramina spino-occipitalia present at the
stage modelled, the two anterior of which fuse later.

6. The tectum posterius is strongly developed, exhibiting a
processus ascendens of large size and a stout processus posterior
which limits the foramen occipitale magnum dorsally.

7. The fissura metotica becomes wide at its antero-ventral
end at which point the nervus vagus and vena jugularis leave
the cranial cavity.

8. The planum basale projects laterally beyond the lateral
portions of the occipital region as the crista inferior which thereby
forms the ventral side of the sulcus supracristularis.

9. The fenestra cochleae opens into the sulcus supracristularis
at its anterior end (recessus scalae tympani) immediately in
front of the fissura metotica.

10. The fenestra cochleae is the posterior opening of the canalis
perilymphaticus which has a horizontal direction and opens into
the cavum cochleae on the median wall of the latter.

11. A canalis hypoperilymphaticus is found in embryos older
than that modelled, passing parallel and ventral to the canalis
perilymphaticus and communicating between the recessus scalae
tympani posteriorly and the cavum cochleae anteriorly. It con-
tains nothing but loose connective tissue. It develops anteriorly
from its posterior end.

DEVELOPMENT OF THE SKULL OF EMYS 757

12. The nervus glossopharyngeus passes through the cavity
of the otic capsule in the suleus glossopharyngeus which is situ-
ated on the posterior wall of the cavum cochleae immediately
ventral to the cavum ampullare posterius. The nerve perforates
the median capsular wall by means of the foramen glossopharyn-
geil internum and the outer wall by the foramen glossopharyngei
externum.

13. The cavum cochleae is developed in a posterior direction
so that the nervus glossopharyngeus passes through it.

14. The cavity of the otic capsule is not divided by a septum
intervestibulare as in Lacerta.

15. The median capsular wall is perforated anterior to the
foramen glossopharyngei internum by the foramina acustica
anterius and posterius and endolymphaticum; besides these are
one or two smaller foramina one of which is for a blood vessel.

16. The foramen facialis lies anterior to the otic capsule and
opens laterally into a depression, the fovea geniculi.

17. The foramen abducentis passes horizontally forward through
the base of the pila prootica.

18. The fenestra prootica is large and unclosed dorsally because
of the absence of a taenia marginalis to unite the distal end of the
pila prootica and the otic capsule.

19. The fenestra hypophyseos is large and accommodates,
besides the hypophysis cerebri, the arteriae carotides internae as
they pass into the cavity of the cranium.

20. The fenestra metoptica, for the accommodation of the
nervi oculomotorius and trochlearis, is in the form of a narrow
slit and, like the fenestra prootica, is unclosed dorsally.

21. The planum basale projects laterally in the region of the
crista sellaris beyond the attachment of the otic capsule to form
the crista basipterygoidea which is apparently homologous with
the processus basipterygoideus and with which articulates a
rudimentary imperfectly chondrified cartilago articularis. The
ramus palatinus nervi facialis passes ventrally from the ganglion
geniculi a short distance behind the cartilago articularis and then
extends anteriorly on the median side of the cartilago articularis
in the sulcus cavernosus.

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

158 B. W. KUNKEL

22. The septum interorbitale is imperforate and is quite evi-
dently of double origin since in younger stages than that mod-
elled it is seen in cross section to have a distinct Y-form, the two
limbs of the Y being parallel to each other and then suddenly
diverging to form the planum supraseptale.

23. The fenestra optica is large and is situated in fog of the.
foramen ophthalmicum through which the arteria ophthalmica
passes into the region of the orbit of the eye. The two foramina
ophthalmica are separated from each other by the subiculum
infundibuli.

24. The septum interorbitale passes without interruption for-
ward into the septum nasi which divides the ethmoidal region
into two symmetrical halves.

25. The olfactory capsule is of compact form and is composed
of continuous plates of cartilage not separated by extensive fora-
mina.

26. The capsule is bent somewhat ventrally as is indicated
by the plane of the fenestra olfactoria which is not horizontal,
as in Lacerta, but inclines downward in front; and also by the
fact that the fenestra basalis is situated below the level of the
ventral margin of the septum interorbitale.

27. The olfactory capsule is connected with the septum pos-
teriorly only by the slender commissurae spheno-ethmoidales
which bound the fenestrae olfactoriae laterally and separate
them from the fissurae orbitonasales.

28. The fissura orbitonasalis extends the entire length of the
posterior cupula of the capsule, as in the Lacertilia, the planum
antorbitale being entirely separated from the septum.

29. The nervus ethmoidalis passes into the capsule from the
orbit through the fissura orbitonasalis.

30. The plane of the fenestra narina is transverse so that the
fenestra faces directly anteriorly.

31. The lamina transversalis anterior is represented by the
portion of the floor of the capsule in front of the foraman prae-
palatinum.

32. The shortness of the cartilago paraseptalis of the stage
modelled is a secondary condition, since in earlier stages the fora-

DEVELOPMENT OF THE SKULL OF EMYS 759

men praepalatinum is not closed posteriorly by a fusion of the
cartilago paraseptalis with the septum, but extends freely pos-
teriorly for a considerable distance.

33. The septum nasi in the stage modelled is continuous with
the solum nasi except in the region of the foramen praepalatinum.

34. The cartilago paraseptalis, limiting the fenestra basalis
medially, does not fuse with the planum antorbitale dorsally so
that in the stage modelled the fenestra basalis is not completely
enclosed with cartilage.

35. A rudimentary concha is present in the form of a shallow
longitudinal infolding of the paries nasi. In older stages this
is increased in height by a thickening of the wall which projects
in a medial direction.

36. The septum nasi is thickened by means of a longitudinal
crest, the crista longitudinalis septi, which separates the pars
respiratoria from the pars olfactoria of the olfactory sac.

37. A slender rod, the pila supraglandularis, attached poste-
riorly to the septum in the region of the foramen praepalatinum,
extends forward parallel to the crista longitudinalis septi above
the glandula nasalis media and ends freely in front a short dis-

* tanee behind the level of the fenestra narina.

38. The septum nasi is provided with a small fenestra very
near its dorsal margin in the region of the fenestra olfactoria.

39. The prenasal cartilage is present as a median ridge on the
ventral side of the anterior part of the capsule.

40. A small foramen epiphaniale is present.

41. The palatoquadratum is entirely separate from the cra-
nium and exhibits a distinct processus pterygoideus which in turn
supports a processus ascendens which is the homologue of the
‘columella’ of the Lacertilia.

42. The pars quadrata of the palatoquadratum may be differ-
entiated into a large hollow pars mastoidea, enclosing an exten-
sion of the tympanic cavity, and the pars articularis.

43. The posterior margin of the quadratum is deeply incised
by the incisura columellae for the accommodation of the columella
auris.

760 B. W. KUNKEL

44. The columella auris originates wholly external to the otic
capsule and is made up of stapes and extracolumella. The foot
plate occupies the fenestra vestibuli and rests upon the crista
substapedialis; the stalk of the columella is slender and slightly
sinuous; the extracolumella is mushroom-shaped.

45. A distinet processus interhyalis extends medially and ven-
trally from the ventral angle of the extracolumella.

46. The ramus hyomandibularis nervi facialis extends pos-
teriorly dorsal to the columella and at the level of the latter gives
off the chorda tympani in a lateral direction. This nerve then
curves anteriorly, crossing the columella dorsally, and bends
ventrally in front of the Eustachian tube.

47. Meckel’s cartilage is strongly developed. The two car-
tilages meet by a strong symphysis anteriorly. The processus
retroarticularis is poorly developed.

48. The corpus hyale is pentagonal, with the apex directed
anteriorly. The short cornua hyalia are not segmented from
the corpus. The cornua branchialia primum and secundum are
both large and segmented from the corpus. A separate epibran-
chiale primum is present.

49. The processus lingualis arises as a pair of anteriorly directed °
rods from the anterior end of the corpus hyale, but at the stage
modelled the two have grown together.

50. The cartilago entoglossalis appears for the first time at
a later stage than that modelled.

51. The squamosum is a thin plate of bone lying parallel to
the outer side of the quadratum.

52. The quadrato-jugale is not laid down in the embryo till
a later stage than that at which most of the investing bones
are laid down.

53. The zygomaticum, together with the postfrontale, com-
pletes the orbit of the eye posteriorly.

54. The maxillare, as in the adult, exhibits a processus palati-
nus,-a processus alveolaris, and a processus praefrontalis. Its
posterior end is perforated by a foramen for the passage from
above of the ramus maxillaris nervi trigemini which continues
forward lateral to the processus alveolaris.

DEVELOPMENT OF THE SKULL OF EMYS 761

55. The praemaxillare is of small size, situated entirely ventral
to the olfactory capsule.

56. The parietale and frontale are wholly lateral in position
and do not begin at this stage to arch in the cranial cavity dor-
sally. The parietale exhibits a large processus inferior which
extends downward, lateral to the pila prootica, to come at length
into close relation with the free end of the processus ascendens
palatoquadrati. A cavum epiptericum is hereby formed.

57. The praefrontale. is exceedingly large and becomes very
thick at its anterior margin, dorsal to the olfactory capsule.

58. The pterygoideum exhibits a pronounced crista ptery-
goidea on its dorsal side which separates the fossa supraptery-
goidea from the sulcus cavernosus.

59. The palatinum and vomer complete the roof of the mouth
at the stage modelled although later a separate parasphenoideum
is present in the region of the fenestra hypophyseos. The vomer
is unpaired but is derived from a paired condition.

60. In the lower jaw all the investing bones of the adult except
the complementare are present at the stage modelled.

61. Along the lateral side of the dentale is the sulcus alveolaris
inferior which becomes converted into the canalis alveolaris
inferior of older embryos.

62. The goniale in the stage modelled is penetrated by the
eanalis gonialis in which the chorda tympani passes from the
posterior end of the goniale obliquely forward and laterally to
attain the canalis primordialis.

BIBLIOGRAPHY

No attempt is made in this bibliography to give a complete list of the works
on the morphology of the skull, as such a list is to be found in Gaupp, 1905 b, and
more specifically for the chelonian skull in Nick, 1912. Only the more recent and
important works which are referred to in the body of the paper are here mentioned.

BenperR, O. 1911 Ueber Herkunft und Entwickelung der Columella auris bei
Testudo graeca. Anat. Anz., Bd. 40, pp. 161-177.

Bosanus, L. H. 1819 Anatome Testudinis Europaeae. Vilnae, 1819-1821.

Fitatorr, D. 1906 Zur Frage iiber die Anlage des Knorpelschidels bei einigen
Wirbeltieren. Anat. Anz., Bd. 29, pp. 623-633.
1907 Die Metamerie des Kopfes von Emys lutaria. Zur Frage tiber
korrelative Entwicklung. Morph. Jahrb., Bd. 37, pp. 289-396.

762

Fucus, H.

B. W. KUNKEL

1906 Untersuchungen iiber die Entwicklung der Gehérknéchelchen,
des Squamosums, und des Kiefergelenkes der Siugetiere. Arch. f.
Anat. u. Physiol., Anat. Abt., Supplement-band, pp. 1-90.

1907 a Untersuchungen iiber Ontogenie und Phylogenie der Gaumen-
bildungen bei den Wirbeltieren. Erste Mitteilung. Ueber den Gau-
men der Schildkréten und seine Entwickelungsgeschichte. Zeitschr.
f. Morph. u. Anthrop., Bd. 10, pp. 409-4638.

1907 b Ueber die Entwickelung des Operculums der Urodelen und des
Distelidiums (‘Columella’ auris) einiger Reptilien. Verhandl. d. ana-
tom. Gesellsch., 21 Versamml., pp. 8-31.

1907 ec Ueber das Hyobranchialskelett von Emys lutaria und seine
Entwickelung. Anat. Anz., Bd. 31, pp. 33-39.

1910 Ueber das Pterygoid, Palatinum und Parasphenoid der quadrupe-
den, insbesondere der Reptilien und Saiugetiere, nebst einigen Betrach-
tungen iiber die Beziehungen zwischen Nerven und Skeletteilen. Anat.
Anz., Bd. 36, pp. 33-95.

Gaupp, E. 1905a Neue Deutungen auf dem Gebiete der Lehre vom Siugetier-

schidel. Anat. Anz., Bd. 27, pp. 273-310.

1905 b Die Entwickelung des Kopfskelettes. Hertwig’s Handbuch
der Entwickelungslehre, Bd. 3, Abt. 2, pp. 573-874.

1905 ¢ Die Nicht-Homologie des Unterkiefers in der Wirbeltierreihe.
Verhandl. d. anatom. Gesellsch., 19 Versamml., pp. 125-140.

1905 d Das Hyobranchialskelett der Wirbeltiere. Ergeb. d. Anat. u.
Entwickl., Bd. 14, pp. 808-1048.

1906 Ueber allgemeine und specielle Fragen aus der Lehre vom Kopf-
skelett der Wirbeltiere. Verhandl. d. anat. Gesellsch., 20 Versamml.,
pp. 21-68.

1907 a Ueber Entwickelung und Bau der beiden ersten Wirbel und
der Kopfgelenke von Echidna aculeata. Jena. Denkschr., Bd. 6, T. 2,
pp. 483-538. (Semon, Zool. Forschungsreisen, Bd. 3, T. 2).

1907 b Hauptergebnisse der an dem Semonschen Echidna-Material
vorgenommenen Untersuchung der Schidelentwickelung. Verhandl.
d. anat. Gesellsch., 21 Versamml., pp. 129-141.

1908 Zur Entwickelungsgeschichte und vergleichenden Morphologie
des Schidels von Echidna aculeata var. typica. Jena. Denkschr., Bd.
6, T. 2, pp. 541-788. (Semon, Zool. Forschungsreisen, Bd. 3, T. 2).
1910 a Das Lacrimale des Menschen und der Saéuger und seine mor-
phologische Bedeutung. Anat. Anz., Bd. 36, pp. 529-555.

1910 b Siiugerpterygoid und Echidnapterygoid nebst Bemerkungen
iiber das Siuger-palatinum und den Processus basipterygoideus. Anat.
Hefte, Bd. 42, pp. 311-431.

1911 a Ueber den N. trochlearis der Urodelen und iiber die Austritts-
stellen der Gehirnnerven aus dem Schidelraum im Allgemeinen.
Anat. Anz. Bd. 38, pp. 401-444.

1911 b Beitrige zur Kenntnis des Unterkiefers der Wirbeltiere. I.
Der Processus anterior (Folii) des Hammers der Siiuger und das Goniale
der Nichtsiuger. Anat Anz., Bd. 39, pp. 97-135. II. Die Zusammen-
setzung des Unterkiefers der Quadrupeden. Anat. Anz., Bd. 39, pp.
433-473.

DEVELOPMENT OF THE SKULL OF EMYS 763

Hasse, C. 1873 Das Gehérorgan der Schildkréten. Hasse’s Anatomische Stu-
dien, Bd. 1, pp. 225-329.

Horrmann, C. K. 1890 Schildkréten. Bronn’s Klassen und Ordnungen, Bd.
6, Abt. 3, pp. 1-442.

Howes, G. B. anp SwinnerTon, H. H. 1903 On the development of the skele-
ton of the tuatara, Sphenodon punctatus. Trans. Zool. Soc. London,
vol. 16, pp. 1-86.

Kunxet, B. W. 1911 Zur Entwickelungsgeschichte und vergleichenden Mor-
phologie des Schildkrétenschidels. Anat. Anz., Bd. 39, pp. 354-364.
1912 On a double fenestral structure in Emys. Anat. Ree., vol. 6,
pp. 267-280.

Monks, Sarau P., 1878 The columella and stapes in some North American
turtles. Proc. Amer. Phil. Soc., vol. 17, pp. 335-337.

Nick, L. 1912 Das Kopfskelet von Dermochelys coriacea L. Zool. Jahrb., Abt.
f. Anat., Bd. 33, pp. 1-238.

Noack, H. 1907 Ueber die Entwicklung des Mittelohres von Emys europaea
nebst Bemerkungen zur Neurologie dieser Schildkréte. Arch. f. mikr.
Anat., Bd. 69, pp. 457-490.

Oausui, K. 1911 Anatomische Studien an der japanischen dreikralligen Lippen-
schildkréte (Trionyx japanicus). Morph. Jahrb., Bd. 43, pp. 1-106.

Parker, W. K. 1880 On the development of the green turtle (Chelone viridis,
Schneid.) ‘Challenger Reports,’ Zool., vol. 1, part 5, pp. 1-58.

SEYDEL, O. 1896 Ueber die Nasenhéhle und das Jacobson’sche Organ der Land-
und Sumpf-schildkréten. Festschr.z.70. Geburtstag von Carl Gegen-
baur, Bd. 2, pp. 385-486.

SIEBENROCK,F. 1897. Das Kopfskelet der Schildkréten. Sitzungsber. d. kaiserl.
Akad. d. Wissensch.Wien; math.—naturwissensch.K]., Bd. 106, Abt. 1,
pp. 245-328.
1899 Ueber den Bau und die Entwicklung des Zungenbein-Apparates
der Schildkréten. Ann. d. k. k. naturhistor. Hofmuseums, Wien, Bd.
13, Heft 4, pp. 423-487. d

Tuynea, F. W. 1906 The Squamosal bone in tetrapodous vertebrata. Proc.
Boston Soc. Nat. Hist., vol. 32, pp. 887-425.

Van BEMMELEN,J.F. 1896 Bemerkurgen iiber den Schaidelbau von Dermochelys
coriacea. Gegenbaur’s Festschrift, Bd. 2, pp. 277-286.

Versiuys, J. 1909 Ein grosses Parasphenoid bei Dermochelys coriacea, Linn.

: Zool. Jahrb., Bd. 28, Anat. Abt., pp. 283-294.
1910 Bemerkungen zum Parasphenoid von Dermochelys. Anat. Anz.,
Bd. 36, pp. 487-495.

Voit, M. 1909 Das Primordialcranium des Kaninchens unter Beriicksichtigung
der Deckknochen. Anat. Hefte, Bd. 38, pp. 425-616.

764

B. W. KUNKEL

ABBREVIATIONS

The following list of abbreviations is used throughout in the explanation of

both text figures and plates.

a., angulare

a.c.i., arteria ecarotis interna

a.0., arcus occipitalis

a.p., ampulla posterior

c.a., cartilago articularis

c.d.a., cavum ampullare anterius
c.a.l., cavum ampullare laterale
c.a.p., cavum ampullare posterius
c.b., crista basipterygoidea

c.b.p., cornu branchiale primum
c.b.s., cornu branchiale secundum
c.c., cavum cochleae

c.d., chorda dorsalis

c.e., cartilago ectochoanalis

c.h., cartilago hypochiasmatica
c.hy., corpus hyale

c.i., erista inferior

c.l., erista longitudinalis septi

c.M., cartilago Meckelii

cn.h., cornu hyale

c.o., condylus oecipitalis

col., columella auris

c.p., cartilago paraseptalis

c.per., canalis perilymphaticus

c.pf., commissura praefacialis

c.pt., crista pterygoidea

cr.p., crista parotica

c.s., crista sellaris

c.s.a., canalis semicircularis anterior
c.sac., cavum sacculi

c.s.l., canalis semicircularis lateralis
c.s.p., canalis semicircularis posterior
c.sph., commissura spheno-ethmoidalis
c.8.8., Cavum sinus superioris utriculi
c.st., crista substapedialis

c.t., chorda tympani

c.v., cavum vestibuli

d., dentale

d.e., ductus endolymphaticus

d.p., ductus perilymphaticus

ec., extracolumella

f., frontale

f.a., foramen abducentis

f.a.a., foramen acusticum anterius

f.a.p., foramen acusticum posterius

f.b., fenestra basalis

f.b.p., fenestra basicranialis posterior

f.c., fenestra cochleae

f.e., foramen endolymphaticum

f.ep., foramen epiphaniale

f.f., foramen facialis

f.g., fovea genicularis .

f.g.e., foramen glossopharyngei exter-
num

f.g.i., foramen glossopharyngei inter-
pum

f.h., fenestra hypophyseos

f.m., fissura metotica

f.mp., fenestra metoptica

f.n., fenestra narina

f.o., foramen ophthalmicum

f.ol., fenestra olfactoria

f.o.m., foramen occipitale magnum

f.opt., fenestra optica

f.orb., fissura orbitonasalis

f.p., foramen praepalatinum

f.pr., fenestra prootica

f.s., fossa subarcuata

f.s.n., fenestra septi nasi

f.sp., foramen spino-occipitale

f.v., fenestra vestibuli

g., goniale

g-g-, ganglion geniculi

g.gl., ganglion glossopharyngei

g.s., ganglion semilunare

g.v., ganglion vagi

i.c., incisura columellae

l.t.a., lamina transversalis anterior

m., maxillare

0.c., capsula otica

p., parasphenoideum

p.a., processus ascendens palatoquad-
rati

pal., palatinum

p.ant., planum antorbitale

p.a.p., prominentia ampullaris pos-
terior

DEVELOPMENT OF THE

‘par., parietale

p.b., planum basale

pf., praefrontale

p.t., processus inferior

p.l., processus lingualis

p.met., pila metoptica

p.p., pila prootica

pl.s., planum supraseptale

pr.a., prominentia ampullaris lateralis

pr.b., crista basipterygoideus

pr.t., processus interhyalis

prm., praemaxillare

pr.p., processus posterior

pr.pt., processus pterygoideus

pr.s., processus supraparinus

p.s.a., prominentia semicircularis an-
terior

p.sg., pila supraglandularis

pst., postfrontale

pt., pterygoideum

q., quadratum

r.c., ramus communicans n. facialis
cum glossopharyngeo

r.h., ramus hyomandibularis n. facialis

r.o., ramus ophthalmicus n. trigemini

r.p., ramus palatinus n. facialis

s., Squamosum

sad., Supraangulare

sac., sacculus

SKULL OF EMYS 765

s.c., Sulcus cavernosus

sep., septum nasi

s.g., sulcus glossopharyngei

s.t., septum interorbitale

s.0., Sinus oculomotorius

s.s., sulcus supracristularis

$.8.a., septum semicirculare anterius
s.s.l., septum semicirculare laterale
S.S.p., septum semicirculare posterius
sub.i., subiculum infundibuli

t., trabecula cranii

t.n., tectum nasi

t.p., tectum posterius

u., utriculus

v., vomer

v.c.l., vena capitis lateralis

v.7., vena jugularis

z., zygomaticum

I, nervus olfactorius

IT, n. opticus

ITT, n. oculomotorius

IV, n. trochlearis

V, n. trigeminus

VI, n. abducens

VIT, n. facialis

VIII, n. acusticus

IX, n. glossopharyngeus

X,n. vagus

XIT, n. spino-occivitalis

Fig. 1 Cross section through the posterior part of the otic region of an embryo
having a carapace length of 1l mm. The section is slightly oblique, the left side
(reader’s) being anterior to the right side. > 15.

Fig. 2 Cross section through the posterior part of the otic region of the same
embryo as figure 1, and slightly anterior to it. On the right side the fissura meto-
tica (f.m.) and sulcus supracristularis (s.s.) are shown and on the left side the cana-
lis perilymphaticus opens into the cranial cavity by the extreme anterior end of
the fissura metotica. 15.

766

Fig. 3 Cross section through the otic region of the same embryo, showing the
nervus glossopharyngeus on the right side emerging from the foramen glosso-
pharyngei externum (f.g.e.) and on the left entering the otic capsule through the
foramen glossopharyngei internum (f.g.1.). The canalis perilymphaticus (c.per.)
is shown on the right side extending forward from the fissura metotica and on the
left side it is opening into the cavum cochleae on the median side of the latter.
x 15.

Fig. 4 Cross section through the same embryo, slightly in front of the previ-
ous figure. The relation of the squamosum (s.) external to the quadratum (q.)
is shown and the arteria carotis interna (a.c.7.) and ramus communicans n. facialis
cum glossopharyngeo (r.c.) lying ventral to the crista substapedialis. > 15.

767

Fig. 5 Cross section of the same embryo, through the otic region, showing, on
the left side, the chorda tympani (c.t.) passing from the n. facialis laterally on the
dorsal side of the columella auris. The processus interhyalis (pr.7.) of the extra-
columella is also seen. X 15. '

Fig. 6 Cross section of the same embryo through the anterior part of the otic
capsule showing, on the right side, the foramen facialis (f.f.), and on the left side
the fovea geniculi (f.g.) and the ramus palatinus n. facialis leaving the ganglion
geniculi. The chorda tympani (c.t.) is lying in a groove on the lateral surface of
the goniale. > 15.

768

.Fig. 7 Cross section of the same embryo through the anterior part of the otic
region in front of the previous figure, showing the nervus abducens immediately
before it enters the foramen abducentis and also the processus basipterygoideus
Cords) a6 LS:

Fig.8 Cross section of the same embryo through the posterior part of the orbito-
temporal region, showing the large ganglion semilunare (g.s.) lying in the fenes-
tra prootica; the nervus abducens (VJ) is seen on the right side lying in the fora-
men abducentis and on the left it is ventral to the base of the pila prootica (p.p.).
The cartilago articularis (c.a.) is to be seen attached to the ventral surface of the
processus basipterygoideus. X 15.

769

Y By
g Rigi
TS 3
pee FE 3 ”
K
° )
OS,
®
¢

m------ Ae

Fig. 9 Cross section through the anterior part of the orbital region of the same
embryo, showing the septum interorbitale (s.7.) and the posterior part of the olfac-
tory capsule, the planum antorbitale (p.ant.) and the cartilago ectochoanalis
(c.e.) alone being cut. X 15.

Fig. 10 Cross section through the posterior part of the ethmoidal region of the
same embryo showing the lack of continuity between the olfactory capsule pos-
teriorly and the septum interorbitale. X 15.

770

Fig. 11 Cross section through the ethmoid region a few sections in front of the
preceding figure, showing the separation of the capsular wall dorsally from the
septum, the thickening of the paries nasi to form a rudimentary concha, and the
fusion of the planum nasi and cartilago paraseptalis with the septum posterior to
the foramen praepalatinum which is seen in the following figure. 15.

Fig. 12 Cross section a short distance in front of the preceding figure, showing
the foramen praepalatinum (f.p.) and the commissura spheno-ethmoidalis (c.sph.).
Web.

Fig. 13 Cross section a short distance in front of the preceding figure showing
the crista longitudinalis septi (c.l.), the pila supraglandularis (p.sg.), and the
lamina terminalis anterior (l.t.a.). X 15.

Fig. 14 Cross section through the extreme anterior end of the olfactory capsule
of the same embryo, showing the glandula nasalis externa situated ventral to
the cavity of the capsule. X 15.

771

I4

Fig. 15 Cross section through the extreme posterior end of the occipital con-
dyle of an embryo slightly younger than that represented in the preceding figures,
showing the chorda dorsalis (c.d.) incompletely surrounded by the occipital con-
dyle which is here hypochordal in position. X 15.

Figs. 16 to 19 A series of four cross sections through the ethmoidal region of
an embryo having a carapace 13.5 mm. long. The series is from anterior to pos-
terior. In figure 16, on the right side, a cartilaginous nodule is lying in contact
by its median surface with the pila supraglandularis (p.sg.). In figure 17 the pila
supraglandularis has extended ventrally and fused with the lamina terminalis
anterior (l.t.a.). The thickening of the paries nasi to form a concha is greatest
in figure 19. X 15.

772

eS _ @ /
e a.
© % ~ / Son PU
2|

Fig. 20 Cross section through the posterior portion of the orbito-temporal
region of an embryo having a carapace 13.5 mm. long, showing the pterygoideum
(pt.) extending medially beneath the arteria carotis interna (a.c.7.) and the ramus
palatinus n. facialis (r.p.) to form the sulcus cavernosus of the adult pterygoi-
deum. A thin osseous lamella may also be seen along the ventral crest of the
planum basale which extends further forward and fuses with the parasphenoi-
deum in the region of the fenestra hypophyseos. X 15.

Fig. 21 Cross section through the otic region of an embryo having a carapace
length of 7 mm. showing the columella auris (col.) in its relation to the first vis-
ceral cleft (1) to the ramus hyomandibularis n. facialis (7.h.), and to the chorda
tympani (c.t.).. The processus interhyalis (pr.7.) in its relation to the cartilago
Meckelii (c.M.) is shown. X 15.

Fig. 21a A portion of figure 21 more highly magnified. X 35.

773

JOURNAL OF MORPHOLOGY, VOL. 23, NO. 4

Fig. 22 Cross section through the anterior part of the orbito-temporal region
of an embryo having a carapace 7 mm. long, showing the septum interorbitale
much thicker from side to side than in an older embryo. 15.

Fig. 28 Cross section through the same embryo as the preceding, somewhat
further anterior than the previous figure, showing the septum interorbitale (s.7.)
made up of two parallel plates which later become pressed together on the middle
line. X 15.

774

DEVELOPMENT OF THE SKULL OF EMYS PLATE 1
B. W. KUNKEL

EXPLANATION OF FIGURE

24 Dorsal view of a model of the chondrocranium of an embryo having a cara-
pace length of 1lmm. The membrane bones of the right side only are represented.
x 20

775

PLATE 2 DEVELOPMENT OF THE SKULLZOF EMYS
B. W. KUNKEL

EXPLANATION OF FIGURE

25 Ventral view of the same model. X 20
776

PLATE 3

DEVELOPMENT OF THE SKULL OF EMYS

B. W. KUNKEL

‘0G

x

‘oPqrpuvul

oY} OS[R SUIMOYS ‘opts JUSTE oY} UWOAT

GHYOSIt AO NOILVNV1d Xl

Jopour ourRs OY} JO MOLA 97

777

DEVELOPMENT OF THE SKULL OF EMYS

PLATE 4

B. W. KUNKEL

velayea ==

"0G X

‘OpIS Jo] OY} WOT JOpOUL OUTS OY} JO MoTA 2

GHuUooId AO NOILVNVIdxXa

778

DEVELOPMENT OF THE SKULL OF EMYS PLATE 5
B. W. KUNKEL

EXPLANATION OF FIGURES

28 View from left side of the same model showing the occipital and otic regions
only, with the palatoquadratum removed in order to expose the lateral surface of
the otic capsule. X 20.

29 Lateral view of a model of the cavity of the right otic capsule of the same
embryo as the above. X 24.

TUS)

PLATE 6 DEVELOPMENT OF THE SKULL OF EMYS
B. W. KUNKEL

ie: fol.
- f.orb.

EXPLANATION OF FIGURES

30 Ventral view of the olfactory capsule of an embryo having a carapace length
of 7 mm. showing the cartilago paraseptalis (c.p.) separated from the septum nasi
as far anterior as the lamina terminalis anterior (/.(.a.) and consequently the
foramen praepalatinum not enclosed posteriorly. X 20.

31 Posterior portion of the right otic capsule viewed from in front and slightly
from the median line, showing especially the groove in which the n. glossopharyn-
geus passes through the capsule between the foramen glossopharyngei internum
(f.g.i.) and the foramen glossopharyngei externum (/.g.e.), and also showing the
opening of the canalis perilymphaticus (c.per.) into the cochlear cavity. X 24.

780

SUBJECT AND-AUTHOR INDEX

CTINOBOLUS radians. Observations on the
A life-history of two rare ciliates, Spathidium
spathula and 349
Amphibian order Diplocaulia. The skull struec-
ture of Diplocaulus magnicornis Cope and ws
Arteries of Polyodon. The heart and 409
ARTELMEZ, Greorce W. The bilaterality of
the pigeon’s egg. A study in egg organiza-

tion from the first growth period of the
oocyte to the beginning of cleavage. 269
Bilaterality of the pigeon’s egg. A study in egg
organization from the first growth period of
the oocyte to the beginning of cleavage. ee

269

Blepharisma undulans St. The paedogamous con-
jugation of 667
Body size and cell size. 159

Co Gary N. The paedogamous con-
jugation of Blepharisma undulans St. 667

Cell size. Body size and 159
Chromodoris zebra Heilprin, a distinct species. 625
Chromosomes in cross-fertilized Echinoid eggs.
The behavior of the 17

of the Reduviidae. II. The nucleolus in
the young oocytes and origin of the ova in
Gelastocoris. I. A further study of the 331
Ciliates, Spathidium spathula and Actinobolus
radians. Observations on the life-history of
two rare 349
Criark, EvizaBpetH G., SMALLwoop, W. M. and
Chromodoris zebra Heilprin, a distinct species.

625

Clinostomum marginatum, a trematode parasite
of the frog, bass, and heron. On the structure

of 189
Conjugation of Blepharisma undulans St. The
paedogamous 667
ConkKLIN, Epwin G. Body size and cell size. 159

Corydalis cornutus L. The structure and meta-
morphosis of the fore-gut of 581
Cross-fertilized Echinoid eggs. The behavior of
the chromosomes in 17
Cryptobranchus allegheniensis, including com-
parisons with some other vertebrates. IT.
General embryonic and larval development,
with special reference to external features.
The embryology of 455
allegheniensis. The embryology of 61

ANFORTH, C. H. The heart and arteries of
Polyodon. 409

Degenerations in the secondary spermatogonia
of Desmognathus fusea. The 231
Desmognathus fusca. The degenerations in the
secondary spermatogonia of 231
Diplocaulus magnicornis Cope and the amphibian
order Diplocaulia. The skull structure of 31

CHINOID eggs. The behavior of the chro-
mosomes in cross-fertilized 17

Eggs. The behavior of the chromosomes in cross-
fertilized Echinoid 17

$1

Embryology of Cryptobranchus allezheniensis.
The 61
of Cryptobranchus allegheniensis, includ-
ing comparisons with some other vertebrates.
If. General embryonic and larval develop-
ment, with special reference to external fea-
tures. The 455

Emys lutaria. The development of the skull of 693

Epithelium of Turbellaria. The 255
ORE-GUT of Corydalis cornutus L. The
structure and metamorphosis of the 581

Fowl. Modifications in the testes of hybrids from
the guinea and the common 45

ELASTOCORIS. I. A further study of the
chromosomes of the Reduviidae. II. The
nucleolus in the young oocytes and origin

of the ova in 331

Guinea and the common fowl. Modifications in
the testes of hybrids from the 45
Guyer, MicHarLt F. Modifications in the testes
of hybrids from the guinea and the common

fowl.

EART and arteries of Polyodon. The 409
Heilprin, a distinct species. Chromodoris
zebra 625

Heredity in heterogeneous hybrids. 1
HirsH, Pavurtine, Kincspury, B. F. and The
degenerations in the secondary spermatogonia
of Desmognathus fusca. 231
Hybrids from the guinea and the common fowl.
Modifications in the testes of 45
Heredity in heterogeneous 1

INGSBURY, B. F. and Htrrsu, Pautine.
The degenerations in the secondary sperma-
togonia of Desmognathus fusca. 231

KunKEL, B. W. The development of the skull of

Emys lutaria. 693

ARVAL development, with special reference
to external features. The embryology of

Cryptobranehus allegheniensis, including
comparisons with some other vertebrates.

II. General embryonic and 455
Life-history of two rare ciliates, Spathidium
spathula and Actinobolus radians. Observa-
tions on the 349

Logs, JACQUES.
brids.

Heredity in heterogeneous hy-

The development of the skull of Emys
693

Lutaria.
A ATHESON, Roperr. The structure and
M metamorphosis of the fore-gut of Corydalis
cornutus L. 581
Metamorphosis of the fore-gut of Corydalis ecornu-
tus L. The structure and 581
Moopy, Jutta EB. Observations on the lifes hiss
tory of two rare ciliates, Spathidium spathula
and Actinobolus radians. 349
Moopiz, Roy L. The skull structure of Diplo-
caulus magnicornis Cope and the amphibian
order Diplocaulia. 31

woes AN oe
; ‘ : ie, : ye p ere ‘
; oom ‘i aac ¢ ‘ - : connie
F +‘ - J a « y ;
al . . . iy , s
. > @ & ar as
5! a4. > ' ¥ ~ 5 ¥ ‘
782 OS : Wr ge Bet
* | | i eee
UCLEOLUS in the young oocytes and origin | KULL of Emys Witata: The development, of s i #3 yp
of nie oe 2 ieee E nh ae + VS) thee Oa OA i : dae
study of the chromosomes of the edu- :
> ———structure of Diploeaulns ‘magnicornis. Co ae en
viidae. I, The _ 881. sie and the prope pin onger pegs Ey The me Se E
E MALLWOOD, . and CiarKk, Evizaperu G. 6 SE cece
4 OOCYTE OS and origin of the ova in Gelastocoris. ae « ang rah
ae é aL TPA tisthar gindwaat the chipmqsomestor as Chromodoris zebra Wy ra dlsiinths poo a Not 2
: the Reduviidae. II. The nucleolus in the wawraat Rena The embryology of: Crypto- a; pare
young ~ 331 branchus all i ae . : Poe a
» Oocyte to the beginning of cleavage. The bilater- The Subeelory of CO ate! ee uni
ality of the pigeon’segg. A study in egg organ- gheniensis, including comparisons with some rv STE S
ization from the first growth period of the 269 other vertebrates. II. General embryonic nae “f
Osporn, Henry Lestre. On the structure of and larval development, with special refer-
: Clinostomum marginatum, a trematode para- ence to external features. : ‘
% _ site of the frog, bass, and heron, 189 Spathidium spathula and Actinobolus radians. :
ae Ova in Gelastocoris. I. A further study of the Observations on the life-history of two rare ;
a PY é chromosomes of the Reduviidae. II. The — - ciliates, _ Me gt ae
nucleolus in the young CoEyyes and origin of — §permatogonia of Desnivhathad fusea, The de ese ae,
< . ‘ the 331 2 tee an in the oephg o at a ae G2
a AEDOGAMOUS _ Structure and metamor uae of the fore-gut_o 2 *,° bs
~ conjugation of Blepharisma FA it
roe. P undulans St. The ~ el 667 ” Cosa atts Lt ie -* ae “dh Bs ae
Parasite of the frog, bass, and heron. rent the 4 Pe
structure of Clinostomum marginatum, ENNENT, Daviv H. The behavior of the :
trematode : 89 chromosomes, in Bes Cae Echinoid Me
* PayNE, FERNANDUS. oy. ‘x further study of the ; eggs. i y 1740
‘i chromosomes of the Reduviidae. TI. The Testes of hybrids Rate the guinea and the common
nucleolus in the young oocytes and origin of — fowl. Modifications in the _ 45
u sae ‘ the ova in Gelastocoris. 331 Trematode parasite of the frog, bass, and ee
ry Pipe _ Pigeon’s egg. A study in egg’ organization from On the structure of Clinostomum eT a :
Me: 5 ee the first growth period of the oocyte to the," 189 ae
yt has ae? beginning of cleavage. ae Te routy. ae. ~ Turbellaria. The neg of | ; :
+ Ae : *
We 20 3 Polyodon. The heart and arteries oe ar ¥
: ; Primitive xepéiles,. Cae ake a Stam yee: S. W. "Primitive: reptiles. a
OG =

roe SueuUg ney ES ae the peels in - the as

young ooc and origin of the ova in

Gelastocoris. I. A further ‘study of the — oa G, R. T The snttom of Turbelt =
chromosomes of oe ae e 2 2931 Yess ‘ 255

Reptiles. Primitive oe 637 cy yiee Sp ME iene See ee etre ‘rive

»!

ts) ‘2 Le

rele

"e
ra

“a

Pd!

2 ~

iy ak ae

mer “iat Sate

LAGS
{

f

ee ~ a — mn a-<- ~ = - ; = = : > ae “
: ~ —_ — 3 —engee - - r< ¥ , . ons Ser >) ep 2 5 = FP —" a= — > - ~~ s ——— ov x