m ...

Bgg : ; . , :

iHcmoivs from tl)c Biological Caboratorg

OF THE

JOHNS HOPKINS UNIVERSITY

IV

SELECTED

x

MORPHOLOGICAL MONOGRAPHS

MAr

BIOLOG
LABO

(

OS HOLE, MASS.
H. 0. I.

BALTIMORE

THE JOHNS HOPKINS PRESS
1900

PRINTED BY
THE FRIEDENWAI.D COMPANY

BALTIMORE. MDM U. S. A.

CONTENTS.

I. The Cubomedusse. By FRANKLIN STORY CONANT.
II. Synapta Vivipara. By HUBERT LYMAN CLARK.

III. The Anatomy, Habits, and Embryology of Yoldia Limatula. By

GILMAN ARTHUR DREW.

IV. Physiology and Histology of the Cubomedusse, including Dr. F. S.

Conant's Notes on the Physiology. By EDMUND WILLIAM
BERGER,

V. Ophiura Brevispina. By CASWELL GRAVE.

VI. The Anatomy and Development of Cassiopea Xamachana. By
ROBERT PAYNE BIGELOW.

Jttcmoivs from tl)c Biological Cnboratory

OF THE

JOHNS HOPKINS UNIVERSITY
IV, 1

WILLIAM K. BROOKS, EDITOR

THE CUBOMEDUS^:

A DISSERTATION PRESENTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY, IN THE

JOHNS HOPKINS UNIVERSITY, 1897

BY

FRANKLIN STORY CONANT

A MEMORIAL VOLUME

BALTIMORE

THE JOHNS HOPKINS PRESS
1898

PRINTED BY

THE FRIEDENWALD COMPANY

BALTIMORE, MD. , U.S.A.

•

FRANKLIN STORY CONANT

/

SEPTEMBER 21, 1870— SEPTEMBER 13, 1897

A BIOGRAPHICAL SKETCH

This Treatise is printed after the author's death, as a Memorial by
his friends, fellow-students and instructors, with the aid of the Johns
Hopkins University. It consists of his Dissertation, reprinted from the
copy which was accepted by this University at his examination for the
degree of Doctor of Philosophy in June, 1897.

As he had made many notes on the embryology of the Cubomedusae,
and had hoped to complete and publish them together with an account
of physiological experiments with these medusae, he had described the
Dissertation on the title-page as Part I, Systematic and Anatomical, and
he went to Jamaica immediately after his examination to continue his
studies and to procure new material, and he there lost his life.

FRANKLIN STORY CON ANT was born in Boston on September 21, 1870,
and he died in Boston on September 13, 1897, a few days after his arrival
from Jamaica, where he had contracted yellow fever through self-sacri-
ficing devotion to others.

He was educated in the public schools of New England ; at the
University of South Carolina ; at Williams College, where he received
the degree of Bachelor of Arts in 1893 ; and in the Johns Hopkins
University, where he received the degree of Doctor of Philosophy in 1897,
and where he was appointed a Fellow in 1896 and Adain T. Bruce Fellow
in 1897.

Most of his instructors have told us that they quickly discovered that
Conant was a young man of unusual intelligence and energy and upright-
ness, and as his education progressed he secured the esteem and the
affectionate interest of all who had him in charge, so that they continued
to watch his career with increasing pride and satisfaction.

He entered the Johns Hopkins University in the spring of 1894, and
at once joined the party of students in zoology who were working, under
my direction, in the marine laboratory of the University at Beaufort,
North Carolina ; and from that time until his death he devoted himself
continually, without interruption, to his chosen subject — spending his
winters in the laboratory in Baltimore, and devoting his summers to out-
of-door studies at Beaufort and at Wood's Holl, and in Jamaica.

It is as a student and not as an investigator that we must remember
Conant, for most of his time was given to reading and study on subjects
of general educational value ; although he had begun, before his death,
to make original contributions to science and to demonstrate his ability
to think and work on independent lines.

His study of the Chaetognaths was undertaken only for the purpose
of verifying the account of their anatomy and development in the text
books, but it soon showed the presence at Beaufort of several undescribed
species. Without interrupting his more general studies, he employed
his odd moments for three years in their systematic analysis, and at
last published two papers, " Description of Two New Chaetognaths,'1 and
" Notes on the Chaetognaths," which show notable power of close and

X FRANKLIN STORY CONANT.

accurate observation and of exact description ; and, while short, are
valuable contributions to our knowledge of this widely distributed but
difficult group.

As he appreciated the value to one who has devoted himself to zoology
of thorough acquaintance with physiological problems and the means for
solving them, he wished, after he had completed his general course in
physiology, to attempt original research in this field ; and, at the sugges-
tion of Professor Howell, he, in company with H. L. Clark, his fellow
student, undertook and successfully completed an investigation of which
Professor Howell gives the following account :

In connection with Mr. H. L. Clark, Mr. Conant undertook to investigate the
character of the nervous control of the heart beat in decapod crustaceans. They
selected the common edible crab, Callinectes hastatus, and made a series of most
careful experiments and dissections which resulted in proving the existence of one
inhibitory nerve and two accelerator nerves passing to the heart on each side from
the thoracic ganglion. They not only demonstrated the physiological reaction of
these nerves, but traced out successfully their anatomical course from the ganglion to
the pericardial plexus. It seemed hardly probable from an a priori standpoint that in
an animal like the crab there should be any necessity for an elaborate nervous mecha-
nism to regulate the beat of the heart, but their experiments placed the matter
beyond any doubt, and have since served to call attention to this animal as a promis-
ing organism for the study of some of the fundamental problems in the physiology of
the heart. As compared with previous work upon the same subject it may be said
that their experiments are the most definite and successful that have yet been made.

His chief completed work, the Dissertation 011 The Cubomedusse, is
here printed ; and through it the reader who did not know Conant must
decide whether he was well fitted, by training and by natural endow-
ments, for advancing knowledge. I myself felt confident that the career
on which he had entered would be full of usefulness and honor. I was
delighted when he was appointed to the Adam T. Bruce Fellowship, for
I had discovered that he was rapidly becoming an inspiring influence
among his fellow students in the laboratory, and I had hoped that we
might have him among us for many years, and that we might enjoy and
profit by the riper fruits of his more mature labors.

Immediately after his examination for the degree of Doctor of
Philosophy in June, 1897, he set out for Jamaica to continue his
studies at the laboratory which this University had established for the
summer at Port Antonio, and he there worked for nearly three months
on the development, and on the physiology of the sense-organs, of the
Cubomedusas.

His notes and specimens are so complete that I hope it will be
possible to complete in Baltimore, at an early day, the work which he
had expected to carry on this year.

FRANKLIN STORY CONANT. xi

After the sudden and alarming death of the director of the expedi-
tion, Dr. J. E. Humphrey, Conant took the burden of responsibility upon
himself, and while he fully appreciated his own great danger, he devoted
himself calmly and methodically to the service of others who, in their
afflictions, needed his help, and he fell in the path of duty, where he had
always walked, leaving behind him a clear and simple account of all the
business of the laboratory and of his scientific work, and of his own
affairs, complete to the day before his death.

Immediately after the opening of the University in October his
friends and companions and instructors assembled to express the sorrow
with which they had heard the sad news of his death, and to record their
love and esteem for the generous, warm-hearted friend who in all the
relations of life had proved himself so worthy of their affectionate
remembrance. At this meeting those who had worked at his side in our
laboratories recalled his steadfast earnestness in the pursuit of knowl-
edge, and the encouragement they had found in his bright example;
while those who had been his instructors spoke of him as one who had
bettered their instruction and enriched all that he undertook by sound
and valuable observations and reflections. While all united in mourning
the untimely loss of one who had shown such rich promise of a life full
of usefulness and honor and distinction, it was pointed out with pride
that his end was worthy of one who had devoted it to the fearless pursuit
of truth, and to generous self-sacrifice and noble devotion to others ; and
it was resolved, " That we prize the lesson of the noble life and death of
Franklin Story Conant."

TABLE OF CONTENTS.

PAGE

INTRODUCTION 1

PARTI: SYSTEMATIC 3

Family I: CHARYBDEID^E 3

Charybdea Xaymacana 4

II: CHIRODROPID.E 4

" III : TRIPEDALID^ 5

Tripedalia cystophora 5

PART II : GENERAL DESCRIPTION OF THE ANATOMY OF THE CUBO-

MEDUS^E 7

A. CHARYBDEA XAYMACANA 7

a. Environment and Habit of Life 7

b. External Anatomy 8

2. Form of Bell 8

3. Pedalia 8

4. Sensory Clubs 9

5. The Bell Cavity and its Structures 10

(a) Proboscis 11

(b) Suspensoria, or Mesogonia 11

(c) Interradial funnels, or funnel cavities 11

(d) Velarium 12

(e) Frenula 12

(f ) Musculature 12

(g) Nerve ring 13

c. Internal Anatomy 13

6. Stomach 13

7. Phacelli 14

8. Peripheral Part of the Gastro-Vascular System 14

(a) Stomach Pockets (Valves and Mesogonial Pockets) 14

(b) Marginal Pockets 17

(c) Canals of the Sensory Clubs and Tentacles 17

9. Reproductive Organs 19

10. Floating and Wandering Cells 20

XVI CONTENTS.

PAGE

B. TRIPEDALIA CYSTOPHORA 22

a. Habitat 22

b. External Anatomy 23

c. Internal Anatomy 24

PART III: DESCRIPTION OF SPECIAL PARTS OF THE ANATOMY 27

A. VASCULAR LAMELLAE 27

B. NERVOUS SYSTEM 37

LITERATURE 57

TABLE OF REFERENCE LETTERS 58

DESCRIPTION OF FIGURES 60

VITA. 63

INTRODUCTION.

Jelly-fish offer to the lover of natural history an inexhaustible store
of beauty and attractiveness. One who has studied them finds within
him a ready echo to Haeckel's statement that when first he visited the
seacoast and was introduced to the enchanted world of marine life, none
of the forms that he then saw alive for the first time exercised so powerful
an attraction upon him as the Medusas. The writer counts it a rare
stroke of fortune that he was led to the study of a portion of the group
by the discovery of two new species of Cubornedusae in Kingston Harbor,
Jamaica, W. I., while he was with the Johns Hopkins Marine Laboratory
in June of 1896.

The Cubomedusae are of more than passing interest among jelly-fish,
both because of their comparative rarity and because of the high degree
of development attained by their nervous system. One fact alone suffices
to attract at once the attention of the student of comparative morphology
—that here among the lowly-organized Coelenterates we find an animal
with eyes composed of a cellular lens contained in a pigmented retinal
cup, in its essentials analogous to the vertebrate structure. Perhaps this
and other facts about the Cubomedusae would be more generally known,
had they not been to a certain extent hidden away in Claus's paper on
Charybdea marsupialis ('78), which, while a record of careful and accu-
rate work, is in many respects written and illustrated so obscurely that
it is very doubtful whether one could arrive at a clear understanding of
its meaning who was not pretty well acquainted with Charybdea before-
hand.

Before Claus's paper was received at this laboratory, H. V. Wilson
went over essentially the same ground upon a species of Chiropsalmus
taken at Beaufort, N. C. When the article on Charybdea marsupialis
appeared, however, the results were so similar that Wilson did not com-
plete for publication the careful notes and drawings he had made.

Haeckel's treatment of the Cubomedusas in his " System " ('79) in the

Challenger Report ('81) is much more lucid than Claus's ; but the extended

scope of his work and the imperfect preservation of his material prevented

a detailed investigation, and for a more complete and readily intelligible

l

2 JOHNS HOPKINS UNIVERSITY MOEPHOLOGICAL MONOGRAPHS.

account of the structure of the Cubomedusse a larger number of figures
is desirable.

In the foregoing facts lies whatever excuse is necessary for repeating
in the present paper much that has already seen print in one form or
another.

PART I: SYSTEMATIC.

It seems advisable first of all to establish the systematic position of
the two newly found species, Charybdea Xaymacana and Tripedalia
cystophora. Haeckel's classification, as given in his " System der
Medusen," is an excellent one and will be followed in this case. One
of the new species, however, will not classify under either of Haeckel's
two families, so that for it a new family has been formed and named the
Tripedalidse. In showing the systematic position of the two new forms,
an outline of Haeckel's classification will be given, so far as it concerns
our species, together with the additions that have been made necessary.

(Haeckel, 1877).

Characteristics : Acraspeda with four perradial sensory clubs which
contain an auditory club with endodermal otolith sac and one or several
eyes. Four interradial tentacles or groups of tentacles. Stomach with
four wide perradial rectangular pockets, which are separated by four
long and narrow interradial septa, or cathammal plates. Gonads in four
pairs, leaf -shaped, attached along one edge to the four interradial septa.
They belong to the subumbrella, and are developed from the endoderm
of the stomach pockets, so that they project freely into the spaces of the
pockets.

Family I : CHARYBDEA (Gegenbaur, 1856).

Cubomedusse with four simple interradial tentacles ; without mar-
ginal lobes in the velarium, but with eight marginal pockets ; without
pocket arms in the four stomach pockets.

Genus: Charybdea.

Charybdeida3 with four simple interradial tentacles with pedalia ;
with velarium suspended, with velar canals and four perradial frenula.
Stomach flat and low, without broad suspensoria. Four horizontal
groups of gastric filaments, simple or double, tuft or brush-shaped,
limited to the interradial corners of the stomach.

4 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Species : Charybdea Xaymacana (Fig. 1).

Bell a four-sided pyramid with the corners more rounded than
angular, yet not so rounded as to make the umbrella bell-shaped. The
sides of the pyramid parallel in the lower two-thirds of the bell, in the
upper third curving inward to form the truncation ; near the top a slight
horizontal constriction. Stomach flat and shallow. Proboscis with four
oral lobes, hanging down in bell cavity a distance of between one-third
and one-half the height of bell ; very sensitive and contractile, so that it
can be inverted into the stomach. The four phacelli epaulette-shaped,
springing from a single stalk. Distance of the sensory clubs from the
bell margin one-seventh or one-eighth the height of bell. Velarium in
breadth about one-seventh the diameter of the bell at its margin. Four
velar canals in each quadrant ; each canal forked at the ends, at times
with more than two branches. Pedalia flat, scalpel-shaped, between one-
third and one-half as long as the height of bell. The four tentacles, when
extended, at least eight times longer than the bell. Sexes separate.
Height of bell, 18-23 mm.; breadth, about 15 mm. (individuals with
mature reproductive elements) ; without pigment. Found at Port Hen-
derson, Kingston Harbor, Jamaica.

As may be seen from the above, C. Xaymacana differs only a little
from the C. marsupialis of the Mediterranean. Glaus mentions in the
latter a more or less well denned asymmetry of the bell, which he
connects with a supposed occasional attachment by the proboscis to algee.
In C. Xaymacana I never noticed but that the bell was perfectly sym-
metrical. C. Xaymacana is about two-thirds the size given by Glaus for
his examples of C. marsupialis, which were not then sexually mature.
It has 16 velar canals instead of 24 (32), as given by Haeckel, or 24 as
figured by Glaus. Difference in size and in number of velar canals are
essentially the characteristics upon which Haeckel founded his Chal-
lenger species, C. Murrayana.

Family II : CHIRODROPID.E (Haeckel, 1877).

Cubomedusse with four interradial groups of tentacles ; with sixteen
marginal pockets in the marginal lobes of the velarium, and with eight
pocket arms, belonging to the exumbrella, in the four stomach pockets.

This family is represented in American waters by a species of
Chiropsalmus, identified by H. V. Wilson as C. quadrumanus, found at
Beaufort, North Carolina.

F. S. CONANT ON THE CUBOMEDUS^E. 5

Family III : TRIPEDALID^E (1897).

Cubomedusse with four interradial groups of tentacles, each group
having three tentacles carried by three distinct pedalia ; without marginal
lobes in the velarium ; with sixteen marginal pockets ; without pocket
arms in the stomach pockets.

Genus : Tripedalia.

For the present the characteristics of family and genus must neces-
sarily be for the most part the same. The genus is distinguished by
having twelve tentacles in four interradial groups of three each ; velarium
suspended by four perradial f renula ; canals in the velarium ; stomach
projecting somewhat convexly into the bell cavity, with relatively well-
developed suspensoria ; four horizontal groups of gastric filaments, each
group brush-shaped, limited to the interradial corners of the stomach.

Species: Tripedalia cystophora (Fig. 17).

Shape of bell almost exactly that of a cube with rounded edges ; the
roof but little arched. The horizontal constriction commonly seen near
the top of the bell in the CubomedusaB not present. Proboscis with four
oral lobes ; hanging down in the bell cavity generally more than half the
depth of the cavity and at times even to the bell margin. In the gelatine
of the proboscis an irregular number (15-21) of sensory organs resem-
bling otocysts, from the presence of which comes the specific name.
Phacelli brush-shaped, composed of from seven to thirteen filaments
springing from a single stalk in each quadrant, or rarely from two sepa-
rate stalks in one of the quadrants. Distance of the sensory clubs from
the bell margin about one-fifth or one-fourth of the height of bell.
Breadth of velarium about one-sixth the diameter of bell at margin ;
with six velar canals in each quadrant; the canals simple, unforked.
Pedalia flattened, shaped like a slender knife blade, about half as long as
the height of the bell. Tentacles at greatest extension observed two and
a half times the length of pedalia. Sexes separate. Height of bell in
largest specimens (reproductive elements mature) eight or nine mm.
Breadth same as height or even greater. Color a light yellowish brown,
due in large part to eggs or embryos in the stomach pockets. The repro-
ductive organs especially prominent by reason of their similar color.
Found in Kingston Harbor, Jamaica.

It will be seen from the above that Tripedalia possesses two of the

6 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

characteristics of the Charybdeidaa and two of the Chirodropidae. The
family was named from the prominent feature of the arrangement of
the tentacles, in groups of three with separate pedalia. The small size
of T. cystophora is worthy of note in connection with the fact that of the
twenty species of Cubomedusas given by Haeckel in his " System " only
two are smaller than 20 mm. in height, and those are the two representa-
tives of Haeckel's genus Procharagma, the prototype form of the Cubo-
medusaB, without pedalia and without velarium. While Tripedalia has
both pedalia and velarium, it may be perhaps that its small size, taken
in connection with characteristics just about midway between the
Charybdeidas and the Chirodropidas, indicate that it is not a recently
acquired form of the Cubomedusas.

PAET II: GENERAL DESCRIPTION OF THE ANATOMY OF THE

CUBOMEDUS^.

A: CHARYBDEA XAYMACANA.
a. Environment and habit of life.

1. The Cubomedusse are generally believed to be inhabitants of deep
water which come to the surface only occasionally. Both of the Jamaica
species, however, were found at the surface of shallow water near the
shore, and only under these circumstances. Whether these were their
natural conditions, or whether the two forms were driven by some
chance from the deep ocean into the Harbor and there found their
surroundings secondarily congenial, so to speak, can be a matter of
conjecture only. C. Xaymacana was taken regularly a few yards
off-shore from a strip of sandy beach not ten minutes row from the
laboratory at Port Henderson. It was seen only in the morning before
the sea-breeze came in to roughen the water and to turn the region of its
placid feeding-ground into a dangerous lee-shore. Some of the speci-
mens taken contained in the stomach small fish so disproportionately
large in comparison with the stomach that they lay coiled up, head over-
lapping tail. The name Charybdea, then, from the Greek %apvfi<>ig (a gulf,
rapacious), seems to be no misnomer. It is worth mentioning that the
digestive juices left the nervous system of the fish intact, so that from
the stomach of a Charybdea could be obtained beautiful dissections, or
rather macerations, of the brain, cord, and lateral nerves of a small fish.

In size C. Xaymacana agrees very well with the average of the
genus. The four single tentacles characteristic of the genus are very
contractile, varying from two or three to six or seven inches in length,
and probably if measurements could be taken while the animal was
swimming freely about, the length would be found to be greater still.
Charybdea is a strong and active swimmer, and presents a very beau-
tiful appearance in its movements through the water, the quick, vigorous
pulsations contrasting sharply with the sluggish contractions seen in
most Scyphomedusae. With its tentacles streaming gracefully behind,
an actively swimming Charybdea presents a fanciful resemblance to a

8 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

comet or meteor. When an attempt is made to capture one, it will often
escape by going down into deeper water — as indeed do other jelly-fish.
Escape from observation is all the more easy by reason of the entire
absence of pigment excepting for the small amount in the sensory clubs.
The yellowish or brownish color usually stated as common in the Cubo-
medusae is nowhere present in C. Xaymacana.

b. External Anatomy.

2. Form of Bell. C. Xaymacana shows the typical division of the
external surface into four almost vertical perradial areas (Figs. 1-3, p),
separated by four stoutly arched interradial ribs or bands (Figs. 1-3, i).
These ribs thus play the part of corners to the Cubomedusan pyramid.
They are formed by the thickenings of the jelly of the exumbrella, and
serve to give the necessary strength to the four interradial corners, each
of which bears one of the four tentacles at its base. Each rib is further
divided into two longitudinal strips by a vertical furrow lying exactly in
the interradius (Fig. 2, ifr). The surface of the exumbrella is thus marked
by twelve longitudinal furrows, as seen in the same figure (2). Of these,
four are the interradial furrows just mentioned ; the other eight are the
adradial (afr) furrows, which set off the four perradial surfaces of the
pyramid from the four interradial ribs or bands of the corners, each of
which is again subdivided, as mentioned above, by the shallower inter-
radial furrows. Each interradial furrow ends above the base of the cor-
responding pedalium, at about the level of the sensory club ; each adra-
dial furrow diverges toward the perradius in the lower third of its course,
and thus with its companion furrow narrows down the perradial surface
of the pyramid in the lower part of the bell to an area of not much
greater width than the niches in which the sensory clubs lie. The pro-
jecting interradial corners are of course correspondingly enlarged in the
lower part of the bell, and in this way the contours of the surface are
changed from those figured in the view of the bell from above (Fig. 2)
to those of Fig. 3, which represents a view of the bell margin from below.

3. Pedalia. From the base of the interradial corner bands spring
the four pedalia (Fig. 1, pe), gelatinous appendages of the margin having
much the same shape as the blade of a scalpel. These in turn bear on
their distal ends, as direct continuations, the long, contractile, simple
tentacles. The relatively stiff pedalia have the same relation to the
flexible tentacles that a driver's whip-stock has to the long lash. In the
living animal the pedalia are found attached to the margin at an angle

F. S. CONANT ON THE CUBOMEDUS^E. 9

of about 45° with the longitudinal axis of the bell. In the preserved
specimens they are bent in toward the axis by the contraction of the
strong muscles at their base, in which position they are figured by Glaus
for C. marsupialis (78, Taf. I., Figs. 1 and 2).

The pedalia are in reality processes belonging to the subumbrella, as
will be shown in the section treating of the vascular lamella. They are
composed chiefly of gelatine covered with thin surface epithelium and
carrying within the gelatine the basal portion of the tentacle canals.
They have received various names at the hands of the writers. Gegen-
baur called them " Randblatter." Glaus gave them the name of "Schirm-
lappen," and incorrectly homologized them with the marginal lobes of
other Acraspeda. Claus's error was corrected by Haeckel, who termed
them "Pedalia" or "Gallertsockel," and homologized them with the
pedalia of the PeromedusaB. Besides furnishing a base of support for the
tentacles they may perhaps also serve as steering apparatus, a function
for which their thin blade-like form would be admirably adapted.

Internal to the base of each pedalium, between it and the velarium,
is found a funnel-shaped depression of the ectodermal surface. This is
shown in Fig. 5 (ft) in longitudinal section, and in cross-section in Fig. 16.
In the latter figure the epithelium of the outer wall of the funnel (mt) is
shown much thickened, the result of a stout development of muscle
fibres. These are the muscles that in the preserved specimens cause the
inward contraction of the pedalia referred to above.

4. Sensory Clubs (marginal bodies, rhopalia). In spite of their posi-
tion above the bell margin, the four sensory clubs, representing as they
do transformations of the four perradial tentacles, are properly classed
with the pedalia and interradial tentacles as appendages of the margin.
They lie protected in somewhat heart-shaped excavations or niches in
the perradial areas of the exumbrella. Each sensory niche is partially
roofed over by a covering scale, a hood-like projection from the exum-
brella. Below the covering scale the water has free access to the niche
and to the sensory club within it. The sensory club consists of a
hollow stock directly homologous with tentacle and canal, and a termi-
nal, knob-like swelling, the sensory portion proper. The latter con-
tains on its inner surface — the surface turned towards the bell cavity —
two complicated unpaired eyes with lens, retina, and pigment, lying one
above the other in the median line ; and at the sides of these, two pairs
of small, simple, pigmented, bilaterally symmetrical eye spots. At the
end of the club, that is, on its lowermost point, lies a sac that contains a
2

10 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

concretion and is usually considered auditory. The canal of the stalk is
directly continuous with the gastro-vascular system. In the swollen
knob of the sensory club it forms an ampulla-like terminal expansion.

As was pointed out by Glaus, the bottom of the sensory niche — by
bottom is meant the vertical wall that separates the space of the niche
from the bell cavity — is formed from the subumbrella only. This
arrangement of parts, apparently impossible for a structure so far
removed from the bell margin as the sensory niche, will be explained
more fully under the special topic of the vascular lamella?, or cathammal
plates. It is sufficient at this point to refer to Fig. 44, which shows the
shield-shaped area mapped out by a vascular lamella that connects the
endoderm of the stomach pocket with the ectoderm of the bottom of
the niche. By this the exumbrella is completely cut off from any part in
the formation of the bottom of the niche. Cross and vertical sections
through the niche (Figs. 39 and 37) help to a better understanding of
these relations. Since the base of the stalk of the sensory niche lies
within the ring of vascular lamella, the whole organ as well as the
bottom of the niche belongs to the subumbrella, and so in spite of its
position some distance upwards from the bell margin the sensory club is
very properly called a "marginal body" (Randkorper).

The epithelium of the sensory niche consists entirely of the flattened
ectodermal surface layer common to the whole exumbrella. No differen-
tiation suggestive of nervous function in addition to that of the sensory
clubs can be discovered, although it would be quite natural to expect to
find something of the sort, as intimated by Glaus ('78, p. 27).

It is worth while to mention again the fact that the eyes are directed
inwardly toward the cavity of the bell. The larger and lower of the two
median eyes looks into the bell cavity horizontally ; the smaller upper
eye is turned upward toward the region of the proboscis. This is in the
normal pendant position of the sensory club. The stalk, however, is very
flexible, and a range of other positions of the sense organs is possible,
although nothing was observed to suggest that such positions were
within the control of the animal. The eyes evidently have as their chief
function to receive impressions of what is going on inside the bell, not
outside. Perhaps the strongly biconvex, almost spherical lenses of the
median eyes also point to a focus on near and small objects.

5. The Bell Cavity and its Structures. In general, the bell cavity
repeats the external form of the bell, being almost cubical. In cross-
section it appears very nearly square with the angles in the interradii as

F. S. CONANT ON THE CUBOMEDUS^. 11

seen in the series of drawings that figure sections of the whole jelly-fish
at different levels (Figs. 6-16). Above, the bell cavity is roofed over by
the stomach ; below, it is open freely to the water, the opening being
narrowed somewhat by the diaphragm-like velarium (Fig. 3, v) ; the four
flat perradial sides are bounded by the walls of the four broad stomach
pockets, to be described when we come to the internal anatomy.

(a) The Proboscis. From the stomach there hangs down into the
bell cavity the proboscis or manubrium, which consists of a short funnel-
shaped stalk bearing on its distal end the four mouth lobes or lips. The
latter are somewhat broadly V-shaped processes lying in the perradii
with the convexity directed outwards, and with the concavity on the
inside forming the beginnings of four perradial furrows that are con-
tinued upwards to the stomach. The four furrows are shown in the
stalk of the proboscis in Fig. 11, which represents a section taken a little
above the level of the mouth lobes. The same cross-shaped section of the
stalk shows the four perradial prominences or ridges overlying the
furrows, which are the direct continuations of the four projecting mouth
lobes.

(b) The Suspensoria or Mesogonia. The stomach (leaving out of
consideration the proboscis) hangs down into the bell cavity as a slightly
sagging saucer-shaped roof (Figs. 4 and 5). In the four perradii it is
attached to the lateral walls of the subumbrella by four slenderly devel-
oped mesentery -like structures, the suspensoria or mesogonia. These are
simple ridges of gelatine, covered of course with the epithelium of the
bell cavity, which serve to keep the stomach in position much in the way
that a shelf is supported by brackets (Fig. 4, su). The suspensorium
accordingly has two parts, curved so as to lie at right angles with each
other : a vertical portion lying along the wall of the subumbrella, and a
horizontal which passes over from the vertical on to the basal wall of
the stomach. In Fig. 10 the suspensorium in each quadrant is shown
cut across just below the angle between the two parts, so that the two
appear in the section as projections on the wall of the stomach and on
the wall of the subumbrella.

(c) The Interradial Funnels or Funnel Cavities. It will be seen at
once that the four suspensoria serve as partitions to divide the upper
portion of the bell cavity, the part that lies between the stomach and the
lateral walls of the subumbrella, into four compartments. These com-
partments extend upwards in the four interradii like inverted funnels,
whence their name. In the series of cross-sections they can be traced

12 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

upwards with constantly diminishing area from the level of the suspen-
soria, Fig. 10 (if), to Fig. 6, which is taken very near the top of the bell.
Homologous structures exist in all the Scyphomedusse, and in some of
the LucernaridaB they are continued up even into the stalk of the
attached jelly-fish.

(d) The Velarium. Charybdea, like most of the Cubomedusas, pos-
sesses a velum-like structure around the opening of the bell cavity
(Fig. 3, v). The velarium is a thin muscular diaphragm, resembling the
true velum in position and essential structures, but differing from the
velum in its origin, and in the possession of diverticula from the gastro-
vascular system, the velar canals. Of these there are in C. Xaymacana
very regularly sixteen, four in each quadrant. Their outline is seen in
Fig. 3 to be forked with small irregular accessory processes. As for its
origin, the velarium of the CubomedusaB is commonly accounted to have
arisen by fusion of marginal lobes, as in the case of the velarium of the
DiscomedusaB. Pending decisive ontological evidence, the slight notches
in the four perradii seen in Fig. 3 may perhaps be taken as slight indica-
tions of a primitive unfused condition, but the question will be brought
up again when the vascular lamellae are discussed.

(e) The Frenula. Just as the stomach is attached to the walls of the
subumbrella in the four perradii by the suspensoria, so in the lower part
of the bell cavity the velarium is attached to the wall of the subumbrella
in the perradii by four structures similar to the suspensoria, the frenula
velarii. The frenula, like the suspensoria, resemble the brackets of a
shelf, with the difference that in the case of the frenula the bracket is
above the shelf, their purpose being evidently to keep the velarium stiff
against the outflow of water produced by the pulsations of the bell.
According to the greater need of strength in this case, we find the frenula
stouter, more buttress-like than the suspensoria. The gelatinous ridge
that gives them the necessary firmness is thickened so as to be triangular
in section, as shown in Fig. 16 (frn).

(f) Musculature. As is general in medusaB, the muscular system, so
far as known, is restricted to the subumbrella. It has a very simple
arrangement, consisting of a continuous sheet of circular (/. e. horizon-
tal) striated fibres, which is interrupted only in the four perradii by the
radially directed muscle fibres of the suspensoria and the frenula. In
each quadrant, between the muscle of the suspensorium above and that
of the frenulum below, in an area just internal to the sensory niche,
there lies a space free from muscle. This interruption of the muscle

F. S. CONANT ON THE CUBOMEDUS^. 13

layer is shown in Fig. 39. Under the head of musculature belonging to
the subumbrella must be included also the radial, or longitudinal muscles
at the bases of the pedalia, which were mentioned before (Fig. 16, mt}.
The mouth lobes and proboscis also are highly contractile and muscular.
(g) Nerve Ring. It is in the possession of a clearly denned nerve
ring that the Cubomedusas differ from all other ScyphomedusaB whose
nervous system has been carefully studied. The nerve ring shows very
plainly on the surface of the subumbrella as a well-defined clear streak.
Its course is zig-zag or festoon-like. In the interradii, at the basis of the
tentacles, it lies not far from the bell margin. In the perradii it rises to
the level of the sensory clubs. This very striking arrangement is under-
stood at once when it is remembered that the sensory clubs represent the
four perradial primary tentacles, and were originally situated on the
margin. When all the rest of the margin grew down and away from
the four sensory clubs, fusing below them to form the present intact
edge of the bell, the four portions of the nerve ring that lay in the
perradii were left at the level of the sensory clubs, and the originally
straight nerve ring was thus bent into a bow in each quadrant. The
finer structure of the nerve will be treated of in the special part to be
devoted to the nervous system.

c. Internal Anatomy.

6. Stomach. The shape of the stomach is approximately that of a
biconvex lens, as seen in Fig. 4, which represents a Charybdea cut in
halves longitudinally in the perradius. The lumen of the proboscis (the
buccal stomach according to Haeckel's terminology) communicates
directly by a funnel-shaped enlargement with the stomach proper, or
central stomach of Haeckel. The term basal stomach is carried over by
Haeckel from the Stauromedusee, where it has considerable significance,
to the Cubomedusaa, and applied to the upper part of the central stomach.
In the stalkless Cubomedusas, however, it has no significance so far as
actual structure goes, and our knowledge of the development of the
CubomedusaB is as yet too simple for us to say that the upper part of the
main stomach represents what remains of the basal stomach of an
earlier pedunculated stage.

The epithelium of the roof of the stomach is not specially differenti-
ated and apparently has little or no part in digestion. The epithelium of
the floor, on the other hand, is composed chiefly of very high and thickly
crowded columnar cells which are usually described as coarsely granular,

14 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

but under high powers appear to be filled with vacuoles surrounded by a
network of cell substance. Thickly interspersed among these columnar
cells are goblet cells filled with mucus. The floor is thrown into
numerous wrinkles by ridges in the supporting gelatine resulting in
increase of digestive surface. The four perradial grooves of the proboscis
are continued in the perradii along the floor of the stomach as four fairly
deep furrows, which lead directly to the gastric ostia and stomach
pockets — structures to be described presently. These furrows are lined
with crowded columnar cells, smaller and denser than the other cells of
the digestive epithelium, containing no granules and but little beside the
relatively large, compact, deeply staining nuclei. The furrows probably
represent special ciliated courses.

7. Phacelli. Lying in the four interradial corners of the stomach are
the four phacelli or tufts of gastral filaments to the number of thirty or
thirty-five in each tuft. The filaments are attached to a single stalk, like
the fringe of an epaulette or the hairs of a coarse brush. The stalk
bearing the filaments is an outgrowth of the lower wall of the stomach
just at the point where it fuses with the upper. The phacelli are there-
fore structures of the subumbrella, proof of which will be found under
the special topic of the vascular lamellae. The stalk, an indication of
which appears in sph. Fig. 6 (the section being a little below the axis of
the stalk, which lies horizontally), consists of a firm core of gelatine
covered with the high columnar epithelium of the floor of the stomach.
The filaments themselves are slender processes repeating the structure of
the stalk and having a central axis of gelatine for support covered with
glandular epithelium, which in the case of the filaments bears numerous
nettle cells. These processes are extremely contractile, and in the living-
animal show a continuous, slow, squirming movement like a mass of
worms. The section just referred to (Fig. 6) shows diagrammatically
three of these filaments (fpli) cut across in each quadrant.

8. Peripheral Part of the Gastro-vascular System. The proboscis and
stomach proper comprise the central part of the gastro-vascular system.
In direct communication with the central is a peripheral part composed
of pouches or pockets lying in the vertical sides of the cube-shaped bell,
just as the central stomach lies in its roof. The peripheral part may be
subdivided for convenience of description into the stomach pockets, the
marginal pockets, and the canals of the tentacles and sensory clubs.

(a) Stomach Pockets. These are four broad, thin pouches lying
between the exumbrella and the subumbrella in the four perradii (e. g.

F. S. CONANT ON THE CUBOMEDUS-ffi. 15

Fig. 9, sp) and separated from one another in the interradii merely by
four thin vertical strips of vascular lamella (ivl) or fusion between the two
endodermal surfaces of a primitively single undivided peripheral cavity.
The structure is exactly that which we should have if in a Hydromedusa,
for example Liriope (Trachomedusse), the four radial canals broadened
out and the intervening cathammal plates correspondingly narrowed,
until the relations in size were just reversed, and instead of four narrow
radial canals separated from one another by four broad cathammal
plates, we had four broad radial canals or pouches separated by four
narrow cathammal plates.

The stomach pockets communicate at their top with the central
stomach by means of four moderately large openings, the gastric ostia.
These are seen in a side view of the whole animal as triangular spaces
(Fig. 1, g, o.) near the top of the broad perradial sides. In Figures 7 and
8 they are seen in cross-sections, in Fig. 4 in vertical section.

The communication between the stomach and each stomach pocket
is guarded by a valve that can cut the one entirely off from the other.
The valve is simply the flexible lower margin of the gastric ostium, a thin
vertical fold of the floor of the stomach, semilunar in shape, just at the
point where it is passing over into the stomach pocket. A longitudinal
section, such as is shown in Fig. 4, gives the best idea of the form and
position of the valve that can be obtained from any simple section.
Internal to the valve is seen a depression of the stomach wall, almost
worthy to be called a pocket. The valve itself lies as a wall across the
end of this depression, obstructing a free course to the stomach pocket.
It will be seen at once that any pressure of fluids in the stomach against
this vertical wall, or valve, would serve only to press it against the inner
surface of the exumbrella, and thus effectually close the entrance into
the stomach pocket. Such a closure would both keep the juices of the
stomach from entering the pockets and the embryos in the pockets from
entering the stomach before the proper time.

The depression of the floor of the stomach just internal to the valve
may possibly be a structure of some morphological significance. In one
series of sections it was found that in two of the quadrants the depression
was deeper than that represented in Fig. 4, and extended perceptibly into
the outer or vertical portion of the suspensorium. Fig. 32 is a diagram
giving a vertical reconstruction in the perradius of the cross-sections in
which this deepened depression was noticed. Fig. 31 is a drawing (the
outline by camera lucida) of one of the cross-sections, through the lower-

16 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

most point of the depression. The figure gives the wall of the stomach
lined with high columnar epithelium (ens), and the wall of the stomach
pockets, with the suspensorium (su) connecting them. The section is
taken just above the broad angle that lies between the two parts of the
suspensorium, that is, in a plane parallel to the arrow a-b in Fig. 32, but
a little lower down. At the points to which the reference letter x (Fig.
31) refers are seen the first indications of the division into two parts, i. e.
of the apex of the angle. The next section or two lower down show the
relation seen in Fig. 10 (su). There can be no doubt in this case that the
depression or pocket lies in the outer vertical limb of the suspensorium.
It is the position that gives it at least the appearance of some mor-
phological significance. In two genera of Lucernaridas named and
described by Clark ('78), Halicyathus and Craterolophus, the mesogonia
or suspensoria in all four perradii contain broad pockets. These mesogo-
nial pockets in the Lucernaridae have given rise to considerable misun-
derstanding owing to the fact that in some forms the reproductive
organs bulge out from the stomach pockets in which they structurally
lie, and come to take up a secondary position in the walls of the mesogo-
nial pockets. The sections of Charybdea above referred to indicate that
among the Cubomedusas we may have the same structure reduced to its
lowest terms, and may be a feather's weight in favor of the view that the
Cubomedusae are descendants of an attached Lucernaria-like form.

Two more diagrams, Figs. 33 and 34, are added in order to give a
more complete understanding of a gastric ostium and its neighboring
structures, the rnesogonial pocket and the valve. Fig. 33 is a view of the
gastric ostium and valve from the stomach side, and represents the
appearance that would be given by a thick section through the arrow x-y in
Fig. 32, in a plane at right angles to the paper. The heavy lines outlining
the gastric ostium (enr and enfl) represent the place where the plane of the
section has cut across the epithelium of the roof of the stomach above the
ostium and the epithelium of the floor of the pocket-like depression
internal to the valve. The continuation of the two heavy lines in either
side of the ostium represents the i*egion where the roof and floor of the
stomach meet ; •/. e., the edge of the lens-shaped stomach. The semilunar
outline of the valve (rg) is shown by alight line just above the epithelium
of the depression. As is seen by the reference arrow in Fig. 32, the valve
lies a little external to the immediate plane of the section, and hence it is
that its inner surface is seen in Fig. 33 and not a section of it. The
vertical part of the suspensorium (su) is seen in section below the epithe-

F. S. CONANT ON THE CUBOMEDUS^E. 17

Hum of the depression. The reference numbers 1, 2, 3 and 4 denote the
same points in Figs. 32 and 33. Fig. 32 referred to Fig. 33 would lie in a
plane at right angles to the paper through the reference arrow x-v of the
latter figure.

Fig. 34 represents a horizontal section through the gastric ostium at
the level of the arrow a-b in Fig. 32, or arrow c-d in Fig. 33. The reference
numbers 5, 6 and 7, 8 denote similar points in the two figures 33 and 34.
Fig. 32 as referred to Fig. 34 is through the arrow e-f ; Fig. 33 is through the
arrow c-d. In the series of cross-sections, Fig. 9 is taken at a level a little
below that of Fig. 34, and passes through the basal part of the valve (vg).
(b) Marginal Pockets. The part of the peripheral portion of the gastro-
vascular system in each quadrant which is called the stomach pocket
extends downwards as far as the sensory niche. Here by the coming
together of the walls of the exumbrella arid subumbrella the space between
them is obliterated (Fig. 15) in the immediate perradius. From the
sensory niche downward to the margin each stomach pocket is thus
divided into two smaller pouches, the marginal pockets (mp). In each
side of the Cubomedusan cube there are, then, in Chary bdea two
marginal pockets ; or in all eight, a characteristic of the family Charyb-
deidse. The marginal pockets as the name implies extend downwards to
the bell margin, and are continued into the velarium as the velar canals.
Of these (Fig. 3) there are two from each marginal pocket, or sixteen in
all. The constancy in their number is one of the characteristics that
distinguish C. Xaymacana from the very closely related C. marsupialis
of the Mediterranean. (Compare Fig. 3 with the similar one by Glaus
for C. marsupialis, '78, Taf. I., Fig. 6.) The forked shape, while to be sure
the common form in C. marsupialis, is an almost invariable character-
istic in C. Xaymacana. It may be mentioned again that the presence of
these canals is one of the chief features that distinguish the velarium
of the Scyphomedusae from the velum of the Hydromedusse.

(c) Canals of the Sensory Clubs and Tentacles. The four interradial
definitive tentacles and the four perradial transformed tentacles, the
sensory clubs, are hollow, and their canals communicate directly with
the peripheral part of the gastro-vascular system. The canal of the
sensory club in each quadrant leads directly out from the stomach by a
somewhat funnel-shaped opening formed by the approximation of the
two walls of the stomach pocket. The relation of the canal of the sen-
sory club to the stomach pocket is seen at a glance in Fig. 37. It is given
by means of cross-sections in Figs. 12-14. Figure 12 shows the inner
3

18 JOHNS HOPKINS UNIVERSITY MOEPHOLOGICAL MONOGRAPHS.

walls of the stomach pocket approaching the outer at two points, leaving
between them a concavity freely open to the rest of the stomach pocket
above and at the sides. Fig. 13, a little lower down, shows the two walls
fused together at two points, making the interspaces a definite canal
communicating with the stomach pocket above only. This canal lies
directly over the sensory niche, and in the next figure (No. 14) the canal
is seen to have passed through the roof of the sensory niche and to have
entered the base of the stalk of the sensory club. In the enlarged end of
the club, the part which bears the sensory structure, the canal widens
into a terminal ampulla-like sac.

The endoderm lining the canal of the sensory club is specially differ-
entiated. In the stalk it is more columnar than the epithelium of the
stomach pockets, and is made up of cells containing a brightly staining
nucleus with very little trace of cytoplasm. The cell bodies appear as if
filled with a clear, non-staining fluid. Perhaps these cells give the stalk
elasticity to act in connection with the thin layer of longitudinal muscle-
fibres that are found just external to the supporting lamella. The epithe-
lium of the terminal enlargement of the canal is composed of very high
narrow cells, many of which show two nuclei of equal size and staining
quality lying side by side.

In continuation of the specialized epithelium of the perradial fur-
rows in the floor of the stomach the inner wall of the stomach pocket
shows a strip of similar densely crowded columnar cells leading from
the gastric ostium downwards to the canal of the sensory club. As in
the other case, the strip probably represents a specially ciliated tract, and
perhaps in it we see the reason why the canal of the sensory club is
almost always found to contain either spermatozoa which are shed by
the male reproductive organs directly into the stomach pocket, or else
floating cells of the kind to be described in the next section.

The canals of the interradial tentacles arise from the peripheral
gastro-vascular system much lower down than those of the sensory
clubs, since these tentacles have preserved their primary positions with
reference to the bell margin. Figure 16 represents a section taken at the
level of the base of the pedalia which gives the connection of the tentacle
canals with the gastro-vascular system. At the level below the sensory
niche the four broad stomach pockets have been divided, as we have seen,
into the right marginal pockets (mp). The figure shows that in the
interradial corners the longitudinal septa (ivl, in the preceding figures),
or lines of fusion between the two walls of the peripheral gastro-vascular

F. S. CONANT ON THE CUBOMEDUS^E. 19

space, which divide the primitively simple space into the four stomach
pockets, have come to an end, leaving a connecting canal (cc) in each
corner as all that remains of the primitive uninterrupted communica-
tion between all parts of the peripheral system. It is from these four
connecting canals that the tentacle canals take their origin. From this
point of origin each tentacle canal passes downwards, surrounded by the
gelatine of the pedalium, into the tentacle proper.

The connecting canals are of morphological importance in that they
are supposed, with much reason, to represent in the Cubomedusa3 the
circular canal of the Hydromedusse.

9. Reproductive Organs. The sexes are separate in Charybdea. In
both sexes the reproductive organs consist of four pairs of long leaf -like
bodies, each leaf attached along one edge to the wall of the subumbrella
in an interradius (see Fig. 1, r), and hanging free in the stomach pockets.
From this position in the stomach pockets it is evident that the repro-
ductive organs are endodermal. The lines of attachment of each pair is
just internal to the longitudinal vascular lamella that fuses the outer
and inner walls of the stomach pockets together in the interradius (wl),
and the reproductive organs are therefore structures belonging to the
subumbrella. It is interesting to note how careful examination of the
medusan organization takes away from the importance of the outer cup,
the exumbrella, and adds to that of the inner, the subumbrella. We
have seen that the phacelli and the sensory clubs, from whose position it
would be supposed that they belonged to the exumbrella, are organs of
the subumbrella, and that there is no muscle-tissue in the exumbrella ;
we find now that the reproductive organs belong to the subumbrella,
and it will be shown later that the tentacles, like the sensory clubs, are
structures of the subumbrella also. To the exumbrella are left only the
functions of support and covering.

The mature reproductive organs extend very nearly throughout the
entire vertical length of the bell, and are therefore found in the series of
cross-sections in all but the uppermost and lowermost (Figs. 7-15 r).
The organs consist of germ cells within, covered by an epithelium of
columnar cells that shows here and there nettle cells. The ova are found
with different amounts of yolk, according to age, surrounding a large
nucleus almost devoid of chromatin and an intensely staining nucleolus.
In young ova there appears very plainly in every case at least one small
deeply staining body inside the nucleus, which very much resembles the
nucleolus. These are probably so-called yolk nuclei, and while I have not

20 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

made a special study of the ovogenesis, I infer that the constant presence
of at least one, points to an origin of the ovum from a syncytium (of at
any rate two cells), similar to that which has been recently shown by
Doflein ('96) to occur in the formation of eggs in Tubularia. In the
nearly mature ovary each ovum is surrounded by a layer of gelatine,
which comes from the gelatinous sheet that enters the leaf -like ovary for
its support along its line of attachment just internally to the interradial
septum. It seems as if the ova, arising in the epithelium on the surface,
pushed their way into the gelatine inside and there completed their
development entirely surrounded by a slight investment of gelatine,
which grows thinner around each ovum as it increases in size. In males
the testes always show a similar division into compartments by gela-
tinous meshes, the compartments thus mapped out being filled with the
small brightly staining spermatocytes. Ova and spermatozoa when
mature are set free in the stomach pockets.

10. Floating and Wandering Cells. In the stomach pockets, the
canals of the sensory clubs, and even in the stomach itself, are found in
varying numbers freely floating cells having the appearance of young
ova. They vary in size, the smallest being of the size and having the
general aspect of the small ovocytes found in the ovary. The largest
(Fig. 70) have exactly the same structure as the young ovarian eggs
before they have begun to accumulate yolk. The granular deeply stain-
ing cytoplasm, the clear non-staining nucleus with its bright nucleolus
and the nucleolus-like yolk nucleus, all show beyond doubt that these
freely floating cells originate in the ovary.

In some of my preparations these cells are found not only floating
free, but wandering through the tissues. Fig. 70 shows two such wander-
ing cells fixed just as they were making their way either through the
digestive epithelium into the gelatine of the floor of the stomach, or from
the gelatine into the epithelium. The former seems the more probable,
though why they should want to get into the gelatine is not very easy to
conceive.

Perhaps there is some connection between this and the appearance
that the young ovarian eggs have of pushing their way from the epithe-
lium into the gelatine of the ovary. And of course it is not impossible
that the whole phenomenon is abnormal, due to rupture of the ovaries
which sets free young ova to exhibit their amoeboid tendencies under
new conditions. Against such an explanation, on the other hand, might
be urged the fact that what seem to be the small floating cells are found

F. S. CONANT ON THE CUBOMEDUS^. 21

occasionally in males as well as females, and that in the females a series
can be traced with a good degree of certainty between the small floating
cells like those found in the walls, and the larger ones which have all the
characteristics of young ova.

However that may be, this amoeboid action of cells having the struc-
ture of ova brings to mind the remarkable form of asexual reproduction
described by Metschnikoff for Cunina proboscidea, under the name of
" Sporogonie." Unfortunately Metschnikoffs original paper was not
accessible to me, so that I was unable to obtain more particulars on the
subject than those given in Korschelt and Heider's text-book (p. 33).
The reproductive organs of both males and females of Cunina proboscidea
are said to produce, besides the usual distinctively sexual elements,
neutral amoeboid germ cells, which wander into the endoderm of the
stomach and circular canal, and also penetrate into the gelatine of the
subumbrella. These amoeboid cells divide parthenogenetically. One of
the two cells of the first cleavage continues to divide and eventually
forms an embryo of Cunina ; the other remains amoeboid and serves for
movement, attachment and nourishment of the embryo.

Charybdea, however, has shown no sign of any such reproductive
process on the part of its floating and wandering cells. The only indica-
tion that I get as to their use points to a possible nutritive function.
The enlarged terminal portion of the canal of the sensory club almost
invariably contains a number of the small-sized floating cells. These
have a vacuolated, half disintegrated appearance, with the nucleus always
compact and brightly staining. Now, examination of the high columnar
cells that line the enlargement of the canal shows the presence in the
cells of bodies of exactly the same appearance as those in the lumen. In
one case a floating cell was found just at the end of an epithelial cell, to
all appearance half ingested. The identity of the bodies inside the cells
and those in the lumen is shown very clearly in some sections of material
fixed in formalin, which preserves nuclei, cell walls and general outlines
well enough, but does not retain the cytoplasm, and hence is useless for
most purposes of histology. In the endodermal cells of the terminal
enlargement thus preserved are found all the more distinctly the bright,
compact, degenerated nuclei of the ingested cells, while in the lumen are
seen other bright, compact nuclei with the poorly preserved remains of
cell substances around them. In addition to the evidence from the
appearance of the floating cells themselves and their ingestion by the
endodermal cells, a little collateral evidence may perhaps be brought in

22 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGEAPHS.

from the Tripedalia about to be described. From the ovaries in this
form are detached masses of cells (Fig. 71) which float free in the
stomach pockets among the developing embryos, and to judge from the
vacuolation that appears, are used up in their favor. These cell masses
are described more fully in the part on Tripedalia.

B: TRIPEDALIA CYSTOPHORA.
a. Habitat.

The species upon which the new family was founded was obtained in
great abundance in one locality in Kingston Harbor in the summer of
1896. The environment was even more unlike that in which Cubome-
dusae have been found heretofore than in the case of Charybdea Xayma-
cana. On the west side of the Harbor there is a part more or less cut off
from the main body of water, and so from the ocean, by a peninsula.
This sheltered bay is dotted with small mangrove islands which toward
the head of the bay become so numerous as virtually to convert it into a
mangrove swamp. The water is shallow and discolored with organic
matter, showing that the tide does not exercise much influence here, and
the bottom is for the most part a black mud, deep enough to make
wading very uncomfortable but not impossible near shore. The islands
rise but slightly above the level of the waters, and the thick vegetation
that covers them, for the most part mangroves, grows out into the water
on all sides, forming a fringe of overhanging boughs. It was here in the
shelter of the boughs, among the roots and half -submerged stems of the
mangroves, that the small Cubomedusa was found to thrive. It could
be obtained in great abundance almost any day, and of all sizes from the
largest adults with stomach pockets filled with eggs or embryos down to
small specimens only about two millimeters in diameter. In but one
other place was Tripedalia found, and that was a similar region of half
landlocked water skirted with mangroves, situated near Port Royal,
across the harbor from the locality just mentioned. It would be hard to
find places in which the conditions of life were more strikingly different
from those of the pure deep sea in which the Cubomedusa3 have been
generally found before. The slight brownish yellow pigment made the
small medusa? a little difficult to see in the discolored water, but like the
pellucid Charybdea in the clear water of the harbor, their active move-
ments gave away their presence. The swimming was very vigorous and
was effected by quick, strong pulsations (as many as 120 per minute were
counted), very different from the slow, rhythmic contractions of the

F. S. CONANT ON THE CUBOMEDUS.E. 23

Discomedusan Cassiopea which was found in the same region over by
Port Royal. Whether or not the animal made intentional efforts to
escape capture could not be decided satisfactorily, but certain it was that
they did escape often enough by swimming quickly below the surface of
the semi-opaque water.

Tripedalia endured captivity much more hardily than the Charybdea,
and would live in aquaria happily enough for a number of days — no
attempt was made to see how long. Specimens with their stomach
pockets filled with ripe spermatozoa, or with young at any stage from
egg to planula, were taken in plenty from the latter part of June to the
latter part of July. In each female the young were all at the same
stage. The embryos were thrown out in the aquaria as free-swimming
planulse, which settled down on the bottom and sides of the glass in a
day or two, and quickly developed into small hydras with mouth and
typically with four tentacles (and four tsenioles, W. K. B.), though three
and five were by no means uncommon. In this condition they lived for
three weeks without essential change, and they were still giving no
promise of further development when the laboratory broke up and the
jars had to be emptied.

b. External Anatomy.

The structure of the Cubomedusse seems to be that of a type well
established, and accordingly offers no very wide range of diversity among
the different genera. The Charybdea that has just been described is a
very typical form and will serve well as a standard with which to com-
pare our species of Tripedalia. The resemblances are so close that a
detailed account of the anatomy of the second form would involve much
needless repetition. It is hardly necessary to do more than merely point
out in what points Tripedalia resembles Carybdea and in what points it
differs.

The form of the bell is less pyramidal than in Charybdea. Some
measurements even gave the breadth greater than the height. The
external surface is divided, as typical for the Cubomedusae, into the four
perradial sides and the four convex interradial ridges, and the furrows
that separate these areas are with one small exception exactly the same
as those of Charybdea, as may be seen by comparing the series of
sections of Tripedalia (Figs. 21-30) with those of Charybdea (Figs. 6-15).
The exception is almost too slight to mention. The adradial furrow in
each octant which sets off the corner rib from the perradial surface in

24 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

the lower part of the bell is not directly continuous, as in Charybdea, with
the corresponding furrow in the upper part of the bell — that is, the afr1
of Figs. 24-27 is not continuous with the afr of Figs. 22 and 23, as is seen
by both being shown in Fig. 24. The upper furrow (afr) is continued
only a short distance, however, below the starting point of the lower (afr').

The pedalia conform entirely to the description given those of
Charybdea, except that there are three attached to the bell margin in
each interradius instead of one, and that the blade of each pedalium is
much narrower.

The sensory clubs also show exactly the same relation to the bell and
exactly the same structure.

In the bell cavity the proboscis has a longer and better denned stalk
than that of Charybdea, and has the further and more important differ-
ence of possessing special sensory organs, to the number of fifteen or
twenty. The suspensoria are much more developed than in Charybdea,
so that the interradial funnels lying between are more marked. In a
corresponding way the frenula are larger and stouter (Figs. 28, 29, frn).
The musculature shows no new features and differs only in being com-
paratively more strongly developed and having a more pronounced
striation. The nerve ring follows the same looped course from the
margin in each interradius up to the level of the sensory clubs in the
perradius.

c. Internal Anatomy.

The stomach offers no peculiarities, and the phacelli also agree with
those of Charybdea except in having a smaller number of filaments in
each tuft. The stomach pockets are not guarded by such well-developed
valves as those described for Charybdea, though the valvular nature of
the lips of the gastric ostia is indicated and the valvular functions
undoubtedly performed. The gastric ostia are smaller (cf. Figs. 7 and
22), and this makes highly developed valves less necessary. No trace of
anything corresponding to mesogonial pockets was noticed.

In the matter of the marginal pockets, however, we find that the
agreement with Charybdea is no longer continued. The regions that
correspond to the eight marginal pockets of Charybdea are formed, as in
that genus, by the coming together of the exumbrella and subumbrella
at the sensory niche (Figs. 25-28), but each of these regions is subdivided,
as it is not in Charybdea, into two marginal pockets, a larger (mp, Figs.
28-29) and a smaller (mp'). In this way sixteen marginal pockets are

F. S. CONANT ON THE CUBOMEDUS^. 25

formed as in the Chirodropidae. Furthermore, as happens in the latter
family but does not in the Chary bdeid;e, the marginal pockets extend
into the velarium. From each of the larger marginal pockets are given
off two velar canals, while each of the smaller gives rise to but one short
one (Fig. 18). Fig. 30 represents one of the last sections of a Tripedalia
cut transversely, in which nothing but the pedalia and the velarium
appear, and in it are shown the velar canals (i>c), which come from the
larger marginal pockets. The velarium appears in four segments because
it is drawn upwards in the four perradii by the frenula (see Fig. 20).
That the canals from the smaller pockets do not appear in the section is
due to their shortness and to the fact that they are pulled upwards above
the level of the sections by the frenula, together with that portion of the
velarium.

The smaller velar canals, a pair in each perradius, seem to have in
the males some function in connection with the storing of matured sper-
matozoa. In specimens with ripe testes they are very often found
crowded to distension with spermatozoa, while the other velar canals
may or may not contain them, and generally do not. The epithelium
lining them is, like that of the others, composed of columnar cells higher
on the wall turned toward the bell cavity than on that turned towards
the exterior, but otherwise not specially differentiated. I searched in
vain for any trace of opening by which the spermatozoa might gain the
exterior. Fig. 29 shows another point which may be mentioned in
passing, namely, that the canal of each of the three tentacles opens
into the peripheral gastro-vascular system independently. The central
tentacle of each group is the homologue of the single tentacle of Chary bdea,
and is formed in Tripedalia before the two lateral tentacles appear.
Its communication with the peripheral pocket system is higher up than
the openings of the lateral tentacles, so that in the section drawn the
latter are just beginning to be indicated (cf).

It remains only to speak of the i-eproductive organs of Tripedalia.
The sexes are separate in this form also, and ovaries and testes have the
same structure as is found in other Cubomedusas. The development of
floating masses of cells in the females, however, is a feature which, so far
as I know, has not been observed before. These masses, of which a
small one is represented in section by Fig. 71, are apparently developed
along with the eggs, and repeat the structure of the ovary to all intents
the same as if they were various-sized fragments of it broken loose.
They consist mostly of high, columnar epithelial cells surrounding a few
4

26 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

central cells and showing here and there a nettle cell just as the repro-
ductive organ does. The epithelial cells differ from those of the ovary in
containing one or more large vacuoles, and this vacuolation increases as
the embryos, among which the masses float, develop. The idea naturally
suggests itself, therefore, that they serve for nourishing and perhaps for
protecting the embryos while they are developing in the stomach pockets
of the mother individual.

PART III : DESCRIPTION OF SPECIAL PARTS OF THE

ANATOMY.

A: THE VASCULAR LAMELLAE.

In Medusas it is a common thing to find that in certain definite
places of the gastro-vascular system two endodermal surfaces that were
primarily separated by a space have come together and fused into a
single lamella or plate. Such a structure is called indifferently a
cathammal plate, an endodermal lamella, or a vascular lamella. In the
adult animal the vascular lamella? are by virtue of their very nature
formations "with a past." They are scaffolding left in the completed
structure, giving us clues as to the way in which that structure was
brought about ; and in the Cubomedusa?, whose development is as yet
unknown, they therefore afford an unusually interesting subject for
special consideration.

The vascular lamella? that are found in Charybdea and Tripedalia
may for convenience be described as forming two systems, the internal
and the marginal. The former comprises the endodermal fusions that
separate the stomach from the stomach pockets (except for the spaces of
communication left free, the gastric ostia) and those that separate the
stomach pockets from one another. The marginal system consists of
the lamella that connects endoderm of the gastro-vascular system with
ectoderm of the surface in a ring all around the bell margin, and with it
also the vascular lamella of the sensory niche, which has already been
referred to in the general description of Charybdea. The lamella? of the
internal system have been described by previous writers, and especially
by Claus in his paper on Charybdea, but they are still in need of compre-
hensive and clear treatment. The lamella? of the margin and of the
sensory niche have also been described by Claus, but not thoroughly or
with entire accuracy, nor did he recognize the vascular lamella? of the
sensory niche as originally a part of the lamella? of the margin. This last
was first determined by H. V. Wilson upon specimens of Chiropsalmus
quadrumanus obtained at Beaufort, North Carolina. Professor Wilson's
unpublished notes on Chiropsalmus were very kindly placed in my

28 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

hands, and so far as the vascular lamellae are concerned my own work is
only a confirmation and amplification of his, since Charybdea and
Tripedalia in this respect agree with Chiropsalmus.

The vascular lamellae of the internal system are the most prominent
and morphologically the most important. They comprise the four
vertical strips of fusion that separate the four stomach pockets in the
interradii (ivl in the figures of the series of cross-sections of Charybdea
and Tripedalia, Nos. 6-15 and 21-29), and four curved horizontal cross-
pieces at the top of these which separate the stomach from the stomach
pockets, and would make the separation complete did they not leave in
each perradius a free space between their ends, which makes possible the
gastric ostia.

The arrangement of this internal system of vascular lamellae is
simple. What they amount to is a certain definite number of linear
adhesions between the two walls of an originally undivided gastro-
vascular space, by which that space is divided up into a central stomach
and a peripheral portion, and the peripheral portion thus further divided
into the four stomach pockets. Perhaps the idea may be conveyed by
likening the whole medusa to a couple of bowls fitting closely one within
another and plastered together at the margins. The exumbrella then
would correspond to the outer bowl, the subumbrella to the smaller inner
bowl, and the original undivided gastro-vascular space to the space
between the two. If now the walls of the space be cemented together in
four horizontal curved lines just in the plane where the bottoms are
bending round to become the sides of the bowls, leaving four interspaces
between the ends of the lines, we should have the original space divided
into a central horizontal somewhat lens-shaped region between the
bottoms of the two bowls that would correspond to the central stomach,
and a peripheral vertical portion between the sides of the bowls that
would correspond to the peripheral gastro-vascular system ; central and
peripheral portions would communicate by the four interspaces between
the lines of fusion, which would correspond to the four gastric ostia. If,
further, the vertical peripheral portion be subdivided by four more lines
of fusion running vertically at equal distances apart, each connecting
above with the middle point of the corresponding horizontal line of
fusion, we should have the simple peripheral portion divided into four
parts, corresponding to the stomach pockets, by four vertical lines of
fusion, corresponding to the four interradial vascular lamellae, the ivl of
the figures.

F. S. CONANT ON THE CUBOMEDUS^E. 29

These mutual relations of stomach, stomach pockets and lamellae
will perhaps be made clearer if a comparison is drawn between them
and the similar structures of a Hydromedusa. Liriope, one of the
Trachomedusae, is a good form to take for such a comparison, since by
reason of its direct development from the egg it is free from the compli-
cations of hydroid medusa?. The young medusa has at first a simple,
undivided gastro-vascular cavity which later is divided up into the
central stomach and the typical radial to circular canals of the Hydrome-
dusae by means of fusions between the two endodermal surfaces. Dia-
grams a, b and c of Fig. 35 represent very schematically this process of
division into stomach and canals. In a we have a projection upon a
plane surface of the primary, undivided gastro-vascular cavity, as seen
from above ; b shows the first four points of fusion in the interradii ;
c represents those four points expanded by growth in all directions into
broad cathammal plates in such a way as to leave the stomach in the
centre, the i^adial canals in the perradii, and the circular canal in the
periphery as all that remains open of the primary simple cavity. These
broad plates of vascular lamella, separating the narrow radial canals,
persist in the adult Liriope to tell the tale of the formation of the defini-
tive gastro-vascular system. It seems to me that we are justified by
analogy in drawing a similar conclusion for the Cubomedusas. In d of
Fig. 35 is represented a projection of a Cubomedusa, in which the
homology of the stomach pockets with the radial canals of the Hydro-
medusa, and of the narrow strips of fusion with the broad cathammal
plates, is shown at a glance. To make the comparison more perfect we
have only to remember that in the Cubomedusa? there exists below each
interradial vascular lamella a connecting canal (Figs. 16, 29 and 35 d, cc)
uniting the two separate adjacent pockets. This, as has been pointed
out by other writers, is the representative of the circular canal of the
Hydromedusa?. Practically the only difference between the structure of
the gastro-vascular system of the Cubomedusa? and that of a form such
as Liriope, is that in the latter the fused areas have broadened out at the
expense of the radial canals, while in the Cubomedusae on the contrary
they have become long and narrow.

One is strongly tempted by the foregoing comparison to speculate a
little as to whether the reproductive organs of the Cubomedusas, which
lie in the stomach pockets and are generally supposed to be endodermal,
may not bear some closer relation to those of the Trachomedusa?, which
lie "in the course of" the radial canals (Lang's Text-book) and by common

30 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

consent are ectodermal. And while we are being led by facts such as
those just mentioned above to wonder just a little whether after all the
position of the Cubomedusee among the Acraspeda is so firmly assured-
doubting some, yet in the frame of mind of one who " fears a doubt as
wrong"— the velarium suggests itself as another point in question.
Haeckel does not hesitate to state emphatically that the velarium of the
Cubomedusae and the velum of the Craspedote medusae are only anal-
ogous, but the reasons that he gives (sie sind unabhangig von einander
entstanden, und ihre Structur ist zwar ahnlich, aber keineswegs iden-
tisch ; namentlich das Verhalten zum Nervenring ist wesentlich ver-
schieden : System, p. 426) somehow do not produce so much impression
upon one as the very velum-like appearance of the velarium itself. The
origin from the fusion of marginal lobes is not as yet a matter of obser-
vation, and the relation to the nerve ring is not essentially different from
that of the velum to the lower (i. e. inner) nerve ring in the Craspedotas.
The four frenula and the diverticula from the gastro-vascular system
seem to be the chief differences in structure after all, and these Haeckel
evidently did not think worth mentioning. This speculation, as to the
possible relation of the Cubomedusa3 to such forms of the veiled medusae
as Liriope, though it may be very tempting, is scarcely fruitful enough
to repay much effort on the part of either reader or writer. The whole
subject must remain uncertain until the facts of the development of the
Cubomedusae are known.

If the structure of the vascular lamella? of the internal system has
been made clear, the appearances of the vertical and horizontal compo-
nents in the figures will be understood without much further explana-
tion. The four vertical strips in the interradii (ivl) have been already
referred to in the figures of the cross-sections of both Charybdea and
Tripedalia. In the longitudinal sections of the two jelly-fish through the
interradii, the vertical lamellae are cut throughout their entire length
from stomach to connecting canals (Figs. 5-20, ivl}. The horizontal
cross-pieces at the tops of the vertical lamellae also appear in several of
the figures. Fig. 36 represents the appearance that would be given by a
longitudinal section taken through any portion of the upper part of the
bell except in the interradii, or in the perradii, through the gastric ostia.
The horizontal vascular lamella (hvl) is shown connecting the endoderm
of the stomach (ens) with that of the stomach pocket (enp). In a longi-
tudinal section directly through an interradius (Fig. 5 or 20) the hori-
zontal lamella is cut just at the point where it joins the vertical, so that

F. S. CONANT ON THE CUBOMEDUS^E. 31

the two are not differentiated. In a section through the region of a per-
radius (Fig. 4 or 19) the horizontal lamella is of course not cut, since the
section passes through the gastric ostium, whose existence is conditional
upon fusion not having taken place between the endodermal surfaces.

The first figure in each of the series of cross-sections (Figs. 6 and 21)
also shows the horizontal vascular lamella, cut across slantingly twice in
each quadrant as it passes between the gelatine of the ex- and of the sub-
umbrella to connect the epithelium of the stomach with that of the
stomach pocket. The fact that more of the lamella does not appear in
such a cross-section only shows that its course is not perfectly horizontal.

The region in which the same lamella lies is indicated in the surface
view of the top of the bell of Charybdea (Fig. 2) by the bent line hvl in
each quadrant. The figure manifests the appropriateness of Claus's
name for the horizontal lamella — " bogenformige Verwachsungs-
Streifen." Haeckel calls the same structures " Pylorus-Klappen," and in
his account of Charybdea Murrayana in the Challenger Report, speaking
of the three divisions of the stomach (buccal, central and basal) which he
traces upwards from the stalked forms of Scyphomedusae, he says: "The
central stomach in this Charybdea, as in most Charybdea, is joined to
the basal stomach, as the pyloric stricture between the two is not devel-
oped and only faintly indicated by the slightly projecting pyloric valves."
Again, in speaking of the valves of the gastric ostia, he says : " These
four perradial 'pouch valves' alternate with the interradial pyloric
valves." It is difficult to understand, however, how the " bogenformige
Verwachsungs-Streif en " of Glaus, which are undoubtedly the same struc-
tures as those which I have called the horizontal lamellae, and are only
strips of endodermal fusion, can be " projecting pyloric valves," or indeed
can properly be spoken of as valves at all. Possibly Haeckel was not
quite able to understand Claus's description, and in his desire to find
something in the stomach of Charybdea which would serve to set off a
central from a basal part, such as is found in the Lucernaridse, hit upon
Claus's " Verwachsungs-Streifen." I have elsewhere given it as my
opinion that in such of the Cubomedusse as I have studied there is no
structure in evidence that would properly serve to mark a limit between
a basal and a central portion of the stomach.

We have next to describe the marginal system. The vascular
lamellae mentioned above in every case connected endoderm of one
cavity with endoderm of another ; those of the margin have the note-
worthy difference that they run from endoderms of some part of the

32 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

gastro-vascular system to ectoderm of the surface. The outermost cells of
the endodermal lamellae make direct connection with the ectodermal
cells, without the usual intervention of a layer of gelatine.

The marginal lamella of Chary bdea lies, as the name implies, just on
the bell margin where the edge is curving round into the velarium. All
around the whole circumference of the bell it is found (in Charybdea) at
this same horizontal bend, except in the eight principal radii, where the
tentacles and the sensory clubs have brought about modifications. In
any place except these a vertical section through the margin will show
the marginal lamella connecting the endoderm of the marginal pocket
with the ectoderms of the surf ace,, as represented by vim in Fig. 38, which
is a vertical section through the sensory niche a little to one side of the
perradial axis.

In the interradii the marginal lamella undergoes modifications due
to the fact that the bases of the pedalia are situated a little upwards
from the exact margin, and that the lamella follows the outline of the
bases. Fig. 1 shows one of the interradial corners of the bell margin
looked at directly from the surface, so that the curved outline of the
junction of the base of the pedalium with the exumbrella is seen. The
trace made by the lamella where it meets the surface ectoderm follows
this outline. The lamella is also shown in the vertical section through
the interradius (Fig. 5 or 20, vim), where it is seen running from the
connecting canals (cc), which joins the two adjacent marginal pockets,
upwards and outwards to meet the surface ectoderm. Its course from
canal to surface is not in a direct line, but curved with the concavity
upwards. Hence, in cross-sections at certain levels through the inter-
radial corner it is met more than once and gives rise to appearances that
seem at first sight too complicated for it to be just the same structure as
the simple marginal lamella described above. That it is the same, and
that the complication is only due to the insertion of the pedalia above
the margin, can be determined by following through a series of cross-
sections, the essential ones of which, as I hope, are given in Figs. 40-43.
The levels of these are shown on Fig. 5 by the letters w, x, y and z,
respectively. Fig. 40 shows the lamella cut but once, just below its highest
part. The section is above the level of the connecting canal and hence
still shows the vertical interradial lamella ivl. Fig. 41, at the next lower
level (x), shows the same portion of the lamella intersected a little nearer
the interior, while the junction with the endoderm of the connecting
canal is shown still further inside. Fig. 42 is at level y, just through the

F. S. CONANT ON THE CUBOMEDUS.K 33

bend of the loop, so that in part of its course the lamella is cut almost
horizontally, i. e. in its own plane. Fig. 43 finally shows the lamella as
it appears below the level of the connecting canal, cut twice, each portion
joining endoderm of marginal pocket with ectoderm of surface. It tl\us
bears exactly the same relations that it had when we first met it in Fig.
38 (vim), except that here in Fig. 43 one finds that a cross-section cuts it at
right angles instead of a vertical as in Fig. 38, as a result of its being
pushed upwards from its former position on the margin by the insertion
of the pedalium above the margin.

The vascular lamella of the sensory niche has already been alluded to
as part of the marginal system, and brief reference has been made to it
in the section on the sensory clubs. Like the rest of the marginal
lamella, it connects endoderm with ectoderm. The line that its fusion
with the ectoderm traces on the surface frames in a shield-shaped area
at the bottom of the sensory niche, which is seen in the drawing of the
outlines of the niche, Fig. 44 (vis). This lamella was observed by Claus,
and was figured by him both in surface view and in cross-section
through the niche. Apparently, however, he omitted vertical sections
through the niche, so that he supposed that the outline traced by the
lamella was not continuous above, i. e. over the stalk of the sensory club
(78, Fig. 41 ; text, p. 28). That the outline is closed above, though
masked in surface view by the roof of the sensory niche, is seen at once
in vertical sections, such as Figs. 37 and 38, one of which is directly
through the perradius, the other a little to one side. Both show the vas-
cular lamella of the sensory niche (vis) intersected twice, above and below
the sensory club, and completely cutting off the exumbrella from any
share in the bottom (or inner wall) of the sensory niche. Fig. 39, which
is a cross-section through the upper part of the niche, and is essentially
like the similar figure of Claus, shows in like manner that the bottom of
the sensory niche belongs to the subumbrella. H. V. Wilson was the
first to point out, in his unpublished notes, that the lamella of the niche
is complete all round.

In the adult structure of Charybdea and Tripedalia the lamella of
the niche is connected with that of the margin by a vertical strip of
endodermal fusion that does not come to the surface like the rest of the
marginal system, but remains just internal to the gelatine of the exum-
brella, connecting the two adjacent marginal pockets. In the cross-
sections of Charybdea it is seen in Fig. 16 (vie) ; in those of Tripedalia it
is seen in Figs. 28 and 29. In vertical section it is found in Figs. 4, 19
5

34 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

and 37. In Fig. 44, which represents the bell margin and velarium of
Tripedalia arranged as if the velarium were vertical and pendant from
the margin (instead of suspended by the frenulum so as to be at right
angles to the vertical plane), the .connecting lamella is shown as a dotted
line (vie) — dotted because it does not come to the surface — joining the
lamella of the niche with that of the margin (vim).

The same figure (No. 44) shows a characteristic difference between
the marginal lamella of Tripedalia and that of Charybdea. While in
Charybdea, as Glaus points out, the marginal lamella keeps at one level,
just a little above the bell margin, all the way round (except where dis-
turbed by the special modifications of the tentacles and the sensory
clubs), and never descends into the velarium, in Tripedalia on the other
hand it describes a sinuous course, following the outlines of the mar-
ginal pockets, as is indicated in the figure by the light parallel line vim.
The course as it would be seen in a surface view is obscured just at each
side of the interradius by the overhanging of the bases of the two lateral
pedalia. This is why the lamella is not indicated at these points in the
diagram. The course is seen to lie almost wholly on the velarium, that
is, in the figure below the line which represents the bell margin proper,
the line at which the angle comes when the velarium is in its normal
position, horizontal to the vertical side of the bell.

In this sinuous course of the marginal lamella we have another point
of resemblance between Tripedalia and the Chirodropidae. H. V. Wilson
worked it out in his sections of Chiropsalmus, and the reconstruction
which I have given in the figure under discussion is in all essentials
similar to his for Chiropsalmus. The differences lie only in the fact that
Chiropsalmus has more velar canals, and that the chief marginal pocket
in each quadrant is not forked peripherally, as is that of Tripedalia (nip),
but presents its distal margin parallel to the edge of the velarium. The
two smaller marginal pockets in the per radii (ni})') are on identically the
same plan in both.

Tripedalia, having three tentacles joining the umbrella in each inter-
radius, shows a disturbance of the course of the marginal lamella in
these regions by just so much the more complicated than in Charybdea.
The plan, however, is exactly the same. The lamella is pushed upwards
from the margin by each of the bases of the three pedalia just as is done
by the base of the single pedalium of Charybdea. Fig. 29 shows the
lamella in the same relation to the canal of the central tentacle (ct) that
it has in the similar sections of Charybdea (Figs. 16 and 43) ; and in

F. S. CONANT ON THE CUBOMEDUS^E. M

addition the first appearances (as the series is traced downwards) of the
arches of the lamella over the two lateral tentacles (rf), which are
inserted a little lower down than the middle one of the group. As con-
cerns these lateral tentacles, the relations of the vascular lamella at this
level are the same as that in the level of Fig. 40 for Charybdea.

It has been stated more than once already that the vascular lamella
of the sensory niche is a part of the lamella that runs round the margin,
and so far the only evidence given has been the strip of endodermal
fusion running from the marginal lamella to that of the niche. This
strip, however, as has been described, does not come to the surface and
consequently seems at first sight to be a different structure from the
lamella of the margin. That it is not, however, I found very prettily
shown in a series of sections of one of my youngest Tripedalia. In this
the lamella of the niche as it was traced in successive sections downwards,
was found not to form a closed ring at the bottom of the niche, but each
side was continued directly and separately downwards to the margin,
whei'e it passed into the corresponding part of the marginal lamella. A
reconstruction of the condition, similar to that of Fig. 44, is given in Fig.
45, and I think explains itself at a glance. Evidently the vascular
lamella? that connect the lamella of the sensory niche with that of the
margin at first come to the surface, like the rest of the marginal system,
but as the animal grows older come to lie within the gelatine. In this
way the condition found in cross-sections just through the margin of my
very small Tripedalia, and represented in Fig. 46, becomes that of the
adult seen in the corresponding portion of Fig. 29. It is as complete a
demonstration as could be required that the lamella of the sensory niche
is at first only a loop of the marginal lamella, a conclusion that had been
already reached by H. V. Wilson on theoretical considerations, based
upon the facts of the adult structure as he found them in Chiropsalmus.

As Wilson pointed out in his notes, these facts have a close bearing
upon the question of the origin of the velarium. Sixteen marginal
pockets are found in both Chiropsalmus and Tripedalia, and all of them
extend into the velarium. It is not unnatural to suppose that these
belong to sixteen marginal lobes, and that these lobes have fused together
to form the velarium. In the Chirodropus figured by Haeckel (Taf.
XXVI) in his " System " gelatinous lobe-like thickenings are shown in
the velarium, corresponding to the sixteen marginal pockets. In Tripe-
dalia no special gelatinous thickenings are found, but the arrangement
of the marginal pockets is the same as that of the Chirodropida?, and

36 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

perhaps I ought, when treating of the systematic relations of Tripedalia
(p. 5, Fam. Ill), to have recognized the analogy to the extent of saying
that marginal lobes may not be completely absent from the velarium of
Tripedalia. At any rate the gelatinous lobes in the case of Chirodropus on
the one hand, and on the other hand the sinuous outline of the margin
still mapped out by the lamella in Chirodropus, Chiropsalmus and Tripe-
dalia, are certainly very suggestive of an ancestral Cubomedusa in which
there was no velarium, but sixteen free marginal lobes instead. Two
more indications favor slightly the same view. In both Charybdea and
Tripedalia a small notch is seen in the edge of the velarium in the perra-
dius (Fig. 44). Its constancy suggests that it may not be a chance or
meaningless feature. The second point is the small size of the two
marginal pockets adjoining the perradius. These are in the position of
the ephyra lobes of the Discomedusas, which always lie on either side of
each sensory club, and which do not keep pace with the other marginal
lobes in development.' In the Rhizostome jelly-fish especially they are
found much smaller than the other lobes, as will be seen by a glance at
such figures as Haeckel's for Lychnorhiza (System, Taf. XXXIV, Fig. 2),
or for Archirhiza (Taf. XXXVI, Fig. 5), or Hesse's figure of the margin
of Rhizostoma Cuvieri ('95, Taf. XXII, Fig. 22). The resemblance between
such margins and that of Tripedalia (Fig. 44), with its simple, unbranched
velar canals, is very suggestive. On the other hand it must be remem-
bered that in considering the vascular lamellae of the internal system we
found the indication pointing rather more to Hydromedusan affinities
than to any other. Charybdea throws no light on the question, since it
has no marginal lobes on the velarium and the marginal pockets end
strictly at the margin, so that the only diverticula of the gastro-vascular
system in the velarium are the velar canals.

Before leaving the subject of marginal lobes and pockets I must
answer a possible objection that may occur to some careful reader. It
may seem that I am wrong in holding that there are two marginal
pockets in each octant instead of three, that just as there is one velar
canal from each of the smaller perradial pockets (mp', Fig. 44), so each
prong of the forked larger pocket (imp), since it is continued into a velar
canal, ought to be called a marginal pocket likewise, the whole number
of marginal pockets then being twenty-four instead of sixteen. Such a
revision of the terminology would not be without some reason in its favor,
and perhaps a study of more forms would show it to be correct. But for
the present, at any rate, it seemed to me best to abide by the analogy of

F. S. CONANT ON THE CUBOMEDUS./E. 37

Chiropsalmus, in which the peripheral edge of the larger marginal pocket
in each octant is not bow-shaped, but runs parallel to the edge of the
velarium. A revision of the terminology of the marginal pockets such
as implied in the suggestion above would also give rise to complications
when applied to Charybdea, since the latter has no marginal pockets in
the velarium.

As to the functions of the vascular lamellae, there is too little known
to say much. It is rather improbable that structures retained so defi-
nitely should be mere scaffolding left over from a previous stage of
usefulness. Claus has found in Chrysaora that the lamella? form a kind
of capillary network in communication with the gastro-vascular system,
and he with others supports the view that they perform an accessory
function in the nutrition of the tissues they penetrate. Upon this point
I have no observations of my own to add.

The marginal vascular lamella is regarded by Claus as perhaps the
vestige of a circular canal around the bell margin. On this subject, too,
I have nothing to add. A lamella of endoderm that connects directly
with the ectoderm of the surface along its whole course is a structure
whose meaning I am wholly unable to understand or even to guess at.
A similar lamella is described by Hesse ('95, p. 430) as occurring in the
ephyra lobes of his Ehizostoma, and he mentions Eimer as the first to
discover this structure, probably meaning the first to discover it in the
Discomedusa3. Whether the lamella is found all around the margin is
not stated. Hesse refers it to the ephyra, and remarks that the investi-
gation of it in the ephyra would undoubtedly give interesting results.

I will close this part upon the vascular lamellas with a very pertinent
suggestion made by Professor Brooks to the effect that the usual way of
speaking of the sensory clubs as having moved up from the margin is
looking at the matter in the wrong way. The level of the sensory clubs
undoubtedly represents the original margin, which elsewhere has grown
down and away from its former level, leaving the sensory clubs like
floatage stranded at high-tide mark. Only in this way can the lamella
of the sensory niche have any meaning.

B : THE NERVOUS SYSTEM.

The nervous system of the Cubomedusa? is the most highly devel-
oped that is found in any of the jelly-fishes. If the position of the group
among the Acraspeda is established, it alone is ample to prove that the
Hertwigs had not sufficient evidence when they stated in their mono-

38 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

graph on the nervous system of the Medusas (78) that the Acraspeda
show a much lower nervous organization than the Craspedota.

The system naturally groups itself under three heads, the nerve
ring, the sensory clubs, and the motor plexus of fibres and ganglia that
underlies the epithelium of the subumbrella. The general relations of
the nerve ring and of the sensory clubs have been given before in the
description of Charybdea Xaymacana, so that we may pass at once to the
consideration of the finer details of the nervous tissues.

In the structure of the nerve ring I have found myself unable to
corne to the same results as those given by Glaus, who so far as I know
is the only one that has studied the nerve with special reference to its
histology. Our difference amounts to this, that he finds two distinct
types of cells in the epithelium of the nerve, sensory and supporting,
which would make it a receiving as well as transmitting organ, while I
have not been able to demonstrate satisfactorily the sensory cells, and,
therefore, so far as my own observation is concerned, I am disposed to
attribute to the nerve simply the function of conducting impulses. I do
not know just how much weight to assign to my inability to find
evidence in my sections of the sensory type of cells. Eimer (mentioned by
Hesse, '95, p. 420), the Hertwigs (78) and Glaus (78) have independently dis-
covered the two types in one medusa or another, and the Hertwigs, at least,
have demonstrated them by macerated preparations. So far as Charybdea
is concerned, however, Glaus had only preserved material and had to rely
upon sections, as have I, since the material which I had preserved with
especial reference to maceration did not turn out well. The results that
we get from sections vary enough for me to believe that Glaus inter-
preted his sections very much by analogy with other forms — as indeed,
is suggested by his own words (78, p. 22) : " Da es mir nicht gegliickt ist
die durch die langere Conservirung in Weingeist fest vereinigten
Elemente zu isoliren, habe ich das muthmassliche Verhaltniss beider
Elemente nach Analogie der mir fur die Acalephen bekanut gewordenen
Verhaltnisse, welche 0. und R. Hertwig so schon auch am Nervenring
der Carmarina zur Darstellung gebracht haben, zu erganzeii versucht."
There can be no doubt of our having the same structures to deal with,
for C. Xaymacana is so much like C. marsupialis as to be perhaps more
worthy of being called a variety of the latter than a distinct species.

The structure of the nerve as I conceive it is given in Figs. 47 and
48. The former represents a cross-section, and shows, as others have
pointed out, that the layer of circular muscle fibres (cm) is interrupted by

F. S. CONANT ON THE CUBOMEDUS.E. 39

the nerve. It is evident that the tissues which elsewhere on the subum-
brella were differentiated into muscle epithelium and muscle fibre have
here become nerve epithelium and nerve fibre, a point that has not been
remarked upon before, so far as I remember, and that may be of interest
in connection with the iieuro-muscular theory. The epithelium of the
nerve (sen) is seen to be made up of cells whose inner ends narrow down
into a kind of stalk or process that runs to the gelatine of the supporting
lamella (gs) and there joins a little cone of the gelatine that juts out to
meet it. The cells are smaller in general than those that overlie the
muscle layer, especially on the two lateral margins of the nerve, where
they are more crowded together and overarch the nerve-fibres. The
fibres are seen in cross-section between the processes of the cells. They
apparently must lie imbedded in some clear, watery fluid that does not
show in the preserved material. The processes of the epithelial cells
give the fibres the appearance of lying in alveoli, or being divided into
strands, and one of these strands (ax) is always discernible among the
others by reason of its more numerous or finer or more compactly
massed fibres. This is the " axis " of Glaus. Here and there in its course
appear ganglion cells having their long axis in the longitudinal direction
of the nerve. Elsewhere, in the nerve as well, and usually nearer to the
surface, are found other ganglion cells, mostly bipolar, some multipolar,
which are readily distinguishable from those of the axis by the fact that
their long axis lies across the nerve. One of these cells is shown in the
figure (gc). Here and there in the epithelium alongside the nerve are
found mucous cells (me), distinguished by their clear contents and by the
small exhausted-appearing nucleus at the base with a few threads of
protoplasm.

In Fig. 48 I have tried to represent the structure of the nerve by
means of a series of five different views such as would be given by
focusing at five successive levels. In the first (1) we have the epithe-
lium of the nerve (sen in Fig. 47) in surface view, the cells appearing
polygonal in outline, with here and there a mucous cell. In (2) we find a
very slight layer of ganglion cells and fibres having a transverse direction
(gc and fp in Fig. 47). These are continuous with the plexus of fibres and
ganglion cells which lie above the muscle layer all over the subumbrella,
and which represent the motor part of the nervous system. This connec-
tion with the nerve shows how co-ordination is effected. At the same
level are found fibres of the axis also having a longitudinal direction.
In (3) is seen the main body of fibres, divided in the osmic preparation

40 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

from which the drawing was made into irregular wavy strands which
are in all probability largely the result of preservation, but are in part
also due to the separation by processes of the epithelial cells, as was seen
in Fig. 47. The axis is seen with one of its longitudinally directed bipolar
ganglion cells ; and at the sides the fibres of the circular muscle of the
subumbrella. These show a slanting direction to the nerve, due to the
fact that the nerve, as mentioned before, has a sinuous course from the
margin in interradius to the level of sensory club in perradius. At the
next focus (4) we come to the gelatine of the subumbrella (gs in Fig. 47),
and below this (5) to the larger polygonal outlines of the endodermal
cells of the stomach pocket (enp, Fig. 47), which like the ectoderm show
mucous cells at irregular intervals.

A comparison, now, with Claus's figures ('78, Taf. II, Figs. 19-21) will
show that, except for the rather unimportant matter of the mucous cells,
which he finds regularly and thickly disposed on each side of the nerve
('78, Fig. 21), our only .essential difference lies in the matter of sensory
cells in the epithelium. His figures show a multitude of spindle-shaped
sensory cells whose central ends are continued in processes that bend
around into the mass of fibres of the nerve. In his Fig. 20 a relatively
small number of nuclei, j ust one-third as many, are seen attached nearer
to the surface, which represent the supporting cells. The plan of struc-
ture (as shown in his Fig. 20) is an alternation of (1) supporting cells
offering a broad peripheral end to the surface and having the central end
continued as a supporting fibre to the gelatinous lamella, and (2) spindle-
shaped sensory cells with nuclei at a lower level, which send their
peripheral process up between the supporting cells to the surface, while
the central process becomes continuous with the nerve fibres, often
branching into two processes. In my sections I have not been able to
see either a regular alternation of nuclei at different levels, or central
processes which unmistakably bend round into the nerve fibres. In
every case in which I could trace the central process of a cell clearly it
ran to the supporting lamella, and this whether the nucleus of the cell
lay near the surface of the nerve or deeper down, as in the somewhat
spindle-shaped cell seen on the left of the centre of the nerve in Fig. 47.
Of course in many cases the central process could not be traced in a
section, and this leaves room for the supposition that such were always
the sensory cells. From my inability to demonstrate sensory cells in the
nerves of Charybdea, I by no means wish to deny their existence ; for
that remains to be proved, or disproved, by macerations. At any rate,

F. S. CONANT ON THE CUBOMEDUS^E. 41

they cannot be so numerous as has been supposed. The position of the
nuclei shows that.

The epithelium of the nerve is said by Glaus to be ciliated. It has
been suggested by Schewiakoff that probably in such cases the sensory
cells bear one long cilium, while the supporting cells have many smaller
cilia. Unfortunately, I made no observations upon the ciliation of the
nervous structures of the living animal, and the traces of cilia that are
shown in preparations of preserved material are a poor basis to speculate
much on. Glaus considers the sensory cells of the epithelium of the
nerve a special seat of tactile sensation.

The way in which the nerve reaches the sensory clubs is interesting.
Under the topic of the vascular lamellae it was explained that the sensory
clubs and the bottom of the sensory niche from which they spring are
parts of the subumbrella. Fig. 37 reminds at a glance better than any
other one drawing how the bottom or inner wall of the niche is com-
pletely cut off from the exumbrella by vascular lamellaa above and below
the stalk of the club. From this figure, now, it will readily be under-
stood that the nerve in order to pass to the base of the stalk has simply
to traverse the gelatine of the subumbrella. This fact, which seems sur-
prising enough at first sight in view of the position of the clubs on the
external surface of the umbrella, was correctly pointed out and explained
by Glaus, but one or two figures will serve perhaps to give a clearer
idea of it.

Fig. 49 is a diagram of the nervous structures in the region of the
sensory niche, as they would be seen on the surface of the subumbrella
turned toward the bell cavity. The outline of the sensory niche as it is
seen through the tissue of the animal is represented by the line osn.
The sensory club (scZ), and its stalk with a conical basal portion are given
by the lightly dotted outline and are also imagined as seen through the
animal. The nerve (n), being on the surface of the subumbrella, is shown
as a heavy line describing an arch over the outline of the niche. In the
middle point of the arch is a slight thickening of the nervous tissue (rg}
which shows in section a large increase in the number of ganglion cells,
and is the radial ganglion of Glaus. The same is seen, exaggerated in
size, in Fig. 12. From it there extends upward a slender strand of
nervous tissue (rn\ the radial nerve of Glaus. In Charybdea this can
be traced but a very short distance. In Tripedalia it is much more
distinct and traceable for a longer distance, and I might say in passing
that this and the sensory organs in the proboscis are the only differ-
6

42 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

ences I have noted between the nervous systems of Tripedalia and
Charybdea.

Nerve ring, radial ganglion and radial nerve all lie on the bell cavity
surface of the subumbrella. The way, now, in which the nerve ring
reaches the base of the stalk is simply by sending two roots through the
gelatine of the subumbrella to the conical base of the stalk. These roots
are seen in the diagram at rns. After passing through the gelatine the
roots come together on the inner side of the base — that is, the side turned
toward the bell cavity — and then pass downwards (nst) on the inner side
of the stalk of the club to the mass of nervous tissue at its end.

This passage of nervous tissue through the gelatine in order to reach
the sensory club is a little hard to grasp at the first, and I have tried to
render it more intelligible by a couple of drawings of sections. Fig. 50 is
a transverse section through the upper part of the region of the sensory
niche, not quite horizontal (i. e. parallel with the bell margin), but slant-
ing so as to lie on the plane of the reference arrow x-y in Fig. 49. The
plane passes just through the top of the niche, and in two areas has cut
through the roof with its epithelium of ectoderm (ece, ecs) so that the
space of the sensory niche (sn) appears. The vascular lamella of the
sensory niche (vis) is shown, as in Figs. 13 and 14, running on each side
from the endoderm that lines the canal of the sensory club (enc) to the
endoderm of the adjacent stomach pocket (enp). By it the gelatine
of the exumbrella is separated from that of the subumbrella, and
one sees that it is only through the latter that the nerve has to pass in
order to reach the base of the sensory club. It is also seen that one part
of the roof of the niche which is cut through lies outside of the ring of
lamella and is therefore lined with ectoderm of the exumbrella (ece)
while the other lies within the ring and is lined with ectoderm of the
subumbrella (ecs). Owing to the slanting direction of the cut only the
root on one side is cut through. The other is indicated, however, on the
right side of the drawing. In this method of passage of nerve fibres,
together with the accompanying ganglion cells, directly through the gela-
tine to the stalk of the sensory club my work is only confirmation and
explanation of Glaus.

Fig. 51 is a vertical section through the base of the stalk in the
plane of the reference arrow w-z in Fig. 49, and therefore passing
through one of the roots of the nerve of the stalk. Here again the
region is seen to, be cut off from the exumbrella by the vascular lamella of
the sensory niche (vis), and the nerve is seen passing through the gelatine

F. S. CONANT ON THE CUBOMEDUS^E. 43

of the subumbrella from the surface of the bell cavity (.sc) to the base of
the stalk hanging in the sensory niche (sn). One of the ganglion cells
(gc) that accompany the nerve is seen to have two nuclei, a not infre-
quent occurrence which has been pointed out by others.

The same figure shows that the axis (ax} of the nerve has penetrated
the gelatine with the other fibres. Here at the base of the stalk it takes
a horizontal course and becomes directly continuous with the similar
structure of the other root, as Wilson, I believe, first pointed out. This
part of the nervous tract which runs horizontally along the base of the
stalk between the two roots (Fig. 49, rns) has been considered by Glaus
the representative in Charybdea of the upper nerve ring of the Craspedota,
which therefore exists in Charybdea in four separate portions. Seeing,
however, that the region in which it is found belongs to the subumbrella,
the homology seems very doubtful. Moreover, the fact that the axis of
the nerve ring runs through this outer portion, instead of remaining on
the inner surface of the subumbrella and passing to the radial ganglion,
rather indicates that the outer portion is part of the original course of
the nerve ring, while the portion that remains on the inner surface is
perhaps a later formation.

A very interesting feature of the nervous system occurs in the same
region in the form of a tract of fibres underlying the endoderm, and
separated from the other fibres by the gelatine of the supporting lamella.
It is seen in vertical section in Fig. 52 (enf\ which is a section through
the base of the stalk in just about its median plane, and, therefore, to one
side of the arrow w-z in Fig. 49 and the corresponding drawing, Fig. 51.
In cross-section it is represented also in Fig. 50 (enf). It varies in size
and prominence very much in different specimens. Fig. 52 is a camera
drawing of it in the case that showed it most developed. Ganglion cells
are found in it, but comparatively infrequently. In some cases the tract
itself can hardly be found with certainty. Hesse has described in a
Rhizostome a much more highly developed tract in a corresponding-
position on the base of the marginal body. Fibres from the " outer
sensory pit" pass through the gelatine to the sub-endodermal tract,
which is described as surrounding the epithelium of the canal of the
marginal body like a collar and is most thickly developed on the under
surface of the canal, at the place that just corresponds with the point
where, and where only, I find the tract in Charybdea. Hesse thinks that
fibres then pass from this region to the nervous epithelium of the " inner
sensory pit" lying underneath the base of the marginal body, which

44 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

•

contains a rich supply of ganglion cells and is considered by him to be
the centre of the nervous system of the medusa. A close comparison
cannot be drawn with Chary bdea in this matter, however, since Charybdea
has nothing to correspond with the "outer" and "inner" sensory pits.
Moreover, the endodermal tract is not found encircling the canal of the
sensory club, nor could I trace fibres passing from it through the
supporting lamella into the fibres of the nerves.

Glaus has figured (78, Taf . V, Fig. 45, Fb) a small bundle of fibres in
the stock of the sensory club lying between the endoderm cells of the
canal and the supporting lamella. The same bundle is found in both
Charybdea and Tripedalia and can be traced in cross-sections up the
stalk to a point which must correspond with that at which the endoder-
mal tract is seen in Fig. 52. Downwards it can be traced only as far as
the entrance of the stalk into the knob of the club where it invariably
becomes lost to view. According to Hesse ('95, p. 427) Schafer found
under the endoderm cells of the whole stalk of the marginal body a
fibrous layer like that under the endoderm cells which he refers to
slender processes from the cells of the crystalline sac. Although Hesse,
as we have seen, finds the layer more limited in extent than Schafer
gives it, and does not trace it to the same source, the observation of
Schafer seems to me worthy of mention here, inasmuch as the trend of
the fibrous bundle under the endoderm cells of the stalk in Charybdea
and Tripedalia suggests quite strongly that the fibres come from the
crystalline sac, as Schafer thought to be the case in his medusa.

Besides the radial ganglion situated in the course of the nerve ring
at its four perradial points there are four other similar ganglia 011 the
subumbrella. These lie in the interradii, at the four lowermost points of
the nerve's course, and undoubtedly send off nerves into the pedalia at
whose bases they are situated. F. Miiller ('59), whose work was not
accessible to me, is quoted by Glaus as recording two ganglia opposite
the base of each pedalium which gave off a great number of nei'ves
partly into the velarium, partly into the tentacles. Glaus observed
nothing of the kind in Charybdea and states that even the interradial
ganglia do not exist.

That they do, however, is shown without doubt in sections of both
C. Xaymacana and Tripedalia, but nerves to the velarium or to the ten-
tacles I was unable to find.

On the two sides of each frenulum and of each suspensoi'ium are
found sub-epithelial ganglion cells in greater numbers than elsewhere on

F. S. CONANT ON THE CUBOMEDUS^E. 45

the subumbrella, and I am inclined to ascribe to them also the import-
ance of special ganglia controlling the musculature of the frenula and
suspensoria. Certainly such ganglia would not be out of place.

It has been mentioned that the greater prominence of the radial
nerve and the possession of special sensory organs in the proboscis were
the only points of difference I had noted between the nervous systems of
Charybdea and Tripedalia. These sensory organs remain to be described.
They are simple ciliated cysts containing a concretionary mass, and are
situated in the gelatine of the proboscis, irregularly disposed of at any
level, from the lips to the beginning of the stomach, and in any radius.
In one series of the adult animal fifteen were counted, of which seven
were situated about interradially, four perradially, two adradially and
two subradially. In another, twenty-one were counted, twelve in the
perradii and nine situated between the sub- and perradii. The one
shown in Fig. 24 is in the perradial position, often seen. In the sections of
the very young Tripedalia in which the vascular lamella had not reached
the adult condition the sensory organs of the proboscis were not found,
although the sensory clubs showed practically no difference from the
adult. Their structure is very simple — merely a round or oval sac lined
with ciliated cells which bear up and keep in constant motion an irreg-
ular coarsely granular concretion. Fig. 53 is a sketch made in Jamaica
from the living specimen. Sections were somewhat disappointing in
that they added but little. Fig. 55 was drawn to show that now and
then a mucous cell (me) is found among the other cells of the sensory
epithelium. An irregular-shaped mass (re) was always found inside the
cysts as the organic remains of the concretion. It gave no trace of
cellular structure and offered no evidence whether the concretion was
the product of one or few or of all the cells of the cyst. The latter would
be unique among the medusee. Even if the otocyst is the result of the
activity of only one or a few cells, it is, so far as I know, the only case
known for the jelly-fish of a free, unsuspended concretion.

As to whether the cysts are of ectodermal or endodermal origin
could not be determined, but there was some evidence in favor of the
latter. Fig. 56 is a drawing of one seen in optical section in a whole
mount of part of a proboscis, and shows a definite connection with the
endoderm of the proboscis. This was the only case when such connec-
tion was satisfactorily established, but in sections it was not uncommon
to find what seemed to be the remains of the broken stalk, as in Fig. 54
(rs ?}. No connection could be traced between the cysts and any other

46 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

part of the nervous system. As to function, the idea that they serve to
give perception of space relations suggests itself as readily as any other
hypothesis.

We come now to the consideration of the terminal knob of the clubs,
the sensory portion proper. A complete and detailed account of the
complex structure of these organs would fill many pages and involve
much useless repetition. Claus ('78) has described them with accuracy,
but not in great detail, and since then Schewiakoff ('89) has given a
careful general description and has supplemented Claus's work by
observations upon the finer structure made with the aid of more recent
technique. It seems in place for me, therefore, to give in the briefest
possible way a general idea of their structure, and to pass then at once to
the points in which my work has led me to different conclusions from
those of Claus and Schewiakoff. In brief, then, the knob of the sensory
club consists of a thick, complex mass of nerve fibres, more or less
imbedded in which lie the special sensory organs, surrounding the
ampulla-like terminal enlargement of the canal. The surface between
the special organs is covered with less specialized sensory epithelium.
The sensory organs are seven in number. Of these, four are simple
invaginations of the surface epithelium arranged in two pairs symmet-
rically to the median line in the proximal end of the knob (the end where
the stalk enters) and having pigment developed in the cells so invagi-
nated, while the space of the invagination is filled with a gelatinous refract-
ing secretion. These are considered simple eyes. Two more of the organs
are complex eyes situated on the median line of the inner surface of the
knob, the upper one smaller than the lower, but having almost exactly
the same structure. Each has a cellular lens over which extends a super-
ficial, corneal layer of cells ; below the lens a refractive "vitreous body";
and below this a retina with pigmented cells. The seventh organ is the
crystalline sac, which lies almost at the end of the knob opposite to the
stalk and contains a large concretion. In view of the fact that the
sensory clubs in toto have been abundantly figured by Claus and
Schewiakoff, it is my intention to give but one simple figure of the
general relations, and I justify that one in that it was made from the fresh
material. Fig. 57 is a camera sketch of the outlines given by a sensory
club seen in optical section from the side. The smaller upper and the
larger lower complex eyes which are situated on the mid-line, are seen in
profile, while the two small simple eyes give the outlines that they would
in a surface view of their side of the knob. Of course it is understood

F. S. CONANT ON THE CUBOMEDUS^E. 47

that two similar ones would appear on the other side, since the four
simple eyes are symmetrically paired on either side of the mid-line. The
sketch seems to show at least this much, that even in the living state the
lens of the larger eye projects out beyond the other contours of the
surface, so that the marked convexity ascribed to it in descriptions is not
to be attributed to the preservation.

It is in reference to the structure of the retina and vitreous body of
the complex eyes that I have found myself unable to come to the same
conclusions as Claus and Schewiakoff . Since the work of the latter goes
much further into the detail of the subject than does Claus's paper, it
will be sufficient for me to compare my results simply with those of
Schewiakoff.

The latter finds that the retina is composed of two kinds of cells,
corresponding to the supporting and sensory cells referred to in the
description of the nerve ring. These he figures ('89, Taf. II, Figs. 12 and
13) as alternating regularly. The two kinds of cells differ as follows :

(1) Shape. The supporting cells like those referred to before, are
cone-shaped, having a proximal fibrous process that runs into the under-
lying stratum of nerve fibres, and on the surface of the retina a broad
distal pigmented termination. The sensory cells are spindle-shaped, the
proximal processes becoming continuous with fibres of the underlying
nervous mass, while the distal process runs up to the surface of the
retina (the part toward the lens) in between the ends of the supporting
cell. The two kinds of cells are accordingly designated as pigment and
visual.

(2) Position of nucleus. This comes in as a corollary of the shape.
The nuclei of the visual cells lie in the enlarged central part of the
spindle-shape, and, therefore, at a lower level than the nuclei of the
alternating pigment cells.

(3) Processes in the vitreous body. The distal processes of the spindle-
shaped visual cells are continued through the vitreous body to the cells
of the lens as rod-like visual fibres which lie in canals in the (supposedly)
homogeneous vitreous body. The pigment cells on the other hand have
no fibres passing from them through the vitreous body, but in the latter
are situated cone-shaped masses of pigment whose bases rest upon the
broad ends of the pigment cells without, however, being a part of the
cell.

(4) Pigment. The distal ends of the pigment cells in the retina are
strongly pigmented, as the name implies. The processes of the visual

48 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

cells, which alternate with these, are pigmented likewise, but the pigment
is not so abundant and lies in the periphery of the cell body, leaving free
a highly refracting central axis.

If the relation of these cells to each other has been made sufficiently
clear, it will be understood that, in accordance with Schewiakoff s scheme
of the structure, sections that cut the retinal cells transversely give very
different appearances at different levels. A section through the very
tops of the retinal cells, that is, the last section of the retina before
striking the vitreous body, would show large polygonal areas of heavy
pigment (the ends of the pigment cells), in between which would lie the
much smaller, less pigmented, highly refracting ends of the visual cells
('89, Taf. II, Fig. 19). A section lower down in the retina, that is, more
toward the centre of the club, would strike the low-lying enlarged central
portion of the visual cells with their contained nuclei, and the smaller,
proximal ends of the pigment cells. It would, therefore, give the reverse
appearance from the preceding section, namely, that of large unpig-
mented (or but slightly pigmented) areas (the swollen bodies and nuclei
of the spindle-shaped cells), and in between them smaller pigmented
areas, the ends of the proximally tapering pigment cells ('89, Taf. II, Fig.
20). A section on the other side of the one first described, that is, one of
the first through the vitreous body, would show pigment areas of the
same size as the large ends of the pigment cells (the cone-shaped streaks
of pigment in the vitreous body which according to Schewiakoff are
associated with the pigment cell), and in between them the cross-sections
of the rod-like processes from the visual cells, lying in canals in the clear
homogeneous ground-substance of the vitreous body ('89, Taf. II, Fig. 18).

Let me give a resume of Schewiakoff's conception of the structure
of the retina.

a. There is an alternation of pigment and visual cells, the nuclei of
the spindle-shaped visual cells lying at a lower level than those of the
cone-shaped pigment cells.

b. From the visual cells extend rod-like processes into the vitreous
body, lying in canals in the latter.

c. In the vitreous body a cone-shaped streak of pigment overlies
each pigment cell of the retina, which is not a part of that cell.

d. Apart from these pigment streaks and the rod-like processes of
the visual cells the vitreous body is structureless, probably a secretion of
the pigment cells.

My own work, now, has led me to a different conception, so that my
conclusions on the same points would be as follows :

F. S. CONANT ON THE CUBOMEDUS^S. 49

a. There is not good evidence of an alternation of cone-shaped pig-
ment cells and spindle-shaped visual cells, with the nuclei of the latter at
a lower level than those of the former.

b. From some of the retinal cells otherwise not distinguished, there
extend rod-like processes into the vitreous body, such as described by
Schewiakoff.

c. The cone-shaped streaks of pigment in the vitreous body belong
to the underlying pigment cells, in fact are direct continuations of them,
and at their distal ends they are prolonged into fibrous processes lying
in canals of the vitreous body exactly like the visual fibres of Schewi-
akoff.

d. The vitreous body is not a homogeneous secretion, but is com-
posed of prisms of refracting substance, each with a denser central fibre.

Let us go over these four points in detail.

(a) As to the first, the question whether there is an alternation of
pigment and visual cells, I am not prepared as yet to make a positive
statement, since my not seeing both kinds as they are described has
little evidential value against the fact that Glaus and Schewiakoff both
claim to have seen them. Perhaps proof could be obtained one way or
the other by maceration of fresh or of specially prepared material,
which none of us had. My evidence for not confirming alternation rests
wholly upon sections. Fig. 58 represents a radial section through part
of the larger eye of Charybdea, made from an osmic preparation which
in this case showed two advantages over the material fixed in corrosive-
acetic (usually by all odds the best), namely, that the vitreous body (vb)
was not shrunken away from the retinal cells, as almost invariably
happens, and that the retinal cells were contracted apart from one
another in some places in such a way as to be almost equal to a macer-
ated preparation. Now, in the figure it is seen that there is an apparent
alternation of two kinds of cells, more regular than I usually find, but
the ones that are undoubtedly the pigment cells of Schewiakoff are the
ones that show the fibrous processes like his visual cells, and the pigment
streaks in the vitreous body are seen to be integral parts of the cells, not
cone-shaped masses lying in the vitreous body, merely associated with
the pigment cells. If these are the pigment cells of Schewiakoff, the
shorter cells in between must be his visual cells, yet they can by no
means be said to conform to a spindle-shaped type, nor are their nuclei
always at a lower level than (that is, internal to) those of the pigment
cells. If the long cells with the fibres are, on the other hand, considered
7

50

JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

the visual cells of Schewiakoff, then again we find nonconformity to a
spindle-shaped type, and nuclei not always at a lower level. The matter
of alternation of nuclei at different levels seems to me any way too slight
a distinction upon which to base a difference in function. It is a neces-
sary mechanical consequence of the crowding together of many cells on
one surface. And in many cases in perfectly radial sections through the
retina I find the nuclei fewer in number and arranged in very nearly a
single level. The retina of the smaller eye represented in Fig. 69 shows
this. In sections further along in the same series the nuclei are found
at different levels, due without doubt to the slanting cut.

t'O

[Dr. Conant did not complete Fig. 72, and the accompanying outline of Fig. 7 of
Schewiakoff' s memoir (Beitrage zur Kenntnis des Acalephenauges, Morph. Jahrb.,
Bd. XV, H. 1) has been substituted. — EDITOR.]

EXPLANATION OF LETTERS IN TEXT FIGURE. — C — concretion cavity ; CO — cornea ;
CP— capsule of lens ; CSC— cavity of sensory club ; EC — ectoderm ; EN— endoderm ;
ENC — endoderm of sensory club ; L — lens ; NC — network cells ; NF — nerve fibres ;
RT — retina ; SLA — supporting lamella ; VF — vitreous body.

Fig. 72 is a horizontal section through the large eye, and shows that
here, too, when the sections pass through the eye just radially, the

F. S. CONANT ON THE CUBOMEDUS^. 51

nuclei are not found at different levels sufficiently definite to suggest two
kinds of cells.

In the inner corner of the retina in the same figure (69) are seen
cells without pigment which show nuclei undoubtedly at different levels.
These cells in this position are a regular feature in the retina of the
smaller eye. Schewiakoff considers them purely visual, because of the
lack of pigment. In so doing it seems to me he forgets his own standard
for discriminating beween pigment and visual cells. The pigment cells
of the retina, according to him, are the same thing as the cone-shaped
supporting cells found elsewhere in the nervous epithelium, and are,
therefore, distinguished from the visual cells primarily by shape and by
position of nucleus, secondarily by the greater development of pigment.
When on the ground of pigmentation alone he calls the cells in the
corner of the retina visual, he judges them by only the second test, and
in so doing virtually admits, as it seems to me, that shape of cell and
position of nucleus are matters of no great moment. His own standards
place him in a dilemma. If on the other hand he judges by the lack of
pigment, the cells are visual ; if by shape of cell and position of nucleus,
they are both visual and pigment cells without the pigment or sup-
porting cells. What use there would be for simple unpigmented cells in
one limited region of the retina is hard to see, so he naturally takes the
other horn of the dilemma and calls them visual because they have little
or no pigment.

The distinction, then, between pigment and visual cells is brought
down to one of pigmentation only. Schewiakoff s test for this is that in
the visual cells "Das Pigment durchsetzt aber nicht das ganze Proto-
plasma des centralen Zellenabschnittes, sondern ist auf seine Oberflache
beschrankt (Fig. 19, sz), so dass der innere, axiale, stark lichtbrechende
Theil vollkommen frei von demselben ist." ('89, p. 37.) That is, in a
section through the ends of the retinal cells each pigment cell will
appear as a uniformly pigmented area, while each visual cell will appear
as a light, strongly refracting spot with a ring of pigment around its
periphery. This is the arrangement given in his Fig. 19.

An arrangement so definite ought to be easily made out in sections,
yet I have not been able to find it so. My sections show considerable
difference in the amount of pigmentation even in material preserved
with the same killing agent. If the retina is heavily pigmented the
ends of the cells have the appearance shown in Fig. 62, which represents
a portion of a cross-section. The ends are seen as clearly defined

52 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

polygonal areas differing among themselves in size, but not showing two
types of size, or two kinds of pigmentation, the one uniform, the other a
ring of pigment around a highly refracting central portion. If the
retina is but slightly pigmented — and some were so light as to make
depigmentation unnecessary — a difference is seen in the pigment, as
shown in Fig. 63, but in no case were areas found that showed a highly
refracting centre surrounded by a ring of pigment. (The unexplained
structures in Fig. 63 will be referred to a little later.)

Figures 59-62 are a series of four successive sections drawn with the
camera lucida for comparison with Schewiakoff's Figs. 20 and 19, and
to show that the presence of two types of cells plainly marked within the
retina by the position of the nuclei at different levels is at least not clearly
demonstrated. Only the nuclei are drawn, since the cell bodies are not
easily distinguished from the surrounding fibres. The eye is the same as
that from which Fig. 72 was made. Fig. 59 shows a relatively small
number of nuclei of slightly larger size than usual. These I take for two
reasons to be nuclei of the ganglion cells that are found in the fibres at
the base of the retinal cells (Figs. 58, gc, 69 and 72). They are the first
nuclei struck in tracing sections toward the retina, and in the series from
which Fig. 58 was taken similar nuclei appeared in both transverse and
radial cuts through the retina stained brightly and clearly with ha3ma-
toxylin, whereas the nuclei of the retinal cells proper were stained a
diffuse brownish-yellow from pigment that had evidently gone into
solution. Fig. 60 shows the closely aggregated, smaller nuclei of the
retinal cells surrounded by the nuclei of the outlying ganglion cells.
Schewiakoff's corresponding drawing ('89, Fig. 20) shows at this level a
definite alternation of the bodies and nuclei of un pigmented visual cells,
with the smaller, pigmented, proximal processes of the pigment cells. In
the next section (Fig. 61) the pigmented ends of a few of the cells have
been struck, and the following section (Fig. 62) shows that, in this
heavily pigmented specimen at least, there is no good evidence within
the retina itself of two kinds of cells, so that it is apparent that at any
rate we cannot accept Schewiakoff's conception of the structure.

(b) Yet the fibres that Schewiakoff observed and associated with
special visual cells occur beyond question. Fig. 64 is a drawing of the
first cut through the vitreous body of Charybdea, and in among the
sections of the pigment streaks are seen sections of processes lying
within clear spaces exactly as Schewiakoff figures his visual fibres ('89,
Taf. II, Fig. 18). That the fibres occur is indisputable, but as to the cells

F. S. CONANT ON THE CUBOMEDUS^E. 53

to which they belong I can say nothing except that from such evidence
as I have given in the preceding paragraph I conclude that they come
from pigmented retinal cells of not very different type within the retina
from the others, if different at all.

(c) On the third point, that the pigment streaks in the vitreous
body belong to underlying cells and are continued distally into fibrous
processes like the visual fibres of Schewiakoff, the evidence is decisive.
Fig. 58 has already shown it, and if this were not enough, a case of
unusual stoutness of the fibres drawn in Fig. 67 is conclusive. The
preparation from which the section is taken was one preserved with
corrosive-acetic, and I have drawn the outlines with the camera in order
to avoid exaggeration of the fibres as far as possible, and also to show
the shrinkage of the vitreous body (vb). It is the shrinkage of the
vitreous body that makes it so difficult to determine the exact relation of
structures seen in the vitreous body to the retina. The fibrous processes
run through the vitreous body to the "capsule" of the lens (cp) (see
also Fig. 72), a layer of homogeneous substance much resembling that of
the vitreous body, which is classed as a part of the vitreous body, but
usually in the shrinking adheres to the lens. The capsule is therefore
regarded by Schewiakoff as a secretion of the lens cells. Some fibres
were found by him to have the appearance of branching upon reaching
the surface of the capsule, others of passing through it and of seemingly
ending among the cells of the lens. The same appearances were given
in my sections. It is altogether impossible in the distal portion of the
vitreous body to distinguish between the fibres of Schewiakoff and those
that come from the long pigment cells. (Figs. 64-66 represent the
appearance of the vitreous body at successive levels, and are from the
same series of sections as Figs. 59-62 and 72.) In Fig. 64 the sections of
the processes that Schewiakoff calls visual are easily distinguished from
the sections of the long pigment cells. In Fig. 65, which is two or three
sections nearer the lens, the pigment cells are shown by their cross-
sections to be tapering down, and in Fig. 66. nearer still to the lens, the
two kinds of processes are no longer to be distinguished from each other.
In a few cases I have found pigment in a fibre which but for this would
be called one of the visual fibres of Schewiakoff. Such considerations as
these, the similar appearance in cross-section, the finding of pigment in
a few cases, and the inability to trace to any readily distinguished special
type of retinal cell, make me wonder whether the visual fibres of
Schewiakoff are anything more than the distal processes of pigment

54 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

cells, into which the pigment granules happened not to be produced at
the moment of fixation.

Fig. 63, however, where the retina was only slightly pigmented,
rather speaks against this view, for the number of darkly pigmented
areas seen here (which are shown beyond question by radial sections to
belong to the long pigment cells) is not great enough to account for the
number of both pigment areas and visual fibres of Schewiakoff seen in
such a section as Fig. 64. This would throw the visual fibres of
Schewiakoff back upon some of the slightly pigmented cells of Fig. 63,
otherwise not distinguished. I think the question cannot be settled
without the maceration of fresh material, and experiments upon eyes
killed in the light and in the dark.

In such cases as that of Fig. 63 it would seem conclusively shown
that the long pigment cells must belong to a different type from the
short, but as I have already said I can find no regularity in either their
shape or in the position of their nuclei. And on the other hand Fig. 58
shows that the reverse relation may obtain and the long cells be less
deeply pigmented on the edge of the retina than their shorter neighbors,
so that it looks as if all the short cells had to do was to project half their
pigment out into the vitreous body in order to become exactly like the
long ones. This they could do if, as is possibly the case, they are pro-
longed into "visual fibres" of Schewiakoff that have escaped observation
and so do not appear in the drawing.

Fig. 58 shows one more thing that is worthy of remark in passing.
In the preparation in which the vitreous body (at this point at any rate)
was not shrunken away from the retina, the fibre from each long
pigment cell does not lie in a clearly defined space or " canal," such as is
usually described as a constant structure of the vitreous body. Very
likely these canals are formed only by shrinkage around the fibres, and
the irregular shape of the spaces around the three fibres in Fig. 67 rather
bears out the same supposition.

As to the structure of the vitreous body, apart from the fibres and
pigment streaks already mentioned, I find it to be made up of prisms
extending from retina to capsule of lens, each containing a central axis
or fibre. Fig. 64 shows that the space around the pigment areas and
" visual fibres," instead of being homogeneous, is wholly filled with the
polygonal cross-sections of these prisms. In Charybdea they are gene-
rally more difficult to perceive than in my best material of Tripedalia
which was killed in acetic acid. In this the polygonal areas stood apart

P. S. CONANT ON THE CUBOMEDUS^E. 55

from each other more plainly. Curiously enough I have been unable to
demonstrate in Tripedalia the " visual fibres " of Schewiakoff. Here and
there were found spaces that at first sight reminded of them (Fig. 68, sh),
but they contained no central fibre, and were probably due to shrink-
age. The polygonal areas themselves, however, often contained a clear
spot in the centre, at one side of which would be found the cross-section
of the fibre, as is shown in many cases in Fig. 68. The clear spot is here
undoubtedly due to shrinkage of the gelatinous substance of the prism.

I think that these prisms and fibres are the direct continuations of
retinal cells. In a section such as that drawn in Fig. 63, which takes
just the very tops of the cells of a slightly pigmented retina, in the
centre of the section just grazing the space that lies between the retina
and the shrunken vitreous body, most of the cells toward the middle
(where especially the extreme tips are taken) show in their centres a dot
exactly corresponding to the dots in the polygonal areas of the vitreous
body. In the exact middle of the section, where only the cell walls
appear, slightly indicated, a dot is seen in each case. The size and shape
of the ends of the cells correspond with those of the polygonal areas in
the vitreous body, and I do not doubt that the latter are continuations of
the former. The vitreous body, then, instead of being homogeneous, is
composed of the clear highly refracting outer ends of retinal cells. The
assumption lies near that these are the true visual rods, but of course it
is assumption only.

To give a brief review, the points in which my conclusions differ
from those of Schewiakoff are as follows : I find (1) that the long pig-
ment streaks are parts of retinal cells continued into processes like his
visual rods ; (2) that the vitreous body is composed of prisms with central
fibres proceeding from retinal cells ; (3) that I am unable to get satisfac-
tory evidence of two types of cell distinguishable within the retina, and
at any rate find considerable evidence against the two types he distin-
guishes.

These results are not wholly satisfactory, for they leave us with
three kinds of fibrous processes in the vitreous body which for the
present we are unable to trace to three, or even two distinguishable
types of cell in the retina. It would be more pleasing if we could
confirm Schewiakoff s simple conception of the structure, with its one set
of visual rods in the vitreous body referable to a clearly marked type of
sensory cells in the retina, but I think the evidence that has been brought
up justifies the conclusion that in some respects he saw too much, in

56 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

other respects too little. This is not to be wondered at, since his mate-
rial, to judge from a single statement, consisted of but twelve marginal
bodies, and, moreover, the work on Charybdea forms but one portion
of a paper that is excellent for the clearness of its descriptions and illus-
trations.

Before leaving the subject I must mention that Wilson suggested
from his observations on Chiropsalmus that the vitreous body had a
prismatic structure, but he was probably mistaken when he thought he
found evidence of nuclei in it. Glaus says that the retina is composed of
pigment and rod cells alternating, and Wilson agrees with him, but
under a sketch of a sense cell from the nerve he makes the express state-
ment " not very well preserved." It seems very probable, therefore, that
he followed Claus's interpretation rather than independent observations,
and Glaus interpreted his results very much by analogy of what had been
found in other forms.

The smaller complex eye which is represented in Fig. 69 agrees in
structure very closely with the larger. The chief differences are that
sections do not show pigment extending into the vitreous body, that
there is no " capsule " to the lens, and that the lens seems to be supported
by a kind of stalk formed by a thickening of gelatine of the supporting
lamella (si). The gelatinous thickening lies between the lens and an out-
growth of endodermal cells (en) from the canal of the club. This
outgrowth is a constant feature, figured by Glaus and Schewiakoff for
Charybdea, and by Wilson for Chiropsalmus, and found in Tripedalia
also. The regularity of its appearance in all three genera leads one to
suspect that it may have some significance not yet understood.

Just above the smaller eye there lies a mass of cells of peculiar
structure (Fig. 69, nc). They are of a rounded polygonal contour, with a
comparatively small circular nucleus in the centre, and are found in
this region only. In and amongst them bundles of fibrous tissue are
found in the sections, which pass from the surface cells to the supporting
lamella. Glaus describes the contents of these cells as coarsely granular
protoplasm and says they cannot be taken for ganglion cells. He is
inclined to believe that they play the part of a special supporting tissue.
Schewiakoff, on the other hand, is convinced that they are ganglion cells,
and finds processes passing out from them ('89, Taf. II, Fig. 22). I find,
however, that the cell contours are perfectly regular and clearly without
processes, and it is incomprehensible to me how, if his material was at all
well preserved, he could for a moment have taken them for the same

F. S. CON A NT ON THE CUBOMEDUS^E. 57

thing as the big multipolar ganglion cells with large nucleus and
nucleolus which lie in about the same region and were correctly described
and figured by Claus but are not specially mentioned by Schewiakoff. I
cannot agree with Claus, however, that their contents are composed of
coarsely granular protoplasm. That which appears such by low magni-
fication shows itself under high powers to be a beautiful network with
thickenings at the nodes of the meshes, which is brought out very plainly
by a cytoplasmic stain such as Lyons blue. Around the nucleus is seen
a more or less well-defined clear zone. What the function of the cell is
remains as unknown to me as to Claus and Schewiakoff.

There is left one more point in reference to the nervous system upon
which I wish to say a word. Claus and Schewiakoff both describe the
wall of the crystalline sac as structureless, formed by the bare supporting
lamella. The credit is due to H. V. Wilson of finding in Chiropsalmus
that it has a special lining of epithelial cells, which he figures as a con-
tinuous, flattened layer. In both Charybdea and Tripedalia I find traces
of the same in nuclei here and there, but whether they are the remains
of a once continuous layer or not the sections do not show satisfactorily.

This ends the account of what it seemed worth while to say at
present upon the nervous system. In concluding, the writer wishes to
express his thanks for the help afforded by Dr. Wilson's notes, in par-
ticular on the subject of the vascular lamellae, and desires to make
especial acknowledgment of his indebtedness to Professor Brooks, whose
suggestions, based upon many years of experience with the Medusa?,
have been most welcome and helpful, and whose evidences of unfailing
kindliness, both in Jamaica at the time the material was obtained and in
Baltimore when it was being studied in the laboratory, take a most
honored part in the pleasant memories associated with the work.

LITERATURE REFERRED TO.

CLARKE, H. J. '78. Lucernarise and their Allies. Washington : Smithsonian
Institution.

CLAUS, C. '78. Ueber Charybdea marsupialis. Arb. aus d. Zool. Inst. d. Univ.
Wien, Band II, Heft 2.

DOFLEIN, F. '96. Die Eibildung bei Tubularia. Zeitsch. f. wiss. Zool., Bd. LXII,
Heft 1.

HAECKEL, E. '79. Das System der Medusen. Jena. — '81. Challenger Report on
the Deep-sea Medusae. Vol. IV.
8

58 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

HERTWIG, O. and R. '78. Das Nervensystem und die Sinnesorgane der Medu-
sen. Leipzig.

HESSE, R. '95. Ueber das Nervensystem und die Siunesorgane von Rhizostoma
Cuvieri. Zeitschr. f. wiss. Zool., Bd. LX, Heft 3.

MULLER, F. '59. Zwei neue Quallen von St. Catherina (Brasilien). Abhand-
lungen der naturf. Gesellschaft zu Halle.

SCHEWIAKOFF, W. '89. Beitrage zur Keiintniss des Acalephenauges. Morph.
Jahrb., Bd. XV, Heft 1.

WILSON, H. V. Unpublished notes.

TABLE OF REFERENCE LETTERS.

afr — adradial furrow.

afr' = furrow in Tripedalia that separates perradial from interrad. regions in
lower half of bell. (In Charybdea the same furrow is directly continuous with afr.}

ax — axis of nerve,
c = concretion.

cc = canal underneath ivl, connecting the two adjacent marginal pockets.

ccl = circular canal.

ci — cilia.
cm — circular muscle.

co = cornea.

cp = capsule of lens.

cs = covering scale of niche,
esc = canal of sensory club.

ct = canal of tentacle.

ct' = beginning of canals of lateral tentacles in Tripedalia.

ec = ectoderm.

ece = ectoderm of exumbrella.
ecs = ectoderm of subumbrella.

ed = distal paired eye.

el — larger unpaired eye.

en = endoderm.

enc = endoderm of sensory club.
enf — tract of nerve fibres underlying endoderm.
en ft = endoderm of floor of stomach.
enp = endoderm of stomach pockets.
enr = endoderm of roof of stomach.
ens = endoderm of stomach.

ep = proximal paired eye.

es — smaller unpaired eye.

fc — funnel leading into canal of sensory clubs.

fp = fibre from subepithelial plexus of subumbrella.
fph = filaments of phacellus.
frn — f renulum.

ft = funnel-shaped depression in ectoderm axial to base of tentacle.
g = gelatine.

F. S. CONANT ON THE CUBOMEDUS^E. 59

gc = ganglion cell.

ge = gelatine of exumbrella.

go — gastric ostium.

ys = gelatine of subumbrella.
hvl = horizontal vascular lamella.
i — interradius.

if— interradial funnel of bell cavity.
ifr = interradial furrow.

ivl = interradial vascular lamella.
I = lens.

lv = lip of valve.

m = bell margin.

me = mucous cell.
mep = mesogonial pocket.
mo — mouth.
mp = marginal pocket.

mp' — smaller marginal pockets, in Tripedalia.
mst — muscle of stock of sensory club.
mt = muscle at base of tentacle.
n = nerve.

nc = network cells, in sensory club.

nf — nerve fibres.
nm = nematocyst.

nst = nerve of stalk of sensory club.
osn = outline of sensory niche.

p =: perradius.

pe = pedalium.

ph = phacellus.

pr = proboscis.

r = reproductive organ.

re = remains of concretion.
rcl — radial canal.

rg = radial ganglion.
rm — radial muscle.

rn — radial nerve.
rns = root of nerve of sensory club.
rs ? = remains of stalk (?) of sensory organ.

rt = retina.
s = stomach.

sc = bell cavity.
scl — sensory club.
sen = supporting cell of nerve.

se = sensory epithelium.

sh = shrinkage space.

si = stalk of lens.
sla = supporting lamella.

sn = sensory niche.

.so = sensory organ in proboscis of Tripedalia.

sp = stomach pocket.
spk — stalk of phacellus.

ss — stalk of sensory organ, in proboscis.

st = stalk of sensory club.

60 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

su = suspensorium.
sub — subumbrella.
tl = lateral tentacle.

tm = median tentacle.
v = velarium.

va — vacuole.

vb = vitreous body.

vc = velar canals.

ve = edge of velarium.
vfs = visual fibres, according to Schewiakoff.

vg = valve of gastric ostium.
vl = vascular lamella.

vie = vascular lamella connecting vis with vim.
vim = vascular lamella of margin.

vis = vascular lamella of sensory niche.
vlst = vascular lamella of sensory niche at base of stalk.

we = wandering cells.
w-x-y-z = successive levels of Figs. 40-43 on Fig. 5.

DESCRIPTION OF FIGURES.

Fig. 1. Charybdea Xaymacana, from one of the four interradial sides.

Fig. 2. The same from above.

Fig. 3. The same from below, the four tentacles cut off.

Fig. 4. The same cut in halves vertically (or radially) through a perradius.

Fig. 5. The same cut in halves vertically (or radially) through an interradius.

Figs. 6-16. Diagrams of horizontal (or transverse) sections through C. Xayma-
cana at successive levels.

Fig. 17. Tripedalia cystophora, from one of the four interradial sides.

Fig. 18. The same from below.

Fig. 19. The same cut in halves vertically through a perradius.

Fig. 20. The same cut in halves vertically through interradius.

Figs. 21-30. Diagrams of horizontal sections through T. cystophora at successive
levels.

(The following are of Charybdea, except when specially stated otherwise.)

Fig. 31. Horizontal section through the suspensorium.

Fig. 32. Diagram of a gastric ostium seen from the stomach side.

Fig. 33. Diagram of a vertical section through a gastric ostium.

Fig. 34. Diagram of a horizontal section through a gastric ostium.

Fig. 35. Diagram to illustrate the formation of the central and peripheral gastro-
vascular systems of a Hydromedusa (a, b, and c) and a Cubomedusa (d).

Fig. 36. Vertical section through the upper part of the bell, adradial, to show
horizontal vascular lamella.

Fig. 37. Vertical section through the perradius, to show vascular lamella of the
niche of the margin.

Fig. 38. Vertical section a little to one side of the last, to show same structure.

Fig. 30. Horizontal section through the upper part of the sensory niche, to show
vascular lamella of the niche.

Figs. 40-43. Horizontal sections through the base of a pedalium at successive
levels, w-x-y-z. Fig. 5, to show marginal lamella.

F. S. CONANT ON THE CUROMEDUS.E. 61

Fig. 44. Diagram to show relations of sensory niche, of bell margin and
velarium in adult Tripedalia. The velarium represented as pendant.

Fig. 45. To show the same structure in a young Tripedalia.

Fig. 46. Horizontal section through the last just at the margin, to compare with
Fig. 29.

Fig. 47. Cross-section through the nerve ring.

Fig. 48. The structure of the nerve as seen by focusing at successive levels.

Fig. 49. Diagram to show the relation of the nerve ring to the sensory club.

Fig. 50. Horizontal section through the upper part of the sensory niche, to show
passage of nerve root through gelatine of subumbrella to stalk of sensory club.

Fig. 51. Vertical section through base of stalk of sensory club, to show same
passage.

Fig. 52. Similar section to last, but nearer to perradius, to show sub-endodermal
tract of nerve fibres.

Fig. 53. Sensory organ in proboscis of Tripedalia, as seen from surface in living
animal.

Figs. 54 and 55. Sections of same sensory organ.

Fig. 56. Vertical section through one side of proboscis, to show sensory organ
attached to endoderm. (Tripedalia.)

Fig. 57. Diagram of the outlines of sensory club seen from the side, by camera
lucida.

Fig. 58. Part of retina of larger complex eye cut radially.

Figs. 59-62. Four sections in direct sequence through retinal cells transversely,
larger eye.

Fig. 63. Transverse section through the tips of cells of a slightly pigmented
retina, larger eye.

Figs. 64-66. Three transverse sections through vitreous body at different levels.
All from same series, but not in direct sequence ; larger eye.

Fig. 67. Radial section through retina, to show fibres from the long pigment
cells ; larger eye.

Fig. 68. Transverse section through vitreous body of Tripedalia near retina.

Fig. 69. Vertical section through smaller complex eye.

Fig. 70. Wandering cells, Charybdea.

Fig. 71. Floating mass, from stomach pocket of Tripedalia.

Fig. 72. Horizontal section through larger complex eye. (See text figure, p. 50.)

VITA.

The writer of the foregoing dissertation, Franklin Story Conant, the
son of Farley Franklin and Emily Wilbur Conant, was born in Boston,
Massachusetts, September 21, 1870. His preliminary education was
received in the public schools of Boston and the vicinity. In 1887 he
entered the University of South Carolina, at Columbia, South Carolina,
where he spent two years. In the fall of 1890 he entered Williams
College and was graduated from there in 1893 with the degree of
Bachelor of Arts. In the spring of 1894 he joined the Marine Biological
Laboratory of this University at Beaufort, North Carolina, and in the
fall of the same year entered upon a course of studies for the degree of
Doctor of Philosophy, comprising Zoology as the chief subject, Physiology
and Botany as the subordinates.

Two months of the summer of 1895 were spent at Beaufort again,
chiefly in investigations upon the Chastognaths. In 1896 three months
were spent with the Marine Laboratory in Jamaica. In June of 1895 he
was appointed to a fellowship in Zoology, and in the following year was
made a Fellow by courtesy.

BIOLOGICAL, LABORATORY, JOHNS HOPKINS UNIVERSITY.

CUBOMEDUSAE.

PLATE

•//. 17.

afr'

Gilman Drew, del

fieliotype Co.,Eoeton.

CUBOMEDUSAE.

PLATE II

ivl

Hel-.otype Co,Boslf>n

CUBOMEDUSAE.

Fif. 6

PLATE

ftf-7-

FBConant.del.

~? >a

CUBOMEDUSAE.

Fif. 14.

PLATE IV

Fig. is.

54-

*\vlm

- jaton

CUBOMEDUSAE.

PLATE V

Fig. 31.

Fig.

Fig. ,3.

Fig. 14.

iff

Fig. 2b.

Fig. 27.

it r

T-? Co.Boetori

CUBOMEDU5AE.

PLATE VI.

, vim

nt.del

•

CU BO MEDUSAE.

PLATE VII.

. -17.

Heliotype Co,Boelon

CUBOMEDUSAE.

• <si

Pi,,. ,M.

PLATE VIM
Ky.63.

'*

Heliotype Co.,Boston

ittcmoirs from tl)c Biological Caboratovy

OF T1IK

JOHNS HOPKINS UNIVERSITY
IV, 2

WILLIAM K. BROOKS, EDITOR

SYNAPTA VIVIPARA

A CONTRIBUTION TO THE MORPHOLOGY OF THE ECHINODERMS

HUBERT LYMAN CLARK

(REPRINTED FROM THE MEMOIRS OF THE BOSTON SOCIETY OF NATURAL HISTORY)

BALTIMORE
THE JOHNS HOPKINS PRESS

1898

3. SYNAPTA VIVIPARA : A CONTRIBUTION TO THE MORPHOLOGY OF ECHINODERMS.

BY HUBERT LYMAN CLARK.

(Read November 3, 1897.)

CONTENTS.

1. Introductory G.'i (i. The development of the pentactula . . 63

2. History and systematic position of Synapta vivipara 54 7. The development of the adult Synapta . . 68

3. Distribution and habits 55 8. The anatomy of the adult 75

4. Fertilization and segmentation of the egg . . 58 9. Conclusion 79

5. Gastrulation and formation of the hydroeoel and 10. Literature 83

coelomic vesicles |>1 11. Explanation of plates . . 85

1. INTRODUCTORY.

In February, 1896, through the kindness of Dr. W. K. Brooks, there were placed in
my hands for study a number of specimens of a small brown Synapta from Jamaica. In
their body-cavities there were numerous young ones in all stages of development, so that
an excellent opportunity was offered for working out the embryology. So interesting
did this prove that I gladly availed myself of the privilege of spending the months of
May, June, and July, 1896, at Port Henderson, Jamaica, in the marine biological labora-
tory of the Johns Hopkins University. For such a privilege I am under the greatest
obligations to the authorities of that institution. During those months, I studied the
segmentation of the egg and the early stages of development from living material and
obtained an abundance of preserved material for further investigations. Since my return
I have been engaged in a detailed study of the development and anatomy of the animal
under the direction of Dr. Brooks, and it gives me great pleasure to acknowledge the debt
I am under to him for his suggestions and help.

The young were easily procured by cutting off the heads of the adults and thus
setting free the contents of the body-cavity. They could then be killed as desired, but
young taken from the body-cavities of preserved adults were fully as satisfactory for all
purposes. The best results were obtained by the use as a fixing agent of " corrosive-
acetic " (four parts corrosive-sublimate and one part glacial acetic acid), but excellent
preservation, especially for the adults, was secured by the use of picro-sulphuric or picro-

54 HUBERT LYMAN CLARK ON

nitric acid. Corrosive-sublimate alone gave poor results, while Perenyi's fluid proved
very unsatisfactory, the material prepared with it showing only fair preservation, and
staining very poorly. Very good material for the study of the calcareous bodies and
the development of the calcareous ring was obtained by the use of absolute alcohol. In
staining, it was found useless to try any solution which contained less than seventy per
cent alcohol, collapse and distortion of the tissues always resulting. Kleinenberg's
haematoxylin and eosin gave good results, but acid borax-carmine (10% alcohol) and
Lyons blue (70% alcohol) proved the most satisfactory. The early stages of the larvae
were oriented and imbedded in celloiden before imbedding in paraffine, and thus kept
their shape very well. Although material was so abundant and easily obtained, it is
impossible to tell the age of the different embryos, for none of the eggs obtained
developed beyond the blastula, and later stages would live outside of the mother for only
a very short time. Accordingly it cannot be determined how many days or weeks are
required for the growth of the pentactula or older stages, though there is reason to
believe that growth, at least early in life, is very rapid.

2. HISTORY AND SYSTEMATIC POSITION OF SYNAPTA VIVIPARA.

The Danish naturalist Orsted ('50) mentioned the discovery in the West Indies of
a new genus of the Synaptidae, the chief characteristic of which was its being viviparous.
To this genus he gave the name Synaptula, and the type species he called vivipara. In
his very brief and unsatisfactory account, for which I am . indebted to Ludwig's German
translation ('81), he mentions the occurrence of the form in " shallow-water," describes
the color as " greenish," and speaks of the presence of '' eyes," " skin-glands," and
" anchors." He never published anything further in regard to the species, and later
writers, as Bronn ('60), Selenka ('67), and Semper ('68), accepted the genus Synaptula
without comment. But Ludwig ('81), in describing a viviparous Chirodota (C. rotifera)
from the coast of Brazil, suggested that it might be the species on which Orsted had
based his new genus. This conclusion was adopted by Lampert ('85), and in his work
we find Synaptula vivipara given as a synonym under kirodota rotifera. Theel ('86)
placed Synaptula vivipara in the list of Synaptidae about which little or nothing is known,
but he suggests in agreement with Ludwig ('81) that the stage of development which the
eggs reach before leaving the mother, is not a satisfactory character upon which to base
a new genus. Ludwig ('86) described a small Sytiapta, found in floating seaweed, west
of the Abrolhos Reef, Brazil, which he regarded as identical with Orsted's species, since
it contained young in the body-cavity and answered Orsted's description, except in color

SYNAPTA VIVIPAHA. 55

In 1891, students of the Johns Hopkins University found at Port Royal, Jamaica, a
viviparous Synapta, and, in 1893, another party from the same institution brought back a
large amount of preserved material of this species, which they found abundantly about
Port Royal. It was this material which came into my hands, and after examining it and
studying the living animal at Port Royal, I had no difficulty in satisfying myself that it,
as \vt?ll as Ludwig's Abrolhos specimen, was indeed Orsted's Synaptula vivipara. A
brief description to establish its position in the genus Synapta, was published by Clark ('96),
and attention was there called to its similarity to S. picta Theel ('96). This species was
described from a single specimen in the Challenger collection, from Bermuda, and the
agreement of all its characters with those of S. vivipara was very striking. Dr. Theel
writes me that he has specimens of a viviparous Synapta from Bermuda in his hands at
the present time, which agree with S. picta and S. vivipara so completely that he has no
doubt that all three are the same species ; but the evidence is incomplete, owing to the
absence of the anchors and plates in his Bermuda specimens, caused by the killing agent
used. Although, as we shall see, Synapta vivipara differs in several important particulars
from all other Synaptas hitherto described, they are not sufficiently obvious to warrant its
separation, under the existing classification of the Synaptidae, as a distinct genus. Synap-
tula must therefore become a synonym of Synapta, and the synonomy of S. vivipara will
be as follows : —

Synapta vivipara (Orst.) Ludw. Zool. Jahrbiicher. Bd. 2, p. 28. 1886.
Synaptula vivipara Orsted. Vid. med. fra d. nat. For. i Kjobenhavn for 1849-50, p. vii.
Synaptula vivipara Bronn. Klas. und Ord. d. Thier., Bd. 2, p. 403. 1860.
Synaptula vivipara Selenka. Zeit. fur wiss. Zool., Bd. 17, p. 365. 1867.
Synaptula vivipara Semper. Reis. in Arch. Phil., 2. Theil, 1. Bd., p. 24. 1868.
Chirodota rotifera (in part only) Lampert. Die Seewalzen. Wiesbaden. 1885.
Synaptula vivipara Theel. Report on the Holothurioidea. "Challenger" Reports.

Zool., Vol. 14, p. 32. 1886.
Synapta picta Theel., p. 10. Report on the Holothurioidea. "Challenger" Reports.

Zool., Vol. 14, p. 32. 1886.

3. DISTRIBUTION AND HABITS.

Orsted's specimens of Synapta vivipara, he tells us, were from the West Indies,
and all of the specimens I have seen, came from Jamaica. Ludwig's single specimen came
from the Abrolhos Reef off the coast of Brazil (18° S. lat.), while all of Theel's specimens
are from Bermuda (32° N. lat.). We may therefore conclude that the species is pretty

56 HUBERT LYMAN CLARK ON

widely distributed throughout the eastern Atlantic ocean, wherever suitable conditions
are found. In Jamaica, however, the species is extremely local and was found in only
one place, the so-called " lakes " at Port Royal. These " lakes " are parts of the harbor
which have been wholly or in part cut off by the growth of mangroves, so that they are
very quiet bodies of water, though not at all stagnant or brackish. On the roots of the
mangroves, which hang down in the water on all sides, is an abundant growth of
vegetable and animal life. In some places, a particular sea-weed, one of the Florideae,
crowds out all the other Algae. In this weed Synapta vivipara makes its home, and though
carefully looked for elsewhere, it was found in numbers only in such situations. The
late Dr. J. E. Humphrey kindly identified this alga for me, as Acanthophora thierii
Lamouroux. This weed also grows in large bunches on the bottom in the shallow water
of the harbor just outside the " lakes,'' and I was told that in 1893 the Synapta was
found in great quantities there, but in 1896 it seemed to have entirely disappeared
from that place. At Montego Bay, on the northwest coast of Jamaica, Acanthophora is
very abundant on the mangrove roots, but a thorough search revealed no sign of
Synaptas there. Even in the Port Royal "lakes," their distribution was very capricious,
and only certain favored masses of Acanthophora contained them in any numbers. They
seem to be quite social in their habits, and usually if one or two were found, there would
be a whole colony of them. They are very sensitive to changed conditions, and I was
unable to keep them alive in aquaria more than twenty-four hours. The anchors in the
body-wall are so abundant and prominent that they cling very tenaciously to anything
with which they come in contact, especially the hands, and it is accordingly no easy task
to disentangle them from the sea-weed without injury. They seem to be able to swim
very little, and it is doubtful if they ever leave the bunch of sea-weed, in which they
have once settled. Their food consists largely of vegetable matter, diatoms being
abundant in the stomach, but probably many small crustaceans and worms are also eaten.
The tentacles are kept constantly in motion, and it was very common to find small
amphipods caught among them, but I was unable to find evidence that these crustaceans
ever served as food. Semon ('87) has called attention to what he considers a mimicry
of coloration in Synapta inhaemns, in relation to the bottom on which it is found. In
this connection, it is interesting to note that the reddish and greenish brown shades
of S. vivipara are almost exactly those of the Acanthophora in which it lives. So close
is the resemblance that it is very easy to overlook Synaptas, even when the sea-weed is
in one's hand. Whether this coloration is actually protective or not is doubtful, for they
seem to have few, if any, enemies. No internal parasites were observed ; externally
however a small brown calcareous sponge, like Grantia, was found firmly attached

SYNAPTA VIVII'AKA. 57

to the .skin, just behind the circle of tentacle's. When placed in aquaria, tliis
Synapta does not break up by muscular contractions, like S. inhaerans, nor does it
ordinarily eviscerate like many holothurians. but after crawling about restlessly for a
while, it stretches out on the bottom and dies, almost without the contraction of a muscle.
In a few cases, evisceration at the mouth took place, when they were thrown into a killing

agent.

. Breeding goes on all through the spring and summer ; and there is no evidence to
show that it does not go on all the year round. My earliest specimens, collected April 30
contained many well-developed young, and up to the end of July in all the specimens
obtained, young were found, while Ludwig's specimen from Abrolhos was collected in
September and contained very young embryos. The number of young in the body-
cavity of a single adult varies greatly, depending more or less on the size of the individ-
ual. Specimens not over a centimeter long may contain a few, while some very large
ones have scores. The largest number I have found is 176. It is a curious fact that the
young are almost always in two broods ; that is, a certain proportion of them will all
have reached a given age, say that of the pentactula, while the remainder will be much
younger, say about that of the gastrula. When the young are very few, they will all be
the same age, while if they are very numerous, they will sometimes show three different
stages. This fact seems to indicate that the eggs ripen and pass into the body-cavity in
lots of from six to a hundred, and that several days elapse before another lot is ripened.
Regarding the length of time during which the young remain in the body-cavity, it is
impossible to make even an estimate. Animals kept in aquaria frequently gave birth to
young only five mm. long, and it was not usual to find much larger specimens in the adults
examined, but sometimes a young one, fifteen or twenty mm. long, with all the charac-
ters of the adult, would be found still inside its mother. Observations made on the living
animals showed that birth occurs normally at the posterior end of the body, apparently
through the anus. Investigation showed that this was accomplished by a rupture of the
body-wall, which may be through the skin some little distance from the anus (Fig. 29),
or, as seems to be more generally the case, through the wall of the rectum close to the
point where it joins the external body-wall, the young passing out through the anal
opening. It may be that the openings in the wall of the rectum (Figs. 30 and 31), to be
described later, are concerned in the birth of the young, but they seem to be too small to
be of any service in this connection, except possibly as starting points for the rupture of
the rectum wall. Under abnormal conditions, I found by experiment, rupture of the
body-wall and consequent birth of the young may occur at other points than near the anus.

58 HUBERT LYMAN CLARK ON

4. FERTILIZATION AND SEGMENTATION OF THE EGG.

Like all other known Synaptas, S. vivipara is hermaphroditic. Ripe spermatozoa and
ova are found in the same genital organ and even in the same branch, and to judge both
from living material and preserved specimens, even at the same time; but this latter
point could not be proven since it is impossible to determine simply by observation, when
the egg is mature. The branches of the genital gland are nile-green in color in the living
animal, while the fully grown ova are brownish yellow, and the spermatozoa, when in
any quantity, appear to be white. In shape, the latter are like those figured by Jourdan
('83) for Holothuria tubulosa. The mature egg (Fig. 1) is about '200/* in diameter and
before fertilization is not provided with any membrane. It is full of yolk material but
comparatively transparent, so that the internal changes could be watched up to a late
stage of development. Artificial fertilization proved unsuccessful, although attempted
several times. Whether self-fertilization takes place or not could not be positively
decided, but that it is at least very improbable seems clear from the structure of the
genital organs and the probable manner of egg-laying. The genital gland of the adult
Synapta lies just above the oesophagus, with one or two branches on each side of the
dorsal mesentery and with the genital duct lying in that mesentery and opening to the
exterior close to the base of one of the mid-dorsal tentacles. A cross-section of one of
the branches (Fig. 40) shows that it consists of an external covering, the continuation of
the epithelium, a very scanty connective-tissue layer consisting only of a few scattered
mesenchyme cells, and a much folded germinal epithelium surrounding the lumen of the
gland. Of the layer of circular muscle fibers which was found in the genital gland
of S. digitata and S. inhaerans by Quatrefages ('42), Baur ('64), and Hamann ('84), I
have found no trace in S. vivipara. The branches of the gland contract vigorously after
being cut from the body, but so far as I could see, these contractions were alwa}rs longi-
tudinal. The germinal epithelium is more or less plainly made up of two or more layers
of cells of which the external are the larger. As will be seen from figures 39 and 40,
from this external layer of cells the ova arise and so come to lie between the germinal
epithelium and the epithelium of the body-cavity, and do not pass into the lumen of the
gland at all. This arrangement is quite the reverse of what Cuenot ('91) has shown to
exist in S. inhaerans, and gives us a clue as to how the eggs get into the body-cavity of
the mother. When the eggs are mature, they press so closely against the external
epithelium of the gland that it bulges out sufficiently to be seen with the naked eye, in
the living gland. Sections show that the epithelium over such ova is so stretched as to
be thinner than elsewhere (Fig. 39), and probably the eggs enter the body-cavity simply

SYNAPTA VIVIPARA. 59

by the rupture of the epithelium at that point. While this has never been actually
observed, the theory is supported by a further examination of the genital gland. The
smaller internal cells of the germinal epithelium give rise to spermatozoa, which are
almost invariably to be found in the lumen of the gland, never on the outside with the
ova. If we trace the lumen forward we find it passes directly into the lumen of the
genital duct, the internal walls of the latter actually being formed by a continuation of
the germinal epithelium, which has become of uniform thickness and ciliated internally
(Figs. 36 and 37). The lumina of the two or three branches of the gland unite on
entering the duct (Fig. 36) and pass through the latter upwards towards the body-wall.
At no point is there any sign of communication between the genital duct and the space
external to the germinal epithelium in which the ova lie. If we follow the genital duct
upwards to where it comes in contact with the body-wall, we find it does not fuse with
the ectoderm and open at once to the exterior, but simply lies in the connective tissue
with its end against the ectoderm (Fig. 38), and although there are probably openings
through which the spermatozoa pass out, they are so extremely small I have not been
able to demonstrate them satisfactorily. In nearly all the specimens examined the
genital duct contained large quantities of spermatozoa, but in no case was there any trace
of an ovum. That the openings at the end of the duct are not directly continuous with
openings through the body-wall is indicated by the occurrence of spermatozoa, sometimes
in large quantities, in the connective tissue surrounding the terminus of the duct (Fig.
38). From all these facts I am convinced that the ova pass into the body-cavity by a rup-
ture of the peritoneal epithelium, while the spermatozoa pass outward through the genital
duct to the exterior. In Cucumaria glacialis, the only other viviparous holothurian
concerning whose breeding we have any information, Mortensen ('94) thinks the eggs
are laid on the bottom and taken up afterwards into the brood-sacks of the mother. But
it is manifestly impossible that the eggs of S. vwipara could get into the body-cavity in
any such way. The next question that arises is, how do the spermatozoa reach the ova
inside the body-cavity of the mother, and the answer brings to light another interesting
modification of structure, adapted to the viviparous habit. Careful examination of the
rectum shows that through its wall there are direct channels of communication between
the body-cavity and the exterior through the anus. The wall of the rectum is folded
and ridged longitudinally, and at certain places, parts of these ridges have pushed out and
fused with invaginations from the surrounding coelomic wall, forming distinct tubes
connecting the interior of the rectum with the interior of the coelom (Figs.
30 and 31). No trace of valves or cilia was found in any of these tubes, but
the passage of water in and out could be easily regulated by the opening and

60 HUBERT LYMAN CLARK ON

closing of the anus. With the anus open, each muscular contraction of the animal
would tend to either force water out or draw it in through these openings, and in
that way spermatozoa could easily get into the body-cavity and thus fertilization
could take place within. This will appear more probable when it is remembered
that the animals are very social, and that the water around a mangrove root on which
there are hundreds of them, must contain countless spermatozoa. There is also a
possibility that spermatozoa enter the body-cavity, through the water-pore and stone-
canal, which, as we shall see, remain open in the adult, and also open into the body-cavity.
After fertilization a membrane forms around the egg and segmentation begins. It
seems probable that the extrusion of the polar bodies occurs before the formation of this
membrane, as Selenka ('S3) found no trace of them in the segmenting eggs of 8. digitata,
and I could not find them in any of the eggs of 8. vivipara which I examined, although
they were looked for with special care. Segmentation and the formation of the blastula
occur in practically the same manner as has been so well figured by Selenka ('83) for
S. digitata. The first plane of division forms two blastomeres of equal size and appear-
ance (Fig. 2). After a resting period of about twenty minutes, the second plane of
division occurs at right angles to the first, giving rise to four similar blastomeres (Fig.
3). The third plane is at right angles to the first two, and we now have an embryo of
eight equal cells with a segmentation-cavity between them (Fig. 4). The sixteen-cell
stage (Fig. 5) soon follows, the division plane being at right angles to the preceding.
The appearance of the embryo at this stage is very peculiar and characteristic, the cells
being arranged in a band or ring and the segmentation-cavity being open at each pole.
Another plane of division, again at right angles to the preceding, doubles the width of
the band and decreases the openings at the poles, but it does not divide the cells exactly
equally, so that the upper- and lowermost rows are of somewhat smaller cells than the two
middle rows and have a less diameter (Fig. 6 and 7). The subsequent divisions occur
with a fair degree of regularity in alternating plane*, each division decreasing the
openings at the poles until at last they are entirely closed. This occurs when the embryo
consists of approximately 256 cells, and so the blastula is formed (Fig. 8). The cells of
the four equatorial rows are somewhat larger than the rest, but the difference is not at
all noticeable and apparently has no significance. The divisions have followed on each
other with great rapidity so that the complete blastula is formed after about four hours,
while in S. digitata, according to Selenka ('83), the blastula is the result of twelve hours'
growth. It is for this reason, that I am inclined to think that the whole process of devel-
opment in S. vivipara is very rapid.

SYNAPTA VIVIPARA. 61

5. GASTRULATION AND FORMATION OF THE HYDROCOEL AND COELOMIC VESICLES.

Invagination of one of the poles of the blastula soon forms the archenteron from the
blind end of which the mesenchyme cells now begin to arise, some of the endodermal
cells being simply crowded out into the segmentation-cavity. I could find no evidence at
all of Selenka's ('83) two primitive mesenchyme cells, but on the contrary, I found cells
all over the archenteron which were in various stages of passing into the segmentation
cavity. On the other hand in not a single case were there found any of the ectoderma-
cells forming mesenchyme, and I feel no hesitation in affirming that the mesenchymel
arises exclusively from the endodermal cells, contrary to Ludwig's ('91) observations on
Cucumaria. The number of mesenchyme cells is comparatively small, and they never
become so numerous or play so important a part in larval structures as they do in S.
digitata, judging from Semon's ('88) account and figures. The completed gastrula (Fig.
9) is covered with cilia which are easily seen in the living specimens, but in none of my
preserved material has it been possible to demonstrate them. In a few cases gastrula
were found free from the egg-menibrane and moving about actively in the fluid of the
body-cavity, but the great majority are still enclosed in the membrane within which they
rotate by means of the cilia. The membrane may be retained, as preserved specimens
show, until long after the coelomic vesicles are formed, and I am inclined to believe that
there is no definite time when it is cast off, but that it ruptures and is lost whenever the
larva has grown too big for it. As the archenteron increases in length it bends to one side
and unites with the wall of what subsequently becomes the dorsal surface of the larva.
Its lumen breaks through the surface, and thus the water-pore is formed, as described by
Selenka ('83) for S. digitata. Meanwhile the gastrula loses its spherical shape and
becomes more or less elongated (Fig. 10) . The cells at the end opposite the blastopore
are already somewhat different in form from those elsewhere and make a sort of plate of
thickened ectoderm (Fig. 21) which may correspond to the so-called "neural plate" of
other echinoderm larvae. This plate, however, does not lie exactly opposite the blastopore,
but somewhat toward the ventral side of the larva, and as the gastrula increases in length,
it comes to lie more and more on that side. At the same time, the archenteron continues
to increase in length and grows forward and at the same time ventralward, thus drawing
away from the water-pore. In so doing, that part of it which grows forward pushes by
the part opening through the water-pore, on the right-hand side (looked at from the
dorsal surface) so that the latter comes to lie on the left side of the larva (Figs. 11 and
12). As the archenteron grows it completely severs its connection with this vesicle, and

62 HUBERT LYMAN CLARK ON

pushing onward and downward against the thickened part of the ectoderm, becomes
attached to it, and the mouth breaks through at that point. The mouth may be formed
before the separation between the archenteron and the other vesicle is completed. Mean-
while the latter increases in size, and its walls become thinner. It grows backward toward
the blastopore especially and soon becomes constricted and divided into two vesicles, the
anterior of which is connected with the exterior by the water-pore, while the posterior is
entirely closed and lies beside the posterior part of the stomach, if we may so designate
the middle section of the archenteron (Fig. 13). Soon afterwards this posterior vesicle
grows out laterally to the right, across the dorsal surface of the larval hind-gut, and forms
on the right-hand side a vesicle like itself, which soon becomes entirely separate from it
(Fig. 14). The anterior of these three vesicles is the rudimentary hydrocoel, while the
pair of posterior pouches represent the right and left coelomic vesicles. Their mode of
formation is essentially the same as that described by Selenka ('83) for 8. digitata, but
the relative size of the various organs is markedly different, so that the figures of the same
stages show almost no resemblance. If figure 15 of Synapta vivipara be compared with
the auricularia of about the same age which Semon ('88, Plate 6, Fig. 2) figures, this
difference will appear in many ways. The rudiments of the two coelomic pouches and
especially of the hydrocoel are very much larger relatively in S. vivijjara than in 8. digi-
tata, while the difference in the digestive tract is even more noticeable. In the Eiu-opean
species there is a well-marked differentiation into fore-, mid-, and hind-gut, while in the
Jamaican form there is no such distinction evident. A remarkable difference is also to
be seen in the external form of the larvae. That of 8. vivipara has retained its elliptical
shape and shows no trace of ciliated bands, while the other has assumed the familiar
auricularia form. Furthermore, there is no trace of calcareous structures of any kind in
8. vivipara to correspond with the plates at the posterior end of auricularia. In the
latter also, Semon ('88) has figured and described a larval nervous system, but after care-
ful investigation of this point, I am convinced there is none in the larva of 8. vivipara.
In external form, the latter is very regular, but with the growth of the coelomic pouches
the ectoderm of the dorsal surface begins to become thinner and more flattened, while that
around the mouth, especially posterior to it, shows a tendency to. increase in thickness.
This difference between the dorsal and ventral surfaces may sometimes be seen in very
young larvae and is clearly shown in figure 22. Before the two coelomic vesicles have
entirely separated, the hydrocoel has become considerably larger and begins to grow
anteriorly and toward the right, while about the same time five outgrowths begin to appear
on that side which is furthest from the fore-gut (Fig. 15). Soon after their appearance
five other much smaller outgrowths arise, one at the right of each of the first five. The

SYl'ATA VIVIPARA. 63

first series gives rise to the five primary tentacles, and the second series corresponds to
those which in X. diyitata u'ivc rise to the radial water-canals. In S. diyittifn however,
there is a sixth outgrowth at the extreme left, the rudiment of the Polian vesicle, but in
S. viuipara this vesicle is not formed until after the closure of the hydrocoel ring. The
water-canal enters this ring at a point just between the fourth primary tentacle (counting
from before backwards) and the fourth secondary outgrowth (Fig. 23). In this particu-
lar my observations confirm Bury ('89) in opposition to the statements of Semon ('88).
Before the outgrowths of the hydrocoel are very evident, that part of it which lies
dorsally and in immediate connection with the water-canal bulges out, becomes thinner-
walled than the rest of the hydrocoel, and gradually separates from it, but before
the separation becomes marked, the outgrowth diminishes in size and with the
increasing growth of the water-canal disappears. This structure, I believe, is the
"anterior coelom " which Bury ('95) has shown to exist in the auricularia of
S. diyitata. It is very marked in some specimens of S. vivipara (Fig. 24), and I
see no reason to doubt Bury's interpretation. He does not make very clear what the
ultimate fate of this coelom is in X. diyitata, but leaves the impression that it is connected
with the subsequent formation of the madrepore plate, as Ludwig ('91) considers it to be
in Cucumaria. In S. vivipara however, there is seldom any trace of it left after the
hydrocoel ring closes, and there is no reason to suppose that it has any connection with
the much later madrepore openings. About the time of the appearance of the primary
tentacles, the larval anus, which was the original blastopore, closes entirely and the
digestive tract ends blindly. Accordingly we now have a regular elliptical larva, about
a third of a millimeter long, with the ventral ectoderm much thicker than the dorsal,
without ciliated bands, calcareous particles, or nervous system, a mouth on the anterior
ventral surface but no anus, a well-developed coelomic pouch on each side of the digestive
tract, and a hydrocoel with five primary tentacles and five secondary outgrowths, opening
to the exterior through the dorsal pore, by means of an adradial water-canal, upon
which may still be seen the vestige of an anterior coelom.

6. THE DEVELOPMENT or THE PENTACTULA.

In such larvae, the ectoderm of the ventral surface continues to thicken and before
long is sharply set off from the ectoderm of the rest of the body, which consists of a
single layer of cells. The thickened ectoderm forms a circular field around the mouth
though the latter does not lie at its center, but nearer to its anterior edge. This circular
disc gradually sinks below the level of the rest of the ectoderm, and the latter grows in

64 HUBERT LYMAN CLARK ON

over it from all sides, until the disc lies at the bottom of a shallow cavity, the so-called
" atrium," which opens on the ventral surface through a small pore (Fig. 16). This
pore does not lie directly over the mouth, but either before or behind or somewhat to
one side. While these external changes are taking place, the hydrocoel has continued its
growth, the anterior end passing across on to the right side of the larva where it bends
downward and backward around the oesophagus to meet the posterior end near the middle
line. The primary tentacles have increased considerably in size and are growing up
around the floor of the atrium, while the secondary outgrowths also grow upward beside
them. No Poliau vesicle has yet appeared, and no marked distinctions between the
different parts of the digestive tract are to be seen, but the latter has become very much
arched toward the dorsal surface, and the lumen of the middle section is larger than at
either end. A new anus may have been formed by the end of the hind-gut growing to
the body-wall on the ventral side and an opening breaking through, but in some cases
the definitive anus does not appear until the pentactula form is nearly attained. The
most important changes have been going on meanwhile in the coelomic pouches,
the growth of which has been very rapid. The left vesicle has grown more ante-
riorly than the right, and sends forward two finger-like processes, one of which passes
across the median line into the right side of the larva, while the other grows up to
the inner side of the hydrocoel ring, above the most posterior tentacle, and follows the
course of that ring around the oesophagus (Figs. 26-28). These anterior prolongations
of the left coelom were observed by Bury ('89 and '95) in 8. digitata, but have appar-
ently been overlooked by other investigators. The one which passes on to the right
side of the body fuses so soon with the right coelom that I have been unable to confirm
Bury's further observations regarding it, and in 8. vivijiam its appearance might easily
be entirely overlooked. But the prolongation which passes to the hydrocoel remains
distinct through all the later stages of the larva, and its subsequent changes are easy to
trace. It soon loses its connection with the left coelom and forms a tube, lying on the
oral surface of the hydrocoel ring. Meanwhile the right and left coelomic pouches have
met and fused on the ventral side of the digestive tract, while dorsally they are still
separate, the right pouch extending considerably further back than the left. The larva
lias now reached the condition shown in Fig. 16. In its subsequent growth the atrium,
with its thick ectoderraal floor and narrow opening to the exterior, moves to the anterior
end of the larva, where it finally comes to lie, along with the mouth, oesophagus, and
hydrocoel ring. The latter has now closed, without the formation of a Polian vesicle,
somewhat to the left of the middle Hue apparently (Fig. 25), but I am not sure that the
point of closure is always on the left side, for it is by no means easy to determine the

SYNAPTA VIVIPARA. 65

exact mid-line. With the closure of the hydrocoel, occurs the union of the two ends of
that coelomic tube which lies on its oral surface, so that we now have a circular sinus
around the oesophagus just above the water-ring. This sinus is very evident in young
Synaptas, and Bury's ('95) surmise regarding its origin from the left coelom is entirely
correct. The primary tentacles are growing upward, not pushing the floor of the atrium
out before them, as Semon ('88) says, but enclosing the atrium within their circle so that
the thickened sensory epithelium, which they subsequently possess, does not arise as
Semon describes. It is clear, from Figs. 16 and 17, that his description could not possibly
apply to S. ririjmra. The secondary outgrowths remain nearly unchanged in size and
show no sign of bending backward to form radial canals. The Polian vessel is formed as
an outgrowth on the inner side of the hydrocoel ring in the left dorsal interradius, as
Ludwig ('23) found it to be in Cucumaria.

The digestive tract grows with greater rapidity relatively than any of the other
organs. Accordingly, the oesophagus pushes upward toward the atrial opening, so that
the thickened ectodermal floor of the atrium lies surrounding it, in the form of a poorly
denned circumoral ring. Continued growth pushes the oesophagus against the upper
ectodermal wall of the atrium, and with that it fuses, leaving the circumoral ring entirely
cut off from the outer ectoderm of the body (Figs. 87 and 88). Meanwhile the growth
of the primary tentacles has pushed this body-wall upward and outward, so that the
narrow slit-like opening of the atrium is gradually widened until it finally disappears,
leaving the ectodermal-covered, anterior end of the oesophagus to form the definitive
mouth, in the center of the circle of tentacles. This process is not completed however,
until the pentactula form is fully assumed. The differences between this development
of the mouth and circumoral ring and that given by Semon ('88) for 8. digitata are
almost irreconcilable, but they are all dependent on the question, whether the five
primary tentacles push up through the floor of the atrium or grow up around it. The
latter is certainly the case in S. cioiparrt. While these changes are taking place anteri-
orly, the hind-gut has increased in length so that it has not only arched still more toward
the dorsal surface but has bent on itself and formed a loop lying to the left of the
stomach. The coelomic pouches already united and forming a single cavity ventrally,
have met in, or close to, the mid-dorsal line and by the union of their walls have formed
the dorsal mesentery. This mesentery follows pretty closely the curve of the intestine
and attaches it throughout its course, to the body-wall. The water-canal lies in the
anterior part of the mesentery, but whether that part was formed in a different manner,
as Bury ('95) thinks probable, it is impossible to say from observations on S. vivipara.

Meantime most important changes are going on in the circumoral ring of ectoderm

gg HUBERT LYMAN CLARK ON

which we have seen was formed from the floor of the atrium. This ring is the
beginning of the central nervous system, and from it the tentacle and radial nerves arise.
As the primary tentacles push upward past the atrium, they lie closely appressed to its
floor and wall. They retain this position even after the circumoral ring is pretty clearly
denned, and before they have grown much above it, the radial nerves appear between
each pair of them as outgrowths of the central ring. These outgrowths pass directly
over the secondary outgrowths of the water-vascular system and bend backwards to run
toward the aboral pole of the body. Very soon after they appear (not before them, as
Semon ('88) describes for >S". digitata), that part of the circumoral ring appressed to each
primary tentacle begins to grow upward with it on its inner side, forming the tentacle
nerves. As the tentacles continue to grow and press the anterior body wall outward, the
ectoderm which covers them at the tip becomes noticeably thickened, especially on the
outer side (Fig. 54), and apparently assumes a sensory function, probably in connection
with the tentacle nerve. The formation of the otocysts of Thomson ('62), the "horor-
gane " of Baur ('64), takes place as described by Semon ('88). They arise by evagina-
tions from the outer side of the circumoral ring close beside the outgrowths which form
the radial nerves. With the growth of the latter, the otocysts come to lie external to
them at the point where they bend backward, and in this position the sense-organs remain
throughout life. It is a very evident and noteworthy fact that the development of the
radial nerves and otocysts does not take place at the same time in the five radii, but there
is a marked difference between them. The first to appear is that nerve which subse-
quently indicates the mid-ventral radius, and with it appear its two otocysts. The two
lateral ventral nerves appear next, and with them the otocyst which accompanies each one
on its ventral side. The lateral dorsal nerves next appear (Fig. 1), and very soon after-
wards the two other otocysts of the right and left ventral radii are formed. Last of all
to develop are the otocysts of the right and left dorsal radii. This sequence in the
appearance of these nerves and sense-organs is probably connected with the fact already
mentioned, that the thickened ectoderm which made up the floor of the atrium did not lie
symmetrically around the mouth, but the greater part of it was posterior or, when the
mouth lies at the anterior end of the larva, ventral to it. What the significance of this
condition may be, I am unable to suggest, but it is interesting to note that in Cucumaria
Ludwig ('91) found the ventral radius the most advanced in development. Any possible
similarity ends here however, for the development of the other radii was quite the reverse
in Cucumaria of what it is in Synapta.

About the time of the completion of the pentactula form, there appear in the ecto-
derm of various parts of the body peculiar invaginations (Fig. 41) which are finally

SYNAPTA VIVIPARA. 67

connected with the exterior only by a very narrow canal (Figs. 42 and 43, and Fig. 17,
Igo.). These organs appear, from their structure and the great variation which they show
in staining, to be of a glandular nature, and I am inclined to think they maybe connected
with the absorption of nourishment from the fluid of the body-cavity of the mother, I'm-
they never increase in size, are most abundant in the young with ten tentacles, and seem
to have entirely disappeared in the adult, and finally, nothing of the kind has been described
for any other holothurian. When fully grown they measure about fifty mikrons in diam-
eter and somewhat less in depth. They consist of very long, clear cells, with nuclei at
the extreme distal ends, surrounding a more or less spacious lumen which opens to the
exterior by a narrow canal of ordinary epithelial cells. Sometimes the clear cells stained
heavily, but often they did not stain at all.

Up to this time the mesenchyme cells have played no part in the development of the
larva. In Fig. 16, they are shown as they appear scattered almost uniformly through the
segmentation cavity. Shortly after this, however, they begin to gather around the lower
and outer edge of the water-vascular ring, and by the time the pentactula stage is reached
they have begun the formation of the calcareous ring. Contrary to Semon's ('88) views
on S. digitata, and in accordance with Ludwig's ('91) observations on Cucumaria, I have
found no evidence at all of any mesenchymatous musculature on the oesophagus. The
first products of the mesenchyme cells to appear are five small straight rods between the
primary tentacles but outside and somewhat below the hydrocoel ring. Soon after these,
five more appear below the bases of the tentacles, so that there are now ten rods, five
radial and five interradial, and they continue in this position so long as there are only five
tentacles. I saw no evidence at all of any such shifting of position of the first five rods
as Semon ('88) records for S. digilata, nor could I consider the position of the second
series as agreeing at all with his description. Very soon after the appearance of the ten
rods, they fork at the ends and begin to branch very irregularly. As will be seen from
Fig. 46 a-i, the divisions occur at all sorts of angles and not only differ decidedly from
Semon's figures of the same rods in S. digitata but show no sign at all of following his
law ('87) for the formation of calcareous plates in Echinoderms. The much-branched
ends of the rods come into very close contact but evidently never mingle, for the plates
into which they develop are at all times easily separable and the line of division between
them is practically straight. Very soon after the appearance of the first five calcareous
rods, the radial nerves grow backward over them and before the completion of the pentac-
tula, run to the posterior end of the animal.

We have now reached the complete pentactula form, a slightly older stage of which
appears in Fig. 18. The pentactula is about half a millimeter long and its characteristics

68 HUBERT LYMAN CLARK ON

may be briefly summed up as follows: — Water- vascular system consisting of a closed
hydrocoel ring or circumoral water-tube with five primary tentacles, between which are
five very much smaller but equally erect secondary outgrowths; a water-tube in the mid-
dorsal interradius connecting the circumoral ring with the exterior; and a Polian vessel
in the left dorsal interradius. Nervous system consisting of a circumoral ring ; five ten-
tacle nerves on the inner face of the primary tentacles, the ectoderm of which is consid-
erably thickened, especially on the outer side; five radial nerves bending backward over
the secondary outgrowths of the water-ring and over the radial pieces of the calcareous
ring, and running to the posterior end of the body; and five pairs of otocysts, lying
external to the radial nerves, where they bend backward. Digestive system, consisting
of a short oesophagus with the mouth opening anteriorly in the center of the circle of
tentacles, a large stomach, a comparatively short intestine with a single loop in it, and
usually an anus formed secondaril}' near or at the aboral pole. Digestive system attached
to the wall throughout its whole course by a mesentery, formed by the union of the two
walls of the right and left coelomic pouches. Calcareous ring consisting of five radial
and five inter-radial pieces with much-branched ends. A few scattered glandular organs
of doubtful function in various parts of the ectoderm.

7. THE DEVELOPMENT OF THE ADULT SYNAPTA.

Since there is no cessation of growth nor any resting period on the assumption of
the pentactula form, it is impossible to draw any hard and fast lines which will always
serve to distinguish that stage. For many larvae, which appear to have only five tenta-
cles, show on careful examination the rudiments of new ones, and other larvae which
show no tentacles externally show the perfect pentactula form, when sectioned. As soon
as the five primary tentacles have pushed out so far as to entirely obliterate the original
opening of the atrium, the secondary outgrowths of the water-ring, which have hitherto
scarcely shown any indication of growth, begin to develop and push upward. As we have
already seen, the radial nerves lie directly over them, so that they cannot grow straight
up but push out to one side or the other of the nerve. The outgrowth which lies in the
mid-ventral radius, however, develops very slightly and does not normally push out on
either side of the nerve which overlies it. The outgrowths which lie in the right and
left ventral radii take the opposite course, and broadening out laterally, grow up on both
sides of the nerves which overlie them, to form accessory tentacles. The outgrowth of
the right dorsal radius pushes out on the dorsal side of its overlying nerve and forms an
accessory tentacle in the mid-dorsal interradius, while the outgrowth of the left dorsal

SYNAPTA VIVIPARA. 09

radius shows as yet little tendency to develop either way (Fig. 85). Consequently we
now have a larval form with ten tentacles, two in each interradius. As the accessory
tentacles grow very rapidly they are soon equal in size to, and cannot be distinguished
from, the five primary tentacles. It is hard to decide positively which of the five accessory
tentacles develops first, for apparently they all begin to grow at about the same time.
In Chirodota rotlfera, Ludwig ('81) found the first two accessory tentacles in the lateral
dorsal interradii, and he does not speak of finding any trace of additional tentacles in the
other interradii. I have not found any stage similar to that in S. vimpara, and I think
the five accessory tentacles appear at practically the same time. It is an important and
interesting fact, however, that the five accessory tentacles are formed in precisely the
same manner and from the same radii as the second series of five tentacles in Cucumaria,
(Ludwig, '91). It seems to me that this fact proves satisfactorily that the radial canals
in Synaptidae are homologous with those of the other holothurians or, more accurately,
the secondary outgrowths of the hydrocoel ring in the Synaptidae are homologous with
the five outgrowths of the hydrocoel ring in the true holothurians. In both cases, the
ten-tentacled young has one primary and one accessory tentacle in each interradius.
While this change is taking place in the number and arrangement of the tentacles, a cor-
responding change is going on in the calcareous ring. As the accessory tentacles push
out into the interradii, the calcareous rod which lies at the base of the primary tentacle
comes to lie between it and the accessory tentacle. I could not see that this came about
by any actual movement of the rod itself, but was due simply to the increase of width
in each interradius. In the further growth of the calcareous ring, the interradial pieces
send up projections between the two tentacles (Fig. 46 h) and at the same- time branch
and divide so rapidly and irregularly that they soon become -plates, with straight sides
but pointed anteriorly and notched behind, made up of a very fine irregular network of
calcareous strands (Fig. 44) . The radial plates develop in the same way but send up two
projections, one on each side of the radial nerve (Fig. 46 i). which finally fuse together
above it and thus form the perforated plates of the ring (Fig. 45). About the same
time, the mesenchyme cells lying between the ectoderm and the wall of the coelom
begin to gather in groups close to the ectoderm and there give rise to anchors
and anchor-plates so chai-acteristic of the adult Synapta. The development of the cal-
careous bodies from a straight rod takes place as described by Semon ('87) for Synapta
inhaerans. While these deposits appear in the body-wall as far anteriorly as the base of
the tentacles, in the walls of the latter, lying parallel to the long axis, there appear
numerous rather long, more or less, knobbed rods (Fig. 48) similiar to those described

70 HUBERT LYMAN CLARK ON

by Seinon ('87) from the tentacles of various synaptids. These become very abundant
in the older ten-tentacled larvae.

The digestive tract meantime has increased in length, and the stomach is more
clearly marked off from the intestine and oesophagus. The nervous system has not
undergone any marked changes, but each of the accessory tentacles is supplied with a
nerve on its inner side, as in the case of the primary tentacles. In various parts of the
skin, especially anteriorly, clusters of ectoderm cells are to be found which later form
the so-called sense-papillae (" Tastpapillen " of Hamann, '83). I have been unable to
find any connection between these spots and the nerves until a very much later period,
and I cannot decide how or when this connection is made. The glandular oi-gans
previously described are very abundant at this stage, especially posteriorly. Soon after
the pentactula form is complete, the walls of the coelom and of the hydrocoel begin the
formation of muscle fibers, always on the side turned from the cavity which they enclose.
The first to appear are the longitudinal muscles of the tentacles and radii. The former
appear as fibers on the outside of the tentacle canals and they soon form quite a thick
layer. The radial muscles arise within a fold of the coelomic wall, which appears along
the inner side of the radial nerves. This fold begins anteriorly near the calcareous ring
and runs backwards with the nerve, enclosing a considerable space between its walls. In
this space the muscle fibers arise from the endodermal cells of the coelom. Later on,
the circular muscles of the body appear, arising from the outer side of the coelomic wall
also. They cross the space in which the longitudinal muscles lie, forming a layer
between the latter and the nerve. At the same time, the longitudinal and circular
muscles of the digestive tract, and the muscles of the water-ring, begin to appear, so that
by the time the ten-tentacled stage is reached, the musculature is practically that of the
adult. With the appearance of the longitudinal muscles of the tentacles, comes the
development of the valves at the openings of the tentacular canals, close to the upper
edge of the calcareous ring.

Before the accessory tentacles have begun to appear, there arises on the right hand
side of the mesentery which fastens the intestine to the wall of the left interradius a
longitudinal fold or evagination of the epithelium close to the intestine (Fig. 58). This
fold follows the course of the intestine, with the mesentery on the dorsal side, and grows
forward along the stomach and backward toward the anus. Later a similar fold appears
in the coelomic epithelium on the opposite (ventral) side of the digestive tract and the
two folds soon become connected around the intestine and stomach by numerous
ill-defined lacunae between the coelomic wall and that of the digestive organ itself.
These vessels are the first stages of the blood vascular system and into them cells from

SVXAITA VIVII'AKA. 71

tin- coelomic epithelium pass to form the blood corpuscles. Theoretically the vessels
ought to be lined with connective tissue of mesenchyme cells but these are so few in the
early stages of the larva, that the connective tissue between the laminae of the mesen-
teries or even around the digestive tract is very hard to demonstrate. The main blood
vessels appear to be purely entodermal in origin and their walls seem to be made up
solely of the coelomic wall. At any rate, the dorsal vessel does not arise in S. mvipara
in the way given by Semon ('88) for its origin in S. digitata, by a, simple split in the
mesenchyme where the two coelomic folds unite above the intestine to form the
mesentery.

Very soon after the pentactula stage is reached the first rudiment of the genital
system appears. It arises on the inner side of the right-hand lamina of the dorsal
mesentery between the stone-canal and the oesophagus. The first appearance is simply
the increased size of the cells at this point, resulting in a thickening of the wall (Fig.
32), but the cells soon multiply rapidly and form a more or less spherical mass within
the mesentery (Fig. 33). As this mass increases in size, cavities appear within it
(Fig. 34) and these increase in size and begin to unite together until they form one
central lumen for the gland (Fig. 35). Meanwhile the right lamina of the mesentery
forms an outer epithelium which soon becomes quite distinct from the central mass of
cells. Between this epithelium and the remainder of the gland a few mesenchyme cells
form an extremely scanty connective-tissue layer. In the full-grown ten-tentaculed
larva, the genital gland is plainly seen (Fig. 19), lying entirely on the right-hand side
of the mesentery, near the stone-canal.

The ten-tentacled larva (Fig. 19) is a more clearly defined stage in the development
of Synnjjfa vicijtara than is the pentacula; that is, it seems to last longer, and the rela-
tive condition of development of the various organs is more constant. The changes
which occur subsequently in the assumption of the adult form must now be considered.
The most obvious of these is the increase in the number of tentacles which arises from
the addition of another tentacle to the right and left dorsal interradii (Fig. 86). The
extra tentacle of the left side arises from the left dorsal " secondary outgrowth " of the
hydrocoel ring, which has hitherto remained in its original position beneath the left
dorsal nerve but now pushes out on its lower or ventral side and forms the eleventh ten-
tacle. At the same time, the extra tentacle of the right side arises from the lower or
ventral side of the right dorsal " secondary outgrowth '' which pushes out on that side
of the right dorsal nerve and forms the twelfth tentacle. Not infrequently individuals
are found with thirteen tentacles, and in these the extra tentacle is usually in one of the
ventral interradii. In such cases it is probable that the mid-ventral " secondary out-

72 HUBERT LYMAN CLARK ON

growth" has grown up on one side or the other of the mid-ventral nerve, although in
normal specimens it does not develop at all. Occasionally the extra tentacle is in the mid-
dorsal interradius, and in such cases it is probable that the left dorsal " secondary
outgrowth," which normally develops only a single tentacle, has given rise to a second
on the dorsal side of the left dorsal nerve. Corresponding to these changes in the num-
ber of tentacles additional plates appear in the calcareous ring, but these plates do not
arise by interpolation of new rods. On the contrary, the plate of the same radius with
the new tentacle increases its length and sends upward a new projection for the support
of the tentacle (Fig. 47), and this subsequently forms the center of the new accessory
plate. In specimens of the calcareous rings cleaned with caustic soda, it seemed to me
that the calcareous plates of the right and left dorsal radii were less easily separated from
those plates on their ventral side than from those on the dorsal, so I am inclined to think
that for a time, if not throughout life, these two plates remain in closer union than any
of the others. In specimens with thirteen tentacles, there is an additional plate in the
calcareous ring corresponding to the extra tentacle. About this same time the miliary
granules (Fig. 50) begin to appear in various parts of the body-wall and in the tentacles.
Like all other calcareous concretions of Synapta, they are formed by mesencl^me cells.
They usually appear in clusters of several hundred, which continue to increase in number
afterwards until it may reach thousands. By the time 'twelve tentacles have appeared,
the genital gland begins to push over on the left hand side, but it is not until long after
the adult form is assumed in all other respects that the left branch of the gland equals the
right in size. As soon as the left branch is well started, the germinal epithelium begins
to push upward in the mesentery beside the stone-canal, and forms the genital duct, but
does not reach the outer body-wall for some time. This account of the development of
the genital duct accords with Mortensen's ('94) observations on Cucumaria glacialis,
although his account of the origin of the genital gland itself differs considerably from my
observations. Important changes are going on meanwhile in the nervous system. From-
the inner side of the circumoral ring nerves or bands of nervous tissue arise and pass
inward to the oesophagus. These will be referred to more fully in describing the nervous
system of the adult. Even in the ten-tentacled stage, before the remaining two tentacles
have made a fair start, there arises on each side of the tentacle nerve at its base a knob-
like outgrowth which becomes covered over with a peculiar mesenchyme layer, and these
form the " eyes," which also appear at the base of the eleventh and twelfth tentacles,
after they receive their nerves from the circumoral ring. With the appearance of these
eyes, the first trace of pigment appears in the mesenchyme not only about them but in
various parts of the body, especially around the calcareous ring. This pigment on its

SYNAPTA VIVIPARA. 73

first appearance is bright green, even about the eyes, so that at this stage the eyes of
the young Synapta are very conspicuous as large green spots at the base of the tentacles.
Very soon, however, a dark reddish brown pigment appears, but this is probably an older
stage of the green, and not a different pigment; for the pigment around the eyes soon
loses its green color and turns brown, and there is no reason to assume that the pigment
in other parts of the body is any different from that around the eyes. All of the pigment
arises in the connective tissue, and is apparently a product of the mesenchyrne cells. It
is especially abundant at the anterior end of the body, and above all other places in and
around the calcareous ring.

Before the number of tentacles is complete the ciliated funnels so characteristic of
Synapta begin to appear on the mesentery, near the body-wall. These funnels arise
from a large cell or group of cells in the endodermal epithelium of the mesentery (Figs.
61 and 62a). The multiplication of these cells soon forms a hemispherical outgrowth
(Fig. 62d) which increases in size and becomes more and more spherical in shape, until
it is finally attached to the mesentery by only a narrow stalk (Fig. 62e). It then begins
to flatten on one side, and the cells of its outer layers become smaller and stain more
heavily than those nearer the stalk (Fig. 62f). The flattened surface at last becomes
concave and the funnel shape begins to be assumed. At the same time, the stalk
becomes elongated and draws up within it some of the connective-tissue layer of the
mesentery, which becomes the supporting layer of the funnel. Even during the ten-
tentacled stage the digits of the tentacles begin to be formed, but they do not become
prominent until the twelve tentacles are all developed. The digits arise as evaginations
of the water-canal of the tentacle (Fig. 56) which very soon become shut off from the
main canal and in the adult have no connection with it (Fig. 55). The earliest ones to
appear are near the middle of the tentacle, and the later ones appear both proximally
and distally to them. The digits form longitudinal muscles on the outer side of the
central cavity in the same way as the tentacles themselves. They are also supplied with
nerves from the main tentacle nerve. The peculiar glandular organs of the larva are no
longer forming but seem rather to be disappearing, and the longitudinal rods of the
tentacles reach their maximum number at this time. The circumoral sinus, which was
entirely cut off from the rest of the coelom in the pentactula stage, has increased greatly
in size (Fig. 89) but is now in open communication with the body-cavity, though strands
of connective tissue traverse it, uniting the oesophagus to the water-ring and, higher up,
to the coelomic wall. With the greatly increased size of the young Synapta, comes a
considerable change in the relative position of the organs in the body-cavity. The body
has grown much posteriorly, drawing out with it that part of the intestine which lies in

74 HUBERT LYMAN CLARK ON

the right ventral infcerradius. The nerve ring is drawn upward with the growth of the
tentacles, so that it comes to lie very near the ectoderm at their hase. The increased
length of the tentacular canals has pushed the water-ring downward so that it lies some
distance below the calcareous ring (Figs. 90 and 91), but it is still in open communica-
tion with the exterior by means of the water-canal (Fig. 66). Mesenchyme cells have
formed a few calcareous rods about the latter (Fig. 49), especially near the point where
it passes into the body-wall. Just within the body-cavity from this point, openings have
appeared on it which place its interior in direct communication with the body-cavity, so
that the water-vascular system combines the primitive external opening of the pentactula
with the internal madreporitic openings of the other holothurians. The mesench_yme
cells around the calcareous ring have formed on its posterior edge a connective-tissue
ring, which later becomes so prominent in the adult as the cartilaginous ring.

In concluding this account of the embryology it may be well to summarize briefly
the derivation of the different organs from the germ layers of the gastrula.

Ectoderm. From the gastrula-ectoderm arise the ectoderm of the adult, the sensory
epithelium of the tentacles, the entire nervous system including all the sense organs,
the larval glandular organs, and a small part of the oesophagus. Possibly the
extreme posterior part of the rectum is also ectodermal.

Endoderm. From the gastrula-endoderm arise first of all the scattered mesenchyme cells
which make up the mesoderm. Soon afterwards the hydroenterocoel is divided off.
The remainder of the archenteron forms simply the lining of the digestive tract,
including most of the oesophagus. From the hydroenterocoel, the coelomic pouches
are constricted off, leaving behind the hydrocoel, from which the entire water-
vascular system, and also the cavities of the digits, arise. The longitudinal muscles
of the tentacles and digits come from the epithelium of the hydrocoel. The coelomic
pouches form the peritoneal lining of the body-cavity and the epithelial covering for
the various organs contained in it. All the muscles of the body-wall, gut, genital
glands, water-ring, and Polian vessels are also derivatives of the endoderm. The
genital organ, including the genital duct, and the ciliated funnels are likewise
derived from the wall of the coelom. The haemal system is also covered by the
epithelium of the coelom and apparently arises as e vagina tions of the same, while
the blood-corpuscles certainly come from that layer.

Mexoderm. From the mesenchyme cells, arising from the archenteron of the gastrula,
come all the connective tissue of the body, the pigment, the covering of the eyes,
all the calcareous concretions (including the calcareous ring), and the cartilaginous

SVXAPTA VIVIPARA. 75

ring. No trace of mesenchymatous musculature was found anywhere, and the part
which the meaoderm takes in the formation of the haemal system is certainly incon-
siderable.

8. THE ANATOMY OF THE ADULT.

Although the anatomy of the European Synaptas is so well known, thanks to the
investigations of Baur ('64), SiMium ('87), Hamann ('83 and '89), Cuenot ('91), and
others, there are so many points in which Synapta vivipara differs from the forms
hitherto examined, it seems desirable to add a few words concerning these and other
points. Except in the case of sense-organs, no attempt has been made to go into the
histology, but my attention has been confined to the more general features of the minute
anatomy. In the structure of the body-wall and the muscular system, there are no
important features to mention, aside from the shape of the longitudinal radial muscle
bands. Each of these bands is forked at its anterior extremity, and the two branches
are attached to the radial calcareous plate, one on each side of the radial nerve. These
branches soon unite as they pass backward, and form a single narrow band, which
extends far out into the body-cavity. But still further back, it decreases in depth and
increases correspondingly in width, and the epithelium which covers it tends to fuse at
the outer edges with the epithelium of the body-cavity, so that at numerous points in its
course the muscle has acquired secondary attachments to the body-wall. During the
greater part of its course, it is a nearly flat band, but as it approaches the extreme
posterior region of the body, it tends to become cylindrical, and where it ends near the
anus the cross-section is circular. These changes in shape will be made clear from Figs.
94-100. The structure of the genital glands has already been given in detail, and the
openings in the wall of the rectum have also been sufficiently described. The blood-
vascular or haemal system is very simple, consisting of a dorsal and ventral vessel on the
intestine and stomach with connecting lacunae in their walls. Posteriorly, both vessels
end about half way down that section of the intestine which lies in the right ventral
interradius (Fig. 92). Anteriorly, the ventral vessel ends a little in front of the stomach,
on the oesophagus. The dorsal vessel runs forward to the water-ring and forms on its
inner side a circumoesophageal ring, from which branches pass on to each tentacular
vessel. The dorsal blood-vessel also seems to open out in the mesentery to form broad
lacunae about the genital gland, such as Cuenot ('91) found in European Synaptas, but I
never found coagulated blood there as in the dorsal vessel, and I do not feel sure that

76 HUBERT LYMAN CLARK ON

such lacunae actually exist. The ventral vessel of the stomach does not lie appressed to
its wall, but entirely free from it and connected with it by several small branches. It is
also connected by a large transverse vessel with the ventral vessel of the intestine (Fig.
92), and the two sections of the latter are also connected by a similar vessel. These
transverse vessels do not appear until the animal is several centimeters long, when they
arise by outgrowths of the coelomic epithelium of stomach and intestine which, lying
close together as they do in the loops of the digestive tract, touch and fuse (Figs. 59 and
60) and with the increased growth of the intestine are finally drawn out to slender con-
necting vessels. Like the vessels of the young Synapta, these are supposed to be lined
with connective tissue, but I have been unable to detect it in their walls.

The ciliated funnels of Synapta vivijmra differ considerably in appearance from those
of S. digitata or S. inhaerans, though they do not differ essentially in structure. Only
one sort seems to be present and these are quite small but extremely numerous on all
three of the mesenteries. They measure from 40^ to 75^ in length, and from 30/x to
4(V in breadth and depth, which is only about half the size of those of S. digitata. They
are broad funnel- or cornucopia-shaped in outline and usually have a short stalk.
Their general structure will be easily understood from Figs. 63-65. The water-vascular
system consists of a circumoesophageal ring from which canals arise and pass to the ten-
tacles, into which their entrance is guarded by well-developed valves. Each tentacle
rests on the calcareous ring in such a way that the outer half of the base is on the out-
side of the calcareous plate, forming a sort of rudimentary ampulla (Fig. 90). There is
not in the adult, any more than in any of the larval stages, the slightest trace of radia.
water-canals. Dependent from the ring-canal there is always present in the left dorsal
interradius a slender Polian vessel five or six millimeters long, and in nearly all adults
additional Polian vessels, sometimes as many as six, are present in the ventral interradii.
The stone-canal leaves the water-ring on the left-hand side of the mid-dorsal interradius
and does not lie in the dorsal mesentery but clearly separate from, and to the left of it.
It soon passes into it, however, on its outward course and runs to the body-wall close
beside the genital duct (Figs. 66 and 67). It enters the body-wall on the right of the
mesentery and bends upward more or less abruptly, opening finally to the exterior close
behind the circle of tentacles (Figs. 66-70). In exceptional cases there are two open-
ings (Figs. 71-73) or rarely the reverse happens and the canal closes before the exterior
is reached. Besides this external opening, the stone-canal also opens into the body-cavity
through a well-developed madrepore (Figs. 66 and 74). Throughout its course the canal
is heavily ciliated, and especially so around these madreporitic openings, the whole
arrangement being admirably adapted for keeping the body-cavity fluid well aerated.

SVXAPTA VIVIPAHA. 77

The nervous system consists as in all Synaptidae of the central circumoral ring with the
live radial branches and the smaller branches to each of the tentacles, but there are some
additional nerves and certain of the sense-organs which have not been figured hitherto.
Each of the radial nerves is divided longitudinally into an outer and an inner band as in
other Synaptas, but, unlike them, there are no canals or vessels of any kind accompanying
the nerves. There are, therefore, in the radii of S. viv'yjara, no spaces or lacunae in
connection with either the blood, water, or nervous systems, but they are marked simply
by the longitudinal muscles and nerves (Fig. 99). Each tentacle-nerve sends off branches
to the digits (Fig. 55), so that almost the whole surface of the tentacle becomes sensory.
On the base of the tentacles and in various parts of the ectoderm all over the body, there
are numerous sense-buds or " taste-papillae," (Fig. 84), such as were first described by
Hamann ('83). The structure of these organs has been well described by him and still
more recently by Cuenot ('91). My observations support the opinion of the latter, that
under each one of these sense papillae there lies a small ganglion. From the lower side
of the circumoral ring, there arises between every two tentacles a broad band-like nerve
(Figs. 75 and 76) which runs inwards towards the mouth, innervating the ectoderm of the
oral disc as well as the muscles of the oesophagus. Hamann ('83) describes a single
nerve to the oesophagus, and Semon ('88) speaks of it in S. digitata, but so far as I can
learn no other nerves from the inner side of the ring have been described in holothurians.
At the base of each tentacle, there are easily seen a pair of reddish brown spots, the so-
called eyes (Figs. 77 and 78). Similar spots are mentioned in various Synaptas by
Miiller ('50), Baur ('64), and Semper ('68), but Semon ('87) and Hamann ('84) seem
to doubt their visual function. There can be little doubt, however, that in Synapta
vimpara these eyes are actually of service as light-detecting organs. In position and
general structure they resemble those described by Ludwig and Barthels ('91) for Synapta
vittata. They consist of a distinct, rather horny mesodermal layer, of a light brown
color, containing scattered nuclei, overlying the swollen end of a large nerve which arises
on each side from the base of the tentacle-nerve (Fig. 77). The ends of these nerves
are made up of large nerve-cells with large nuclei, which are somewhat swollen and
apparently vacuolated at their outer extremities. They are polygonal in outline, when
seen in cross-section (Fig. 79), and the inner ends taper off into fibers which run out into
the nerve (Fig. 80). The mesodermal covering, which also has the appearance of being
vacuolated, is clearly a continuation of the thin mesoderm layer which surrounds all the
nerves. The eyes are about 60/j. in diameter, the mesodermal covering being six or eight
mikrons thick. That this covering may be affected by light is probable, for its color is
due to the pigment it contains. The other noteworthy sense-organs are the otocysts

78 HUBERT LYMAN CLARK ON

(Figs. 81-83), already mentioned as lying external to the radial nerves at the point where
they bend backwards over the calcareous ring. They are much smaller than those figured
by Cuenot ('91) for S. inhaerans and differ from them in having only a single large
vesiculated cell enclosed within them, instead of a number of small ones. The otocysts
of S. vivipara measure only about 60-70/x in diameter, while the contained cell is almost a
quarter as large. In no case have I found more than one cell enclosed in an otocyst of
this species. Hamann ('84) suspected that they were larval organs having no function
in the adult, but Semon ('87) has already proved that idea erroneous, as Hamann ('89,
p. 308) has since admitted. If any further evidence were needed, it could be found in
the increase of size of the organs during the development of the animal (Figs. 81-83) as
well as in the very obvious connection with the radial nerves. But I am inclined to the
view that these so-called otocysts do not function as hearing organs at all, but are of use
to indicate the animal's position. Semon ('87) was unable to find any cilia in them, and
his experiments on living Synaptas brought him to the conclusion that they were deaf to
sound waves. If the enclosed cell is vesiculated, as it appears to be, it must float in the
fluid with which the otocyst is filled and so presses on that part of the wall which is
uppermost. Any change in the position of the animal would cause a corresponding
change in the position of the enclosed cell and thus give rise to a changed sensation.

The fully grown Synapta oivipara (Fig. '20) measures from ten to fifteen centi-
meters in length and from four to nine millimeters in thickness ; the size depending
largely on the state of contraction of the muscles. In color they vary from a pale
reddish brown to a very dark greenish brown more or less spotted and blotched with
white. The ground color is due to the pigment in the connective tissue of the body-wall
and varies greatly with the amount of that pigment, but the while spots and blotches are
due to the aggregation of great numbers of the miliary granules, just beneath the
ectoderm. The pigment is not affected to any extent by pure alcohol, but corrosive
sublimate and all acids destroy or greatly modify it. Just posterior to the calcareous
ring and in connection with it there is a ring of cartilage-like connective tissue (Fig.
90). This structure was described and figured by Theel ('86), who also figured the
anchors and plates from the body-wall, in his account of S. picta. The anchors (Fig. 51)
lie close under the ectoderm and parallel with it, at right angles to the main axis of the
body. Each anchor is much curved or bowed inwards, while its arms or flukes are curved
outward so that the points of the arms are always projecting. The vertex is not toothed
but has five or six almost spherical knobs on its edge. The posterior end is broadened
out into several short, very finely-toothed branches. Beneath the anchors lie the rounded,
smooth-edged, somewhat arched plates, which normally possess seven large toothed holes

SYNAPTA VIVIPARA. 79

(Fig. 52) and two large and three small smooth holes. On the side of the plate next
to the anchor and near the posterior end is an arched bow, which bears a few teeth on
its anterior edge. Increased growth of the plate often increases the number of holes
(Fig. 53), but as a rule they are very constant. The calcareous rods which were so
abundant in the tentacles of the young larvae are so few that for a long time I was led
to consider them entirely wanting. The tentacles (Fig. 57) of the adults are long and
slender with from 12 to 18 pairs of digits, but the number varies greatly with the age
and size of the animal. The glandular organs which characterize the young ten-
tentacled stage seein to be entirely wanting now ; at any rate, I have never found any
trace of them in an adult.

A number of interesting monstrosities were found, chiefly among the older embryos.
One of these is shown in Fig. 93, but some of the others were much more complicated,
consisting of three, four, and, in one case, five young, which had grown together, or
budded from each other in various ways. Among adults, besides the rather common
addition of an extra tentacle, the only peculiar specimen found was one which had only
three radial muscles and nerves and only eleven tentacles. There were three tentacles
in the mid-dorsal interradius (indicated by the mesentery), and four tentacles in each of
the lateral interradii.

9. CONCLUSIONS.

Probably no theory of echinoderm phylogeny has attracted more attention or seems
more plausible than that upon which Semon ('88) determined, as the result of his
studies on the development of the auricularia larva of Synapta digitata. Although it
still finds supporters at the present time, the investigations of Ludwig ('91) on Cucu-
maria and of Ludwig and Barthels ('91) on the anatomy of the Synaptidae have shown
the incorrectness of Semon's views, while the observations of Bury ('89 and '95) have cast
doubt on his interpretation of some of the conditions in auricularia. It is not my inten-
tion to enter here into any discussion of the phylogeny of the echinodernis but only to
suggest some of the points in the phylogeny of the holothurians, upon which the history
of Synrqita vimpara seems to throw some light, and to indicate some of the particulars in
which my studies have apparently offered support to Bury's ('95) theory of the ances-
tral form of the echinoderms.

There are three possible opinions concerning the relationship of the Synaptidae to
the other holothurians : first, Semon's ('88) view that Synapta represents a primitive
form, from which the other holothurians have been derived; second, Cuenot's ('91) view

80 HUBERT LYMAN CLARK ON

that the Synaptas represent a more primitive branch of the echinoderms than and dif-
ferent from the true holothurians ; third, Ludwig's ('89-'92) view that the Synaptidae
are degenerate, pedate holothurians. Semon bases his opinion on the high development
of the nervous system in Synapta, the absence of anything in their manner of life to
cause degeneration, and the fact that no organs appear in the development of the young
Synapta which are not present in the adult. His own observations on the nervous sys-
tem of Synaptas as well as Hamann's ('83) and Cuenot's ('91) show that there is some
tendency to diversity in the nervous system, especially as regards sense-organs, among
the Synaptidae. and it also shows a considerable degree of adaptability to changing
conditions. Moreover, I have found in S. vivipara that the sense-organs are highly
developed, and there are additional nerves to the oesophagus, indicating modifications
to suit the mode of life. It seems from these facts, that too much stress must not
be placed on the opinion that the nervous system of Synapta digitata is primitive.
As regards degeneracy and the absence of anything in the mode of life to cause it in
Synaptas, it seems that Semon has expressed an erroneous opinion of the causes of
degeneration. He says that we know of only three reasons for its occurrence, parasitic,
fixed, or subterranean life, and, since none of these are characteristic of Synaptas, they
cannot be degenerate. Had he given the matter more careful consideration he would
have seen how untenable his position is. Certainly no one will deny that the loss of the
power of flight in certain birds, as the famous New Zealand ground parrot, is degenera-
tion, yet they are neither parasitic, fixed, nor subterranean. Any change in the mode of
life, due to a change in environment, may result in degeneracy. The word has come to
have a bad significance so that we think of it as indicating that the animal is on the
down-hill road, whereas it strictly means that the animal has lost some organ or group of
organs which its ancestors possessed and so has become less complex than they. Such a
loss must necessarily, however, always be a gain to the species involved, otherwise it
could never have come about. Now, it is entirely conceivable that in certain conditions
of life on the ocean bottom, in shallow water near shore, the loss of numerous ambulacral
appendages and the concentration of the water-vascular system in the circumoral ten-
tacles would be a distinct advantage to the animal. Such has certainly been the case
in Caudina (Gerould, '96), for instance, and it is probably true of all the Molpadiidae,
although in these cases it may have been brought about by subterranean life. But
Semon has by no means proved his point that the Synaptas are not, as a rule, subterranean.
Whatever may be the condition at Naples, both on the New England coast and in Jamaica
Synapta inhaerans and its allied forms are found normally buried deep in the sand, while
the large Synaptas, like S. lappa, are found under stones, which is practically a sub-

SYNAI'TA VIVIPARA. 81

terranean mode of life. The absence of anything, therefore, in their manner of life to
cause degeneration is by no means proven and will hardly stand as a good test for con-
sidering the Synaptns primitive. The statement that there is no .structure developed in
the young Synapta which does not appear in the adult is completely i-efuted by the
careful observations of Ludwig and Barthels ('91) on the absence of radial water-canals
in the Synaptidae. Since all observers are agreed that radial canals are developed in the
embryo of S. digitata, it is clear that we have here a most important structure lost in the
adult. For these reasons, it seems to me that Semon's view is no longer tenable.
Cuenot's ('91) view is based chiefly on the important differences in the embryology of the
Synaptidae and that of other holothurians, but it seems to me that he does not take suffi-
cient account of the important evidences of degeneration in the Synaptas.

Ludwig's ('89-'92) view appears to be the one best supported by the facts, and the
anatomy and embryology of S. vimpara offer no little confirmatory evidence. If we
compare its ten-tentacled stage with the hypothetical ancestor which Ludwig describes
for the Synaptidae, the resemblance is extraordinary, almost the only important differ-
ence being that the genital gland in S. vivipara is not equally developed on each side.
In fact, the ten-tentacled stage of Syncqita mvipara represents an actual step in the
development of the Synaptidae from Ludwig's hypothetical ancestor. The Jamaican
species is beyond doubt a highly modified form, and, though in some respects more highly
organized than other Synaptas, in certain particulars, degeneration has gone farther.
Differing from other holothurians in its manner .of life and its mode of reproduction, it
has undergone various modifications to fit it for the changed conditions. Living in sea-
weed near the surface of the water, it has developed pigment in its skin to a marked
degree, and at the same time has acquired additional sense-organs in the eyes at the base
of the tentacles, and an increased innervation of the oral disc. In conformity to its
changed mode of reproduction, important changes have taken place in the structure of
the genital gland, openings have appeared in the walls of the rectum to connect the body-
cavity with the exterior, while the stone-canal has retained or has acquired secondarily
its original connection with the exterior. During the progress of these specializations,
the same causes have led to degeneration in other particulars. The changed mode of
reproduction has modified the genital duct, so that its lumen is no longer open to the
ova, and it no longer has an obvious opening to the exterior. The changed mode of
life has caused a greater concentration of the water-vascular system around the mouth
and a consequent further degeneration of the radial canals, so that they no longer appear
as such even in the embryology, but tentacles develop directly from the secondary out-
growths of the hydrocoel. And, furthermore, the mid-ventral outgrowth has degenerated

82 HUBERT LYMAN CLARK ON

a step further and normally never develops at all, but disappears altogether, which is
especially interesting as the mid-ventral radius is the first to develop its nerve and oto-
cysts, and so seems to be the leader in modifications. The manner of life has also caused
a modification of the tentacles in a way which we may consider as a degeneration from
other Synaptas. Semon ('87) describes calcareous rods in the tentacles of all the Medi-
terranean Synaptas, and these appear abundantly, as we have seen, in the young stages of
S. vivipara, but in the adult they seem to have almost entirely disappeared, the tentacles
and digits being very delicate and flexible and containing no calcareous deposits, except
some miliary granules. The changes in the larva due to the retention and development
of the ova in the body-cavity of the mother, such as closing of the blastopore and absence
of any true metamorphosis, must also be taken into consideration. For these reasons, we
must consider S. vivipara as a highly specialized Synapta, but in its water-vascular system
it has degenerated a step further than S. digltata, although it is neither "parasitic, fixed,
nor subterranean " in its manner of life.

It is very clear from the examination of the literature on the subject that the study
of any one form or class of echinodenns is entirely insufficient to fit one to determine on
a theory of the phylogeny of the group. Notable examples of this may be seen in the
speculations of Semon ('88), Biitschli ('92), and MacBride ('96). The facts added to
our knowledge of echinoderm embryology by all these writers are of real value, but their
hypotheses are for the most part of little importance. The same may be said of any
attempt to determine the entire course of echinoderm evolution by the study of palaeon-
tology alone, a notable example of which has recently appeared by no less an authority
than Haeckel ('96). The only author who has carried on original investigations on all
the classes of echinodenns and has formulated his views on the phylogeny of the group is
Bury ('80 and '95), and I cannot conclude this paper without calling attention to the
support which my observations give to him, on the questions involved in the development
of the Synaptidae. Regarding all the points on which he lays particular stress, I have
confirmed his work completely or in part. The adradial position of the water-tube, the
rudimentary left anterior enterocoel, and the growth of the left body-cavity around the
oesophagus are all very clearly marked in the development of Synapta vivipara. The
only point on which I could not entirely confirm his views was on the formation of the
mesentery of the stone-canal from the left coelom entirely, and on this point what evi-
dence I did obtain indicates the correctness of his position.

SYXA1TA VIYII'ARA. 53

10. LITERATURE.

[Articles marked * contain references ti> ,s'//i«y><« viin)innt.]
Agassiz, A.

'<>». On tlie embryology of echinoilerms. Mem. Amer. acad. arts and sci., Vol. 0, p. l-.'SO.
Baur, A.

'fi4. Hi'itruge zur iiatiirgeschichte der Synapta digitata. Nova acta acad. Leop.-Carol., Bd. :!] , 110 pp., 8 taf.
*Bronn, II. G.

'00. Die klassen und ordnungen der strahlenthiere (Actinozoa). Bronn's Klassen und orduungen des thierrrirlis-

Leipzig und Heidelberg, 434 pp., 48 taf.
Btitschli. O.

'92. Versuch der ableitung des ecliinnderms aus einer bilateralen urform. Zeitschr. f. \viss. zool., I?d. 53, suppl., p. 130-

160, taf. 9.
Bury, H.

'89. Studies in the embryology of echinoderms. Quart, journ. micros, sci., Vol. 29, p. 409-449, pis. 37-39.
Bury, H.

'95. The metamorphosis of echinodevms. Quart, journ. micros, sci., Vol. 38, p. 45-135, pis. 3-9.

* Clark, H. L.

'90. The viviparous Synapta of the West Indies. Zool. anz., Bd. 19, p. 398-400.
Cuenot, L.

'91. Etudes morphologiques sur les echinodermes. Arch, de biol., Tom. 11, p. 313-680, pis. 24-31. See also Arch. zool.

expe>., (2), Tom. 9, Notes et Rev., p. 8-16.
Danielssen, D. C., and Koren, J.

'82. Holothurioidea. Norwegian North Atlantic Expedition, 1876-'78. 6. Zoology. Christiania, 94 pp., 13 pis., 1 map.
Gerould, J. H.

'96. The anatomy and histology of Caudina arenata Gould. Proc. Bost. soc. nat. hist., Vol. 27, p. 7-74, 8 pis. Ainu

Bull. mus. comp. zool., Vol. 29, p. 121-190, 8 pis.
Goette, A.

'80. Bemerkungen zur entwickelungsgeschichte der echinodermen. Zool. anz., Jahrg. 3, p. 324-32C.
Haeckel, E.

'96. Die aniphorideen und cystoideen. Beitrage zur morphologie und phylogenie der echinodermen. Festschrift zum

siebenzigsten geburtstage von Carl Gegenbaur, Bd. 1, p. 1-179, 4 taf.
Ilamann, O.

'83. Beitrage zur histologie der echinodermen. 1. Die holothurien (Pedata) und das nervensystem der asteriden.

Zeitschr. f. wiss. zool., Bd. 39, p. 145-190, taf. 10-12.
Hamann, O.

'84. Beitrage zur histologie der echinodermen. Heft 1, Die holothurien. Jena, 100 pp., 6 taf.
Hamann, 0.

'89. Anatomie der ophiuren und criuoiden. Jena, zeitschr., Bd. 23, p. 23:3-388, taf. 12-23.
Hfrouard, E.

'89. Recherches sur les holothuries des cotes de France. Arch. zool. expfir.. (2), Tom. 7. p. 535-704, pis. 25-32.
Jourdan, E.

'83. Recherches sur 1'histologie des holothuries. Ann. mus. hist. nat. Marseille. Zool., Tom. 1, Mem. 6, 04 pp., pis. 1-6.
Kowalevsky, A.

'07. Beitrage zur entwickelungsgeschichte der holothurien. Mem. acad. sci. St. Pftersbourg (7), Tom. 11,8 pp., 1 pi.

* Lampert, K.

'85. Die seewalzen. Reisen in Archipel der Philippine]!. Von Dr. C. Semper, Bd. 4, Heft 3, 312 pp., pi. 1. Wiesbaden.
*Ludwig, H.

'81. Ueber eine lebendiggebarende synaptide und zwei andere neue holothurienarten der brazilianischen kiiste. Arch.

de biol., Tom. 2, p. 41-58, pi 3.
»Ludwig, H.

'86. Die von G. Chierchia auf der fahrt der Kgl. Ital. Corvette " Vettor Pisani " gesammelten holothurien. Zool. jahr-
biicher, Bd. 2, p. 1-36, taf. 1-2.

84 HUBERT LYMAN CLARK ON

•

* Ludwig, H.

'89-92. Echinodermen. Bronn's Klassen und ordnungen des thierreichs, Bd. 2, Abth. 3. Lieferungen 1-16 (Holo-

thurien).
Ludwig, H.

'91. Zur entwickelungsgesohichte der holothurien. SitzungsbBr. k. Preuss. akad. wiss., No. 10, p. 179-192, No. 32, p. 603-
612. Transl. in Ann. and mag. nat. hist., (6), Vol. 8, p. 413-427.

* Ludwig, II., und Barthels, P.

'91. Zur anatomie der synaptiden. Zool. anz., Jahrg. 14, p. 117-119.
MacBride, E. W.

'96. The development of Asterina gibbosa. Quart, journ. micros, sci., Vol. 38, p. 339-411, pis. 18-29.
Metschnikoff, E.

'69. Studien iiber die entwickelung der echinodermen und neniertinen. Mem. acad. sci. St. Pfitersbourg, (7), Tom. 14,

73 pp., 12 pis.
Mortensen, T.

'94. Zur anatomie und en twioklung der Cucumaria glaoialis (Ljungman). Zeitschr. f. wiss. zool., Bd. 57, p. 704-732,

taf. 31-32.
M tiller, J.

'50. Anatomische studien iiber echinodermen. Miiller's archiv, Jahrg. 1850, p. 117-155.
Miiller, J.

'52. Ueber Synapta digitata und iiber die erzeugung von schnecken in holothurien. Berlin, 1852, 3G pp., 10 taf.
Orsted, A. S.

'49. [Slaegt of Synapta gruppen.] Videns. medd. nat. for. Kjobenhavn, Aarene 1849 og 1850, p. vii. Translated in

Ludwig ('81), p. 48.
Quatrefages, A. de.

'42. Memoire sur la synapte de Duvernoy (S. duvernaea A de Q). Ann. sci. nat , Zool., (2), Tom. 17, p. 19-93, pi. 2-5.
Selenka, E.

'07. Beitrage zur anatomie und systematik der holotliurien. Zeitschr. f. wiss. zool., Bd. 17, p. 291-372, taf. 17-20
Seleuka, E.

'76. Zur entwickelung der holothurien (Holothuria tubulosa u. Cucumaria doliolum). Zeitschr. f. wiss. zool., Bd. 27,

p. 155-178.
gelenka, E.

'83. Die keiiublatter der echinodermen. Studien ueber entwickelungsgeschichte der thiere. Heft 2, p. 28-01, 6 taf.

Wiesbaden.
Semon, R.

'83. Das nervensystem der holothurien. Jena, zeitschr., Bd. 16, p. 1-23, taf. 1-2.
Semon, R.

'87. Beitrage zur naturgeschichte der synaptiden des Mittelmeeres. Mittheil. zool. station Neapel, Bd. 7, p 272-300;

p. 401-422, taf. 9, 10, u. 15.
Semon, R.

'88. Die entwickelung der Synapta digitata und die stammesgeschichte der echinodermen. Jena, zeitschr., Bd. 22, p. 175-

309, taf. 6-12.
Seinon, R.

'89. Die homologien innerhalb des echinodermenstammes. Morph. jahrb., Bd. 15, p. 253-307.
Semper, C.

'68. Reisen im Archipel der Philippine!!. 2. Wissenschaftliche resultate. 1. Holothurien, 288 pp., 40 taf. Leipzig.
Th6el, H.

'82. Report on the Holothuroidea. Pt. 1. " Challenger " Reports. Zoology, Vol. 14, Part 13, London.
Theel, H.

'86. Report on the Holothuroidea. Pt. 2. " Challenger " Reports. Zoology, Vol. 14, Part 39. London.
Thomson, W.

'62. On the development of Synapta inhaerens, O. F. Miiller (sp.). Quart, journ. micros, sci., Vol. 2, p. 131-14(5.

SYNAPTA VIVIPAKA.

85

11. EXPLANATION OF PLATES.

[All figures except 20 and 23 were drawn with the aid of a camera lucicla.]
ABBREVIATIONS USED.

A. — atrium.

AE. — anterior enterocoel.
AN. — anus.
AO. — atrial opening.
AT. — accessory tentacle.
BC. — body-cavity.
BL. — blastopore.
BV. — blood-vessel.
CF. — ciliated funnels.

CH. — circumoesophageal ring of blood-system.
CM. — circular muscles of body-wall.
CH. — calcareous ring.
CT. — connective tissue.
Car.R. — cartilaginous ring.

Cir.S. — circular sinus formed from the anterior-prolon-
gation of the left coelom.
DV. — dorsal vein of blood-system.
E. — enterocoel.
Ect. — ectoderm.
Epi. — epithelium.
Ky- — eyes.
GD. — genital duct.
GG. — genital gland.
H. — hydrocoel.
I. — intestine.
LC. — left coelom.

LGO. — larval glandular organ.

LM. — longitudinal muscles.

M. — mouth.

Mes. — mesenchyme.

MD. — madrepore.

MY. — mesentery.

NR. — circumoral nerve-ring.

0. — otocysts.

OE. — oesophagus.

OEN. —nerve-band to mouth and oesophagus.

FT. — primary tentacle.

PV. — Polian vesicle.

R. — rectum.

RC. — right coelom.

RN. — radial nerve.

SC. — stone-canal.

SO. — secondary outgrowth of the hydrocoel.

SP. — sense-papilla.

T. — tentacle.

TC. — canal of tentacle.

TN. — tentacle-nerve.

TV. — blood-vessel on inner side of tentacular canal.

V. — valves.

WP. — water-pore.

WR. — water-ring.

86 HUBERT LYJMAN CLAKK ON

PLATE 11.

Fig. 1. Mature ovum. 22f>x.

Fig. 2. Two-cell stage of segmenting egg. 225x.

Fig. 3. Four-cell stage. 225x.

Fig. 4. Eight-cell stage. 225x.

Fig. 5. Sixteen-cell stage. 225x.

Fig. 6. Thirty-two-cell stage, seen from the side. 225x.

Fig. 7. Thirty-two-cell stage, seen from one of the poles. 225x.

Fig. 8. Blastula, seen from the side. 225x.

Fig. 9. Gastrula, seen from the side. 225x.

Fig. 10. Older gastrula, seen from left-hand side, to show formation of the water-pore.

Fig. 11. Still older stage, seen from left-hand side, to show the drawing away of the archenteron from the water-
pore. 225x.

Fig. 12. Slightly older stage, seen from left side, to show the formation of the mouth. 225x.

Fig. 13. Older stage, seen from in front (ventrally), to show the formation of the enterocoel. 225x.

Fig. 14. Older stage, seen from in front (ventrally), to show the formation of the coelomic vesicles. 22ox.

Fig. 15. Older stage, seen from in front (ventrally), to show the five primary outgrowths of the hydrocoel. 22.">x.

Fig. 16. Older stage, seen from left side, to show the hydrocoel, body-cavities, and atrium. 225x.

Fig. 17. Very young pentactula, from right side, to show the nerves and sense-organs. 22-5x.

Fig. 18. Older pentactula, seen from dorsal surface, to show the radial nerves, calcareous ring, and rudiment of accessory
tentacle. 130x.

Fig. 19. Ten-tentacled young, seen from right side, to show genital gland and arrangement of organs. Twelfth tentacle
just developing. 22x.

Fig. 20. Adult Synapta, viaiparu, seen from ventral surface. Nat. size.

PLATE 12.

Fig. 21. Vertical section of gastrula, to show the thickened ectoderm at apical pole. 225x.

Fig. 22. Transverse section of Fig. 13, at the line A. B., to show the thickened ventral ectoderm. 22ox.

Fig. 23. Schematic outline of hydrocoel, to show the position of water-canal.

Fig. 24. Transverse sections of hydrocoels of three embryos, to show formation of anterior, enterocoel. a, youngest,
stage; li, somewhat older; c, oldest stage. 225x. In e, only a very small part of the hydrocoel is shown.

Fig. 25. Transverse section of larva, somewhat older than Fig. 10, to show closure of hydrocoel ring, without the forma-
tion of a Polian vesicle. 225x. The section is very oblique, and takes in a part of the floor of the atrium (A), and only a
portion of the hydrocoel.

Fig. 26. Transverse section of larva like Fig. 16, to show anterior prolongation of the left coeloin. 22">x.

Fig. 27. Transverse section of same larva, somewhat higher up, to show the left prolongations of the left, coelom. 225x.

Fij,'. 28. Longitudinal section of larva like Fig. 16, very badly preserved, to show the anterior prolongation of the left
coeloin, ant. I. c. 225x.

Fig. 29. Posterior end of an adult, to show the rupture of the body-wall, caused by birth of the young. 35x.

Fig. 30. Transverse section of the posterior end of an adult, to show one of the openings from the rectum into the body-
cavity. 65x.

Fig. 31. One of these openings more highly magnified. 500x.

Fig. 32. Part of a transverse section of a young pentactula, to show the origin of the genital gland. :~>00x.

Fig. 33. Similar section of an older larva, to show the increased development of the genital gland. 500x.

Fig. 34. Similar section of a still older larva, to show first appearance of lumen and covering epithelium of the genital
gland. (, beginnings of lumen; epi, outer epithelium of gland, formed secondarily from right side of mesentery. 500x.

Fig. 35. Similar section of a young twelve-tentacled larva, to show the genital gland well developed on the right-hand
side of mesentery and confined to that side. 500x.

Fig. 36. Part of a similar section of an adult, to show the formation of the lumen of the genital duct from the lumina of
the^glands. 200x.

SYNAPTA VIVIPARA. 87

Fig. 37. Longitudinal section of a part of genital duct, to show its .structure and position in the mesentery. I, lumen of
duct; g. e., germinal epithelium; in. e., epithelium of mesentery. 950x.

Fig. 38. Transverse section of dorsal body-wall of an adult, to show the termination of the genital duct, sp, sper -
matozoa. 225x.

Fig. 39. Longitudinal section of a small part of genital gland, to show the position of ovum, pressing against the epi-
thelium of the gland, epi, epithelium of gland ; on, ovum. GOOx.

Fig. 40. Transverse section of genital gland, to show its structure and the position of the ova (ou). 500x.

PLATE 13.

Fig. 41. Transverse section of the ectoderm of a pentactula, to show the invagination which forms the larval glandular
organ. 500x.

Fig. 42. Transverse section of the same organ, fully grown, from the body-wall of a ten-tentacled larva. 500x.

Fig. 43. Longitudinal section of one of these organs, to show the peripheral position of the nuclei, and the lumen at the
center. 500x.

Fig. 44. Interradial plate from the calcareous ring of an adult. 35x.

Fig. 45. Radial plate from the same ring. 35x.

Fig. 46. The development of these calcareous plates, a, youngest stage ; b-g, successively older stages ; n, older stage
of an interradial plate ; i, older stage of a radial plate. 225x.

Fig. 47. Part of the calcareous ring of an old ten-tentacled larva, to show the growth of the radial plate (r.p.) of the
calcareous ring, in support of the developing eleventh tentacle, ip, interradial plates. 65x.

Fig. 48. Calcareous rods from the tentacles of a ten-tentacled larva. 225x.

Fig. 49. Calcareous rods from around madrepore. 225x.

Fig. 50. Miliary granules from the skin of an adult. 225x.

Fig. 51. Anchor from the skin of an adult. 120x.

Fig. 52. Normal anchor-plate from an adult. 120x.

Fig. 53. Abnormal anchor-plate from an adult. 120x.

Fig. 54. Longitudinal section of the tentacle of an old pentactula, to show the thickened ectoderm at the tip. 225x.

Fig. 55. Cross-section of a young tentacle, to show the entire separation of the canals of the digits from the central
canal of the tentacle. 225x.

Fig. 50. Longitudinal section of part of a young tentacle, to show the origin of the digits, as outgrowths of the tentacle-
canal. 225x.

Fig. 57. Tentacle of an adult. 22x.

Fig. 58. Transverse section of a mesentery and intestine of a pentactula, to show the origin of the haemal system from
the right lamina of the mesentery. 500x.

Fig. 59. Loop of intestine in a small adult, to show formation of the transverse vessels of the haemal system. :!ox.

Fig. 60. The same from an older specimen. 65x.

Fig. 61. Cross-section of the dorsal mesentery of a larva one mm. long, to show the beginning of the ciliated funnels.
575x.

Fig. 02. The development of the ciliated funnels, ft, the first large cells seen from above looking down on the surface
of mesentery ; b, increased number of cells ; c, slightly older stage seen from the side ; ((, older stage, surface view ; e, older
stage, seen partly from the side ; /, older stage seen from the side. 575x.

Fig. 63. Longitudinal section of a funnel, to show its structure. 950x.

Fig. 04. Ciliated funnel of an adult, seen from in front. 575x.

Fig. 05. Ciliated funnel of an adult, seen from above. 225x.

Fig. GO. Surface view of stone-canal of an adult, to show its course from water-ring to the exterior. 35x.

PLATE 14.

Fig. 67. Cross-section of stone-canal and genital duct of an adult before they enter the body-wall. 225x.

Fig. 08. Cross-section of stone-canal and genital duct of same adult in the body-wall. 225x.

Fig. 09. Cross-section of same stone-canal as it approaches the body-wall. 225x.

Fig. 70. Cross-section of the same, where it opens to the exterior. 225x.

88 HUBERT LYMAN CLARK ON SYNAPTA VIVIPARA.

Fig. 71. Cross-section of the terminus of a stone-canal with two openings, to show the uppermost opening. 500x.

Fig. 72. Cross-section from the same series, two sections lower clown. 500x.

Fig. 73. Cross-section of the same series, two sections further down, to show the second opening. 500x.

Fig. 74. Cross-section of a stone-canal of an adult, taken through the madrepore to show the openings. 225x.

Fig. 75. Longitudinal section of nerve-ring of an adult (transverse section of the animal), to show the nerve-band to the
mouth and oesophagus in cross-section. 200x.

Fig. 76. Transverse section of nerve-ring (longitudinal section of animal), to show the same nerve in sagittal section.
200x.

Fig. 77. Longitudinal section of nerve-ring and transverse section of tentacle-nerve, to show the position of the eyes.
120 x.

Fig. 78. Cross-section of nerve-ring and one of the eyes, to show their relative position. 225x.

Fig. 79. Cross-section of the tip of one of the eyes, to show the vacuolated structure of the mesenchymatous covering as
well as of the polygonal nerve-cells. 950x.

Fig. 80. Sagittal section of one of the eyes, to show the shape and structure of the nerve-cells. 950x.

Fig. 81. Cross-section of an otocystand its single vesiculated cell from a pentactula. 500x.

Fig. 82. Cross-section of the otocysts and a part of the radial nerve from a twelve-tentacled young five mm. long.
500x.

Fig. 83. Cross-section of an otocyst of an adult, showing the single vesiculated cell within. 500x.

Fig. 84. Cross-section of a sense-papilla in the ectoderm of an adult, to show its structure and nerve connections. go,,

ganglion ; n, nerve running to radial nerve. 500x.

PLATE 15.

Fig. 85. Transverse section of an old pentactula, to show the formation of the five accessory tentacles and the oto-
cysts ; the secondary outgrowths of the left dorsal and mid-ventral radii have not developed at all as yet. The section is
slightly oblique, more posterior on the left than on the right. 225x.

Fig. 86. Transverse section of an old ten-tentacled stage, to show position of the eleventh and twelfth tentacles. The
section is somewhat oblique, more posterior on the right than on the left. 120x.

Fig. 87. Longitudinal section of the anterior end of a pentactula ; on the left through a radius ; on the right a little to
one side of a radius. /, indicates the place of fusion between the oesophagus and roof of the atrium. 225x.

Fig. 88. Another section of the same series; on the right, through an interradius, just touching on the side of the
Polian vesicle ; on the left a little to one side of an interradius. /, indicates the fusion of the oesophagus with the roof of
the atrium. 225x.

Fig. 89. Longitudinal section of a ten-tentacled young, to show position of the various organs. G5x.

Fig. 90. Longitudinal section of an adult through one of the tentacles, to show the position of the various organs. 35x.

Fig. 91. Longitudinal section of an adult through a radius, to show position of the organs. 35x.

Fig. 92. Adult laid open in right dorsal radius, to show position of the organs and especially the blood-vessels on
digestive tract. Nat. size.

Fig. 93. Larval monstrosity, a double embryo. 225x.

Fig. 94. Cross-section of longitudinal muscle of adult at the point of juncture with the calcareous ring. 65x.

Fig. 95. Cross-section of same muscle, lower down. 65x.

Fig. 96. Cross-section of same muscle, lower down. 65x.

Fig. 97. Cross-section of same muscle, lower down. 65x.

Fig. 98. Cross-section of longitudinal muscle of a smaller aiult, somewhat anterior to the middle of the body. 200x.

Fig. 99. Cross-section of a longitudinal muscle in the middle of the body. 80x.

Fig. 100. Cross-section of a longitudinal muscle at extreme posterior end of the body. 200x.

Printed, January, 1S9S.

f/J

o

a

a

V. -*

,Si.

f . la «•.-»•

• £•*.&¥•

J '?&•'- ' /£##-*•

' -fe^iS - tW*

. U' X ^S* . \fa**ff

'•>*, *• - :••:•

•s,.:

««•

:?

.-v>* ,

..'•.'

'

it -.

1 '"'

..

f\ . -.- v:.-:..^:'' ./.>
; V . .^ "•:••'.

^ V---V. -"^

'

o

o

o

PQ

m

£

o
S

s

'« 1,5-5

* Jh,

1 It*

CO

K

THE ANATOMY, HABITS, AND EMBRYOLOGY OF

YOLDIA LIMATULA

A DISSERTATION PRESENTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN THE
JOHNS HOPKINS UNIVERSITY, JUNE, 1898

BY

GILMAN A. DREW

[FROM THE MEMOIRS FROM THE BIOLOGICAL LABORATORY OF THE JOHNS HOPKINS UNIVERSITY, 1899]

BALTIMORE

THE JOHNS HOPKINS PRESS
1899

PRINTED BY

BALTIMORE, MD., U.S. A.

THE ANATOMY, HABITS, AND EMBRYOLOGY OF YOLDIA

LIMATULA, SAY.

Yoldia limatula is widely distributed along the coast of Maine, and
large specimens are abundant. Their principal habitat, however, is
in the shallow coves and inlets, where the tidal currents sweep by
without entering. Here, soft mud has accumulated, and is constantly
being added to by sediment, and by the decay of the plants and
animals that are swept in from the surrounding sea and land. Speci-
mens are most abundant in water from one to five fathoms deep, and
probably never occur above low tide mark.

During the two seasons of my stay at Casco Bay, Maine, Mr. C.
B. Wilson generously allowed me the free use of his equipment,
which was of great value to me. Previous to this, specimens from
Woods Holl, Mass., were furnished me through the kindness of Professor
W. K. Brooks, Mr. Richard Eathbun and Dr. James L. Kellogg. Most
of these specimens were collected by Mr. Vinal Edwards.

My own work, like the work of Drs. Mitsukuri8 and Kellogg5 upon
this animal, has been carried on under the direction of Professor Brooks,
who has for many years manifested a great interest in this rather
peculiar lamellibranch. It is a pleasure to acknowledge my indebted-
ness to him. I wish publicly to acknowledge my indebtedness to my
wife, who has materially aided me in securing, tending and preserving-
specimens.

ANATOMY AND HABITS.

The distinctive characters for the genus Yoldia as given by Verrill
and Bush17 are : " shell nearly smooth, compressed, lanceolate, gaping,
more or less prolonged and tapering posteriorly, with a poorly defined
wide rostrum, generally without carinations. The external ligament
is marginal, feebly developed, continuous under the beaks and not
much differentiated from the general epidermis. The chondrophore is
large, concave, and projects within the margin. The pallial sinus is
1

2 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

large and deep. The siphon-tubes and posterior pallial tentacle are
long. The palpal tentacles are long and tapered ; in life they may
extend nearly to the end of the expanded siphon."

SHELL.
FIGURES 1 AND 2.

To the above characters of the shell may be added a few others,
part of which, no doubt, are common to all the species of the genus,
while others are specific.

When the animal is taken from the mud in which it lives, the
anterior portion of the shell is jet black. Posteriorly it gradually
assumes olive green. The black is no doubt a stain, as it bleaches
out in animals kept in aquaria, and the shell assumes a rather uni-
form olivaceous tint, somewhat streaked along some of the more
prominent lines of growth with yellow or brown.

On each valve, two more or less pronounced radial stripes extend
from the beak to the ventral margin, one anterior, the other posterior,
Fig. 1. The margin of the mantle opposite the extremity of each of
these stripes is specially modified and sensitive. Internally, near the
dorsal margin of each valve, are two rows of toxodont teeth, one
extending anteriorly and the other posteriorly from the cartilage pit,
Fig. 2. Near the cartilage pit these teeth are very small and closely
placed. They attain their greatest length about midway along each
row, and become short and rather widely separated at the ends
furthest from the cartilage pit.

Internal markings are rather obscure. Besides the scars of the
foot and adductor muscles, and the pallial line with its deep and
broad sinus, there is on each valve a rather distinct, curved line, ex-
tending from the ventral margin of the scar of the anterior adductor
muscle, nearly to the cartilage pit. Dissection shows that this line
marks the limit of the genital mass and digestive gland.

Tryon* refers to this mark on the right valve, and thinks it is
caused by the loop of the intestine that extends forward, on the right
side, very near the shell. That this view is wrong is shown by the
presence of similar markings on both valves.

* Structural and Systematic Conchology.

G. A. DREW ON YOLDIA LIMATULA, SAY. 3

MANTLE.
FIGURES 1 AND 3.

The lobes of the mantle are free along their ventral borders, and are
ciliated in patches on their inner surfaces. They are modified to
form the siphons, the marginal tentacles, an unpaired, very much
elongated, tentacle situated near the base of the siphons, a flattened
expansion opposite the extremity of each radiating stripe on the pos-
terior end of the shell, and a rounded projection opposite the extremity
of each radiating stripe on the anterior end of the shell. These will
be treated in turn. There are also two pairs of patches that are
apparently glandular. One pair is situated ventral to the anterior
adductor muscle, and the other ventral to the base of the siphons.

Siphons. — The oldest specimens reared from eggs did not begin
to develop siphons. The youngest specimen collected with a dredge
(about .5 mm. long) had already formed the exhalent siphon.

This siphon, Fig. 11, seems to have been formed by the union of
the margins of the lobes of the mantle, followed by their growth into a
tube, and the withdrawal of the tube between the lobes of the mantle.
In withdrawing the tube, its dorsal surface, corresponding to the dorsal
surface of the united lobes of the mantle, is drawn in, forming a
complete septum. A ridge on each lobe of the mantle indicates where
the point of union of the margins of the mantle has been drawn
along its inner surface.

The ventral surface of the base of this siphon arches dorsally,
Fig. 12, the ridges on the lobes of the mantle near its base thicken, and
finally fuse. Thus a second tube is formed lying ventral to the exhalent
siphon, Fig. 13. The wall separating the two siphons remains arched
upward for some time, but subsequent growth straightens it. Even
in the adult the line of fusion along the ventral side of the inhalent
siphon remains distinct, Fig. 14, and offers little resistance to splitting.
Although the inhalent siphon is formed between the lobes of the mantle,
at the base of the exhalent siphon, it may morphologically be considered
marginal in formation, as it is formed by thickenings of, and growth
from, ridges that seem to have been carried back from the margin.
The adult siphons are united along their whole length, Figs. 1 and 3,
and may be extended beyond the shell to a distance considerably
exceeding the length of the shell. Normally the inhalent siphon is
shorter, broader, and has thicker walls than the exhalent siphon, but

4 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

they are frequently rendered of equal length by injury. Both siphons,
Fig. 14, have, between their outer and inner layers of epithelium, large
bundles of longitudinal muscle fibres, separated by sheets of radial
muscles. A few circular fibres lie near the epithelium, but they are not
numerous. The siphons are extended by forcing blood into the spaces
hollowed out in the connective tissue. If the siphon of a young specimen
is examined with a moderately high power of the microscope, small,
conical papillae will be seen projecting from its surface, Fig. 15. Each
papilla bears at its tip a long, rigid filament. Older specimens show
similar papilla^ but instead of bearing a single elongated filament,
several shorter filaments are borne on the tip of each. Similar papilla?
are found in various places on the mantle and its modifications. They
probably correspond to the " Pinselzellen " of Fleming,3 later described
by Dorst'2 and Rawitz.13

Siphonal tentacle and marginal tentacles. — The youngest specimen
collected with a dredge shows a little rounded knob, Fig. 11, st., lying
on one side, between the mantle and the siphon. This knob, the
rudiment of the siphonal tentacle, is an outgrowth from the line of
union of the mantle with the base of the siphon, which line, as has
been shown, seems morphologically to be a portion of the margin of
the mantle.

During the development of the siphons the tentacle is carried
ventrally and, in the adult, comes to lie nearly opposite the ventral
border of the inhalent siphon. This unpaired tentacle was first
described by Brooks1 and has frequently been referred to since.
Pelseneer10 found that it occurred on either the right or left side of
Yoldia isonota, a'n observation that also holds true for Y. limatula.
From its position and innervation Pelseneer10 was led to conclude that
it might be compared to the osphradium of gastropods, a conclusion
that he later modified.11 Another organ corresponding in innervation
and position to the osphradium of other lamellibranchs is present
and will be described later.

When extended, Figs. 1, 3 and 10, st., the siphonal tentacle is a long
and slender filament, gradually tapering to its free extremity, and
generally lying loosely coiled on the bottom of the aquarium, or on
the surface of the mud in which the animal lives. Throughout its
length it is set with small, conical papilhe, Figs. 6 and 7, such as have
already been described for the siphons. Each papilla, like the papillse
on the adult siphon, bears a cluster of filaments at its tip.

G. A. DREW ON YOLDIA LIMATULA, SAY 5

When retracted the tentacle presents a series of transverse wrinkles,
Fig. 6. Beneath the layer of epithelium, Fig. 8, ep., run strands of
longitudinal muscle fibres, 1m., imbedded in connective tissue. On the
side of the tentacle nearest the lobe of the mantle to which it is attached,
inside the layer of muscles, is a large nerve, tn., which can be easily
traced to the tip of the tentacle. This nerve is a branch of the posterior
pallial nerve. It is not given off directly opposite the base of the ten-
tacle, but some distance above it, and the two nerves continue along, side
by side, until the tentacle is reached. The appearance of the two
nerves lying side by side is quite like that of a ganglion, and has,
no doubt, been mistaken for one. On the side of the tentacle
opposite the nerve, also within the layer of muscles, is a more or less
definite blood-space, bs.

Extension of the tentacle seems to be accomplished by forcing-
blood into this space. In favorable cases blood corpuscles can be
seen moving along it when the tentacle is being extended. If
considerable force is exerted in extending the tentacle, swellings,
Fig. 7, may occur in it. Such swellings are filled with blood and
disappear when the pressure is removed. When the muscles of the
tentacle contract, the blood is forced back and the blood-space may
be completely obliterated.

The tentacles that fringe the postero-ventral margins of the
mantle, have, beneath their epithelium, both longitudinal and trans-
verse strands of muscle fibres, and generally several blood-spaces.
While branches of the pallial nerves have not been traced into these
tentacles, such branches can be traced to their bases, and there
can be no doubt that the tentacles are supplied by fibres from these
nerves.

Each tentacle has at least one papilla and frequently there are
several papilla? upon its surface. In development the tentacle is pre-
ceded by a single sense papilla which is carried out by the growth of
the mantle near its base into a conspicuous projection. As this pro-
jection grows, other sense papillae make their appearance on its sides,
and the papilla at its tip may or may not retain its position. In a
few cases the tentacles divide or branch. This is most common with
the tentacles fringing the posterior expansions of the mantle, soon to
be described.

Considering the origin, structure, and innervation of the siphonal
tentacle and the marginal tentacles, there can be but little doubt that

6 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

the siphonal tentacle is a greatly enlarged and specialized marginal
tentacle.*

Both structure and experiment agree in assigning to these tentacles
a tactile function. The siphonal tentacle is not, as might be supposed,
the most sensitive to touch of any of the organs. It seems to be
rather more sensitive than the marginal tentacles, about as sensitive
as the expansions of the mantle opposite the extremities of the stripes on
the posterior end of the shell, and rather less sensitive than the projections
opposite the extremities of the stripes on the anterior end of the shell, the
foot, and the siphons. It may be moved about with the point of a pencil
without causing much disturbance, but if it be pressed slightly, moved
quickly, or jarred, it is quickly withdrawn, and the withdrawal may be
accompanied by the partial or complete retraction of siphons, the closing
of the shell, and, in some cases, by the activity of the foot. It seems
especially sensitive to sudden movements or jars, and its special func-
tion may lie in this direction, in enabling the creature to detect the
approach of enemies, of which flounders seem to be among the most
dangerous.

If a vessel, containing several specimens partially buried in the
mud, is not disturbed for some hours and is then jarred, all will
generally disappear with astonishing rapidity. All of the experiments
that were tried to determine the function of the tentacle resulted in
failure, inasmuch as specimens in which the tentacle had been removed
seemed to be as sensitive to jars as uninjured specimens. It may
be remarked, however, that the jars of a small vessel can at best
bear only a slight resemblance to jars affecting the surface of the
mud on the bottom of the ocean, such as would be caused, for instance,
by the swimming of a flounder.

Posterior expansions of the mantle. — These expansions, Figs. 1, 3 and
10, pe., lie opposite the extremities of the stripes on the posterior end of
the shell. They consist of flat, somewhat triangular extensions of the
edges of the mantle. The anterior border of each is fringed with
tentacles, resembling the ordinary marginal tentacles, but more
frequently branched, and usually with a rather larger proportion of
sense-papilla?. The inner surfaces of the expansions are very densely
ciliated. Near the base of each is a fold or ridge, along the side of
which such dirt as gains access to the mantle chamber is swept.

* See note at the end of the paper.

G. A. DKEW ON YOLDIA LIMATULA, SAY. 7

This dirt is swept back over the expansions and so out of the cavity.
The rich nervous supply comes as branches of the posterior pallial
nerves.

Anterior projections of the mantle. — The antero-ventral margins of
the mantle may be protruded a short distance beyond the margin of
the shell, Fig. 1, and seem to aid in cleaning the foot when it is
being withdrawn from the mud. These margins are destitute of
tentacles, but they are well supplied with sense-papillae.

Arising from each margin, opposite the extremity of the stripe on
the anterior end of the shell, is an oblong projection, Fig. 1, ae. This
projection carries a row of sense-papillae, and is ciliated on either side,
Figs. 16 and 17. Branches of the anterior pallial nerve undoubtedly
enter it, and enlarged pallial muscles act as its retractor.

These organs are very sensitive to mechanical stimulation, and
probably serve as tactile organs, but their function is not known.
They are placed so near the edges of the retracted foot that it seems
possible that they may serve some function in connection with it,
but no observations were made that support this surmise or that give
a clue to any other function.

FOOT.
FIGURES 3, 5 AND 9, f.

The foot, Fig. 3, f., is very large and powerful. Its sides form
two muscular flaps, that lie side by side, and are capable of being
extended laterally, Fig. 9, in which position they form an almost flat or
arched surface. It is this part of the foot that is called the sole.

Notwithstanding its being a rigid muscular organ, known to be
adapted for burrowing5 aDd 16 and executing movements with remark-
able rapidity,1 and 17 this part of the foot is still referred to as a " creeping
sole.'1*

The free margins of the flaps are thin, and are fringed with
flattened papillas that are very sensitive to touch. There are four
pairs of foot-muscles, one posterior and three anterior. The posterior
foot-muscles, Fig. 3, pfm., are inserted on the shell, just in front of
the posterior adductor muscle and very near the bases of the teeth.
They are very large and send all of their fibres forward, along the

*See note on page 9 of the paper.

8 JOHNS HOPKINS UNIVERSITY MOEPHOLOGICAL MONOGRAPHS.

sides of the foot, to be distributed to its anterior and ventral portions.
Their function is to retract the foot.

The anterior foot-muscles, afm., are inserted on the shell just
behind the anterior adductor muscle and very near the bases of
the teeth. Of these muscles there are three pairs. The anterior pair
pass ventrally and posteriorly along the sides of the foot, and are
distributed to its posterior portion. The middle pair pass ventrally,
between the pair just mentioned, and are distributed to an intermediate
portion. The posterior pair pass ventrally and anteriorly, between both
of the preceding pairs and are distributed to the anterior and ventral
portion of the foot, including the muscular flaps. These three pairs
of muscles are undoubtedly used in producing the varied and energetic
movements of the foot.

Besides these special foot-muscles, which together form a large part
of the outer walls of the foot, there are transverse strands of muscle
fibres extending from one wall to the other. Large blood-spaces are
hollowed out between these strands.

Movements of the foot are always very rapid and are remarkably
diversified. The movements of burrowing consist of thrusting the
closed foot, which is wedge-shaped, deep into the mud, reflecting its
margins to form an anchor, and then withdrawing it. These move-
ments follow each other in quick succession and enable an animal to
bury itself with great rapidity. When a specimen is placed upon its
side, the closed foot is bent back under the shell, its margins are
reflected, and it is then withdrawn. If these movements are executed
on mud, the lower reflected margin of the foot buries itself, and gives
the necessary purchase for the animal to right itself. If instead, they
are executed on a smooth, hard surface, where no purchase can be
obtained, these movements are likely to be replaced by leaping
movements.

One leaping movement consists of bending the foot back under the
shell, turning the shell nearly on its dorsal margin, planting the surface
of the expanded "sole" on the bottom, and giving a quick downward
movement. If the foot does not slip, the shell is generally turned end
for end and thrown some inches. If the foot slips, as more frequently
happens in smooth-bottomed aquaria, it shoots forward with wonderful
rapidity and causes a posterior movement. Other movements are
common but will not be described. It seems very hard to imagine that
the foot could possibly be used as a creeping organ, and its present

G. A. DREW ON YOLDIA LIMATULA, SAY. 9

function, for which its structure so admirably fits it, is to burrow in
soft mud.*

BYSSAL GLAND.

FlGUEES 4 AND 5, bg.

In none of the specimens that came under my observation, did this
organ seem to be functional. The youngest specimen collected with a
dredge shows a small pouch, lined with epithelium, which opens into the
posterior end of the groove that separates the muscular flaps of the foot.
The gland receives a rather large nerve from each pedal ganglion, and
it often contains a few fine threads. These threads have never been seen
protruding from the duct of the gland.

ALIMENTARY CANAL.

FIGURES 4 AND 5.

The rather broad and long oesophagus opens into the dorsal end of
a capacious stomach, Fig. 4, sto., which extends from near the dorsal
margin of the animal, well into the foot. The stomach is divided
transversely by a well-marked constriction, and internally by a
prominent ridge, formed by elongated epithelial cells. Above this ridge
the epithelial cells are rather slender, and stain deeply. Below the
ridge the epithelial cells are broader and stain less deeply.

From the ventral end of this portion of the stomach, the intestine
bends back along its posterior side, ascends to a point just beneath the
pericardium, makes a forward turn, comes very near the surface on the
right side, follows the edge of the genital mass and digestive gland

* Specimens of Yoldia limatula, Say, Nucula proxima, Say, and Nucula delphino-
donta, Mighels, have been observed for hours together, in smooth and rough-bottomed
dishes, in dishes containing sand, and in dishes containing the soft mud in which the
animals normally live. They all burrow with rapidity and in the same manner, but
in no case was a specimen observed to creep, even for the shortest distance. The
expanded foot of N. delphiuodonta is very large in proportion to the size of the
animal ; the movements are comparatively sluggish, and the almost spherical shell is
raised or turned over with but little difficulty, but the movements of the foot are
always such as are used in burrowing. Furthermore, in none of these forms does the
foot correspond in structure with that of known creeping organs.
2

10 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

nearly to the anterior adductor muscle, then turns dorsally and
posteriorly, gradually assumes a medium position again, passes through
the ventricle, and opens into the mantle chamber at the base of the
exhalent siphon, just behind the posterior adductor muscle. The
intestinal epithelium is composed of short, deeply staining cells, which
are frequently arranged in longitudinal ridges.

DIGESTIVE GLANDS.

The racemose digestive glands open into the stomach near its dorsal
end by short broad ducts. The lobes are very numerous and form a
rather compact mass in the dorsal region of the body, which latterly
overhangs the dorsal part of the foot. The left gland is larger than the
right.

LABIAL PALPS.
FIGURES 3 AND 5, Ip.

The palps of the adult animal are very large, Fig. 3, Ip. Each
united pair is suspended from the body wall by a thin membrane,
which contains a few muscle fibres. This membrane is attached to the
body wall beneath the overhanging digestive gland. The palps are
triangular in shape and have their inner surfaces ciliated and thrown
into series of ridges and grooves that do not extend quite to their free
margins. Large blood-spaces follow along the bases of many of these
ridges. The supposition that the large expanded palps serve in
respiration seems probable.

Each outer palp is supplied with a long appendage, Fig. 3, pap.,
which originates from its posterior end, near its dorsal margin, and
can be extended far beyond the posterior margin of the shell.

In young specimens, this appendage is flat and is entirely confined
to the external palp, Fig. 5, pap. As it grows it folds longitudinally,
so as to form a groove on its inner side, and, at the time, twists so
that it may appear like a continuation of the united dorsal margins
of the outer and inner palps.

Each palp-appendage is supplied with longitudinal muscles, Fig.
18, 1m., a large nerve, pn., that is continued into it from the dorsal
margins of the palps, and a continuous blood-space, bs., lying beside,

G. A. DREW ON YOLDIA -LIMATULA, SAY. 11

and morphologically ventral to it. Under favorable conditions, blood
corpuscles may be seen moving along this space when the appendage
is being extended. Other blood-spaces occur in the connective tissue
of the appendage, but they are much smaller arid anastomose freely.

The epithelium lining the groove formed by the longitudinal folding
of the appendage is densely covered with cilia.

The animal while feeding, Fig. 10, is usually slightly tipped
ventrally from the perpendicular, and frequently has about one-third
of its posterior end above the mud. While in this position, the palp-
appendages are thrust out of the shell and one, at least, bends over
and inserts its tip in the mud. The cilia lining its longitudinal groove
immediately begin to elevate the mud, which is rich in living organisms.
The stream of particles passing along the groove is large enough to be
distinguished at a distance of some feet. In this way foraminifers,
ostracods, and even small lamellibranchs, and gastropods, together
with the smaller forms and mud, are passed along the groove, finally
between the palps, and so 011 into the mouth. Thus we find that
Mitsukuri's surmise,8 based on finding sand in the grooves of the palp-
appendages of preserved specimens, was right. The palp-appendages
are food collectors.

GILLS.
FIGURES 3, 19 AND 20.

Although the gills have been well treated by Mitsukuri,8 Kellogg5
and others, some additional facts, both on the structure and function of
these organs, seem to make it allowable to go over the already rather
familiar ground.

Each gill, Fig. 3, g., consists of a double series of rather broad, flat
plates, placed laterally edge to edge, and antero-posteriorly with their
flat surfaces opposed.

This double series of plates is suspended from the body wall by a
thin muscular membrane. Anteriorly the gills pass beneath the
overhanging digestive glands, on the under sides of which the plates
gradually become smaller and smaller, and finally fade away. The
extreme anterior ends are continued as ridges which seem to represent
continuations of the suspensory membranes rather than gill plates, and
point to the probable origin of the membranes as folds from the body
wall. Posteriorly the gills likewise become smaller, and are finally

12 JOHNS HOPKINS UNIVEESITY MOEPHOLOGICAL MONOGRAPHS.

attached to the wall that separates the inhalent from the exhalent
siphon. The suspensory membranes are continued posteriorly as
ridges, beneath the posterior adductor muscle, and finally disappear.

With the exception of a few of the plates at the extreme anterior
ends of the gills, which are sometimes much distorted and swollen,
all of the plates are alike in shape and structure. Laterally the plates
of each gill lie opposite each other. Generally in the length of each
gill one or more plates are interpolated on one side or the other, Fig.
20, and thus the plates are made to alternate for a short distance,
but they soon become opposite again. Kellogg's5 opinion that the plates
of each gill are alternate and not opposite, is based on horizontal
sections taken below the upper longitudinal muscle, in which position
the inner edges of the plates are alternate. This seems to be due to
bending, caused, perhaps, by the pressure of blood in them. Horizontal
sections through the bases of the gills, as well as whole specimens, show
the plates to be opposite as a rule.

The plates on the two sides of each gill are separated from each
other except below, where they are fused, either leaving a connecting
ridge, as frequently happens anteriorly, or a smooth intermediate
portion, as is generally the case further back. Each plate is also
joined to the suspensory membrane, Fig. 19, gs., above, and to two
longitudinal muscles, llm. and ulm., that are continuous the length
of the gill. Each plate is hollow and may be likened to a loosely
folded sheet of paper, leaving a space between, which space is traversed
by a few fibres of connective tissue. The unshaded line, Fig. 19,
represents the wall that has been cut in separating the plate from
the plate lying in front of it. The opposite wall of the plate is
continuous with the wall of the plate lying behind it. Along each side
of the ventral portion of each plate, lying just inside and closely applied
to the wall, is a chitinous rod or plate Fig. 19, cr. These rods are
quite heavy near the median line of the gill, become thinner laterally,
and finally disappear. Furthermore the chitinous rod on each side
of each plate is continuous with the chitinous rod of the opposing
side of the plate adjoining. Each rod might then be described as
hairpin-shaped, with two prongs, each of which extends along the
inner side of the wall of one side of one plate. Each plate receives one
prong from each of two hairpin-shaped rods, Fig. 20, cr.

Many of the muscle fibres of the suspensory membrane are
continued down along the inner surface of the anterior wall of each

G. A. DREW ON YOLDIA LIMATULA, SAY. 13

plate, Fig. 19, mp. These were supposed by Mitsukuri8 to be chiti-
nous, but they were later shown by Kellogg1"' to be muscles. They
start from the suspensory membrane as bundles of fibres, which
branch and spread out laterally as they descend. Many of the fibres
apparently find insertion in the wall of the plate itself, but not a few
are extended to the chitinous rods and are inserted over their surfaces.
Kellogg,5 no doubt rightly referred movements of the plates, which
are common in the gills of mutilated specimens, and in the excised
gills themselves, to these muscles. The normal function of these
muscles will be made apparent later, in connection with the special
function of the gills.

Each plate bears cilia along its ventral margin, and for a short
distance along each side, but not over its general surface.

The cilia between the plates are arranged in rows that run
parallel with the margins of the plates, there being one or two rows
on each side of each plate. Near the ends of the chitinous rods,
the cilia on the edges of the plates disappear, and those between the
plates become modified, as described by Kellogg,15 for holding the plates
together. They do not seem to be hooked, but are very numerous,
rather stout, and freely intermingle with the cluster on the adjoining
plate. The area covered by these cilia is rather limited, and above
them the plates are not ciliated. The cilia of the gills seem to drive
water between the plates and aid in lowering the gills.

Each gill has two rather large bundles of muscles running its
entire length, one above, Fig. 19, ulm., and one below, llm. The lower
bundle is somewhat crescent shaped in cross-section, the concave
surface being directed upward and forming part of the ventral wall
of a large blood-space, bs., that is also continuous the length of the
gill. These longitudinal muscles were supposed by Mitsukuri,8 whose
material was poorly preserved, to be fibrous tissue, but there seems to
be no doubt that they are muscles. Their contraction causes the gills
to shorten, a movement already described by Kellogg,5 and a move-
ment necessarily of service in their special function.

The cavities of the plates on the two sides of each gill open freely
into each other, and the cavities of all of the plates are put into
communication through the blood-space, bs., which is continued
upward between the plates as a narrow slit reaching to the upper
longitudinal muscle. Above this muscle, at the base of the suspensory
membrane, is another blood-space, also running the length of the

14 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

gill, separated from the blood-spaces between the muscular walls
of the suspensory membrane only by occasional strands of connective
tissue. This space communicates laterally with the cavity of each
gill plate. Mitsukuri8 supposed that these blood channels in Nucula
were connected with some definite blood circulation. In Yoldia there
seems to be no evidence that the blood follows a very definite course.
The blood from the gills passes into the auricles, which are situated near
the anterior extremities of the gills. It seems probable that the move-
ments of the blood in all of the blood-spaces of the gills tend toward the
auricles. The continuous longitudinal spaces are probably the means
of equalizing pressure throughout the gills, by allowing a free move-
ment of the blood to different parts as circumstances demand.

— s Diagrammatic transverse section of
--m Yoldia, showing the position and at-
'" = ? tachment of the gills, f, foot; g, gill ;
gs, gill suspensory membrane ; m, man-
tle ; s, shell.

It will be remembered that the gills are suspended by muscular
membranes (gs, in the accompanying diagram), which probably repre-
sent folds of the body wall. The gills, g, are composed of wide plates
that are sufficiently wide to span the space between the foot, f, and the
mantle, m, and, behind the foot, unitedly to span the mantle-chamber
itself. Anteriorly they gradually diminish in size and finally
disappear. Posteriorly they are attached to the wall that separates the
inhalent from the exhalent siphon. The mantle-chamber is thus
divided by a movable partition, into a ventral chamber, opening
through the inhalent siphon and a dorsal chamber, opening through
the exhalent siphon.

In young specimens the movements of the brown gills are visible
through the shell. They are gradually pressed ventrally, probably by
the blood forced into them, possibly aided by the cilia, which,

G. A. DREW ON YOLDIA LIMATULA, SAY. 15

as Kellogg5 has pointed out, are so powerful in their action as fre-
quently to cause excised gills to move about in the water. As the
gills descend, the water passes between the plates. After reaching their
greatest ventral depression the gills may remain quiet for a longer
or shorter time. The period of rest is followed by energetic contractions
of the suspensory membranes, resulting in a vigorous discharge of water
through the exhalent siphon, as the chamber dorsal to the gills
diminishes, and a corresponding influx of water through the inhalent
siphon, as the chamber ventral to the gills enlarges. During this
exchange, the siphons are made quite rigid, probably by having blood
forced into them. The movements of the siphons accompanying the
movements of the gills are very conspicuous and have been mentioned
by Brooks1.

The movements of the gills are more or less rhythmic, the time
varying with the needs of the animal. Thus the rhythm is much
faster after a disturbance that has caused all movements to be sus-
pended for some time, than when no such disturbance has occurred.

The currents of water are probably for respiratory purposes, and
to keep the mantle-chamber clear of the dirt that is constantly find-
ing its way in, more especially the faeces, which, being so largely
composed of sand and other heavy particles, would otherwise drop into,
and soon clog, this chamber.

It seems to be a justifiable suggestion that the contraction of the
gills may aid in the movements of blood as well as the movements
of water. The blood-spaces of the plates, and especially of the suspensory
membranes are considerably diminished in size during each con-
traction, and, as the openings where the spaces of the suspensory
membranes communicate with the auricles are guarded by ridges that
probably act as valves, each contraction must modify the blood-supply.

A little study will show how admirably the gills are fitted for
the function of pumping water.

In shape they exactly fit the mantle-chamber, in which they form
a movable partition. Contact is insured by the pressure of the blood
inside the plates, and by the soft dorsal projections of the plates. These
projections must act much like the leather on the plunger of a suction
pump, making good contact when there is pressure from above but not
hindering its descent.

When the gills are depressed, the plates are loosely opposed and
allow water to pass freely between them. In this position the gills

16 JOHNS HOPKINS UNIVERSITY MOEPHOLOGICAL MONOGRAPHS.

are much longer than during contraction, and lie in curves, Fig. 3,
g. When the suspensory membranes contract, the longitudinal muscles,
Fig. 19, ulm., and llm., also contract, the gills are shortened, and the
plates are more closely opposed.

Each plate is rendered rather rigid ventrally by chitinous rods
that are so disposed as to allow flexibility. Many of the muscle
fibres of the plate are attached to these rods in such a way as to
insure an even strain, and keep the plate from bending ventrally
when pressure is applied.

Experiments were tried to determine, if possible, the part taken
by the gills in the collection of food. For this purpose both mutilated
and uninjured specimens were used. No definite results were reached,
but they were not observed actively engaged in collecting food. Con-
sidering the remarkable activity of the palps as collectors of food,
such activity for the gills seems rather unnecessary, and it would
also seem that the pumping action of the gills would seriously
interfere with their normally performing such a function.

As an experiment, specimens were placed in shallow dishes and
supplied with water in which living organisms were very abundant.
Under these conditions many specimens died with all the symptoms
of starvation, and those that still survived after several weeks of this
treatment were very weak and without the usual color. These weak
specimens, when placed in their native mud, where the palp-append-
ages could be used, regained their strength and color very rapidly.

EXCRETORY ORGANS.
FIGURE 23.

The excretory organs of the adult animal are very extensive. They
consist of many loops, the walls of which are much sacculated. They lie
ventral to the pericardium, and are continued into the foot along the
posterior side of the stomach. The inner, pericardial, opening of each
excretory organ is separated from the outer, mantle-chamber, opening
by a very short space.

Fig. 23 is a drawing of a wax model of the ends of the left excretory
organ and genital duct, seen from the ventral side. The inner end, ie.,
opens into the pericardium, pc., by a rather large opening, turns
abruptly towards the middle line of the body, crosses the outer end, oe.,
and is continued anteriorly close to the pericardium. After passing

G. A. DREW ON YOLDIA LIMATULA, SAY. 17

through the many sacculated loops already mentioned, the tube again
comes near the pericardial wall, enlarges, runs alongside the anterior
turn of the inner end, narrows down again and opens into the mantle-
chamber through an antero-posteriorly elongated opening, mco., just
posterior to the point where the outer and inner ends of the tube cross.

The genital duct, gd, following the course of the cerebro-visceral
commissure, turns ventrally when almost in contact with the inner end
of the excretory organ, meets the outer end, and opens with it. The
common opening of the excretory organ and genital duct is, as has been
noticed, elongated in an antero-posterior direction, Fig. 23, mco., and
may represent a fusion of the two rather than an opening of one into
the other.

As has been seen, the condition described by Pelseneer,11 viz: the
genital duct opening into the excretory organ near its pericardial
opening, is not borne out by these observations.

In material that is not well preserved it not unfrequently happens
that the walls of the genital duct and the inner end of the excretory
organ are ruptured in sectioning, at the point where they come in
contact, thus placing their cavities in communication.

It may be well to state in passing that the genital ducts of Nucula
proxima open into, or with, the outer end of the excretory organs, in a
manner similar to that just described for Yoldia limatula.

GENITAL GLANDS.

These, when distended with their products, are very extensive.
They push in between the excretory organs, below the pericardium,
between the stomach and intestine nearly to the pedal ganglia, and
cover and dip in between the lobes of the digestive glands. The opening
of the genital duct, Fig. 23, gd, has already been described in connection
with the excretory organ.

The sexes are separate and, when the genital products are abundant,
they may easily be distinguished by the color of the genital mass
surrounding the digestive glands. This portion is chocolate brown in
females and yellow in males.

NERVOUS SYSTEM.
FIGURES 4 AND 5.

The cerebral ganglia, Fig. 4, eg, united in front of the oesophagus
by a very broad commissure, which differs very little in structure from

18 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

the ganglia themselves, lie just posterior to the ventral side of the
anterior adductor muscle. Each ganglion is almost circular in
transverse section and gradually tapers posteriorly into the cerebro-
visceral commissure. So gradually does this tapering take place, that it
is impossible to determine where the ganglion ends and commissure
begins. Even in structure there is only a difference in degree, there
being proportionally more nuclei in the ganglion than in the
commissure. Posteriorly each commissure gives place, in the same
gradual way, to a visceral ganglion. The visceral ganglia, vg., lie
anterior to the ventral side of the posterior adductor muscle, and almost
equal the cerebral ganglia in size. The commissure connecting these
ganglia is also very broad and, like the cerebral commissure, tends to
share the structure of ganglia.

The pedal ganglia, pg., are rounded and fully as large as the
cerebral ganglia. They lie very close together and are connected by a
very broad commissure.

The cerebro-pedal, unlike the cerebro-visceral commissures are not
surrounded by nuclei and show ordinary commissural structure.

Although a great many specimens of widely differing ages have
been examined, the double origin of each cerebro-pedal commissure
from the cerebral ganglia has not been observed in Yoldia.* Neither
is a separation into cerebral and pleural ganglia evident. It is true
that the cerebral ganglia are not quite smooth and, in some places, the
outer layer of nuclei dips into the fibrous layer, but it seems possible
that this may be accounted for by the origin of nerves in the immediate
vicinity.

Each cerebral ganglion gives rise to several nerves. A large nerve
passes forward, under the anterior adductor muscle, branches, and
supplies the anterior portion of the corresponding lobe of the mantle.
Another large nerve passes down to the dorsal margins of the palps,
follows them posteriorly and is finally continued the whole length of
the palp-appendage. Of the smaller nerves one supplies the anterior
adductor muscle and another passes to the body wall.

Each visceral ganglion gives rise anteriorly to a small osphradial
nerve, and posteriorly to a nerve that branches almost immediately.
Sometimes these branches originate as separate nerves. One of the

* There seems to be no doubt that this commissure does have a double origin in
Nucula proxima, Say, and N. delphinodonta, Mighels.

G. A. DREW ON YOLDIA LIMATULA, SAY. 19

branches sends a small nerve to the posterior adductor muscle and
another below the adductor muscle, to the postero-dorsal portion of the
mantle. The main nerve bends ventrally and is distributed to the
siphons. The other branch passes ventrally, beside the nerve just
described, sends a large nerve to the siphonal tentacle, on the side where
this tentacle is present, and is continued to supply the postero-ventral
margin of the mantle. Several branches are given to the posterior
expansion of the mantle.

Each pedal ganglion gives rise to a nerve that runs directly to the
byssal gland, and to six or more nerves that supply the muscles of the
foot. Frequently one nerve, and sometimes several nerves, may run
up the cerebro-pedal commissure for some distance before issuing to
be distributed to the muscles of the foot.

The nerve that supplies the otocyst issues from the cerebro-pedal
commissure but, judging from the direction of the fibres in the com-
missure, probably has its origin in the cerebral ganglion.

OTOGYSTS.
FIGURES 4, 21 AND 22.

The otocysts are rather large and, as Kellogg5 has pointed out,
each contains a single large otolith, which plainly shows concentric
structure. The epithelial cells of the otocysts carry rather long and
slender cilia. In adult specimens there is a well-defined canal
connected with each otocyst. Pelseneer11 describes this canal as
opening at the surface of the foot and gives a figure indicating the
position of its opening.

I have found that, by the careful dissection of specimens preserved
in formaline, the cerebro-pedal commissures, with the otocysts, otocystic
nerves, and otocystic canals, all held together by connective tissue, can
be removed. Study of these preparations, of which I have made
about thirty, shows the otocystic canals to be cylindrical and of
uniform size for about three-fourths of the length of the otocystic
nerves, Fig. 21, ot., at which point there is nearly always a swelling.
A thin strand of fibrous tissue, probably connective tissue, is continued
beyond the swelling. At the distal end of the swollen portion of each
canal there is generally, probably always, a wall that can easily be
seen in dissected preparations, and verified in sections. This wall is
the rounded end of the otocystic canal.

20 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Sometimes one or more little closed pouches are found, lying in
the strand of tissue that is continued on past the swollen portion of
the canal. Two such pouches are- shown in Fig. 22, which represents
the distal swollen portion of a canal and part of the continuing strand
of tissue. These pouches, when found, show the same structure that
is shown by the canals, and like them contain numerous, very long cilia.

It seems to me that these walls and pouches show conclusively
that in each of the several cases the whole canal was removed
and, that in this species, the canals do not reach the surface of the
foot.*

In both Nucula proxima and Nucula delphinodonta there is a
strand of tissue passing from each otocyst to the surface of the foot.
This corresponds in position to the canal described by Pelseneer11 for
Nucula nucleus, and is probably an open canal, but I have not had
time to demonstrate this point.

In Yoldia limatula, the strand of tissue, with the series of pouches
which it sometimes incloses, seems to indicate that these canals have
undergone degeneration.

The otocystic nerves lie alongside the canals and seem to send
fibres to their walls. In fact it is difficult to demonstrate that the
nerves reach tile otocysts other than through the walls of the canals.

OSPHRADIUM.

On the inner side of each suspensory membrane of the gills, lying
just beneath the visceral ganglion of the corresponding side, is a rather
large patch of modified epithelium, which probably represents an
osphradium.

This epithelium is supplied by a rather large nerve that leaves
the corresponding visceral ganglion near its anterior end.

CIRCULATORY SYSTEM.

The heart, Fig. 4, is situated just posterior to the beaks of the
shell, and very near the dorsal margin. The cavity of each auricle is
connected with the blood-spaces of the corresponding gill, through a

* Perhaps individuals living in widely separated localities may differ in this
regard.

G. A. DREW ON YOLDIA LIMATULA, SAY. 21

slit in a muscular partition that projects dorsally into the cavity of
the auricle. The arrangement of this partition is such as to tend to
stop the backward flow of blood into the gills. The ventricle is per-
forated by the intestine. Where each auricle joins the ventricle, there
is a marked constriction and, internally, a muscular septum perforated
by a single opening. When the ventricle contracts, this septum
probably contracts and more or less completely closes the opening in
it.

Two blood-vessels leave the ventricle, one anteriorly on the left
side of the intestine, the other posteriorly, beneath the intestine.

EMBRYOLOGY.

The eggs of Yoldia limatula, are about .15 mm. in diameter, of a
chocolate brown color, and very opaque. They are not encumbered by
any kind of envelope. The polar bodies are lost soon after they are
formed. The first cleavage, which takes place about two hours after the
eggs are fertilized, results in the formation of sub-equal blastomeres,
Fig. 24.

In the next cleavage, Fig. 25, the smaller blastomere divides equally
and the larger unequally. The eight-celled stage is formed by a cleav-
age at right angles to the two preceding planes of cleavage, and the
two sets of four cells each come to lie in the position indicated by
Fig. 26, the last division being in the plane of the paper.

One of these cells is considerably larger than the other seven and,
with the beginning of the next cleavage, begins to be inclosed by the
other cells, Fig. 27. The small cell formed by the unequal division of this
large cell, Fig. 28, becomes a surface cell. What becomes of similar
cells in later divisions has not been determined, but, from sections,
it appears probable that the next small cell, cut from the large cell,
Fig. 29, crowds in over the tip of the large cell and divides into two,
Fig. 37.

About this time the outer cells, or at least part of them, acquire
cilia, Fig. 30, and the embryo begins to roll around on the bottom of
the dish. It finally rises to the surface of the water and swims freely.

Just what internal changes take place during this interval is
rather hard to determine. Surface cells in the region of the blasto-
pore became especially active, dividing and crowding in around the
large internal cell, which now divides into nearly equal cells, Fig. 39.

22 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Most of the outer cells now take on a decidedly different appear-
ance from the internal cells. There is a layer of protoplasm against
the outer wall of each cell that stains rather deeply, but the remainder
of the cell is occupied by vacuolated protoplasm. These vacuoles become
larger with age.

Beside the vacuolated cells, there is a group of cells, Figs. 38 and
39 eg., and Fig. 40, ap., which are largely surface cells and later give
rise to the cerebral ganglia and the apical plate. The apical plate,
Fig. 40, ap., is already distinguishable.

Whether there is a general migration of surface cells into the
interior is doubtful, but it is certain that their number is reduced.
The reduction is probably not more than is accounted for by the
pushing in of those around the blastopore, and the setting aside of
the group from which the cerebral ganglia and apical plate arise.

The embryo now elongates, Fig. 31, the apical plate acquires a
tuft of apical cilia, ac., and the outer, vacuolated cells, which will here-
after be referred to as test-cells, arrange themselves in five rows
that surround the embryo. Frequently the fifth, last, row of test-cells,
counting from the anterior end, is not complete on the dorsal side at this
stage, but this does not long remain so. Occasionally a sixth more or
less complete row is present. The blastopore and the apical plate occupy
nearly opposite ends of the embryo, but for some time the blastopore
may be twisted around somewhat to one side". The position of the group
of cells from which the cerebral ganglia are formed is indicated in the
external view of the embryo by a depression, x, where the cells of the
first, or anterior, row of test-cells are pushed away from the second row.
As later stages show this to be the ventral side of the embryo, it is easy
to determine that the blastopore is always twisted toward the same, that
is, toward the dorsal, side.

The test-cells are distinctly rounded and the cilia on each of the
three intermediate rows are collected into a band. The embryo now
swims in more or less definite lines, rotating the while upon its longi-
tudinal axis. The rotation is quite rapid, but is not always in the same
direction, there being frequent changes. During activity the apical
cilia are bunched together into a sort of whip, which lashes vigorously
from side to side, and always precedes the embryo in its movements.

Inside the test the cells in the region of the blastopore are still
actively undergoing division and are quite small, Fig. 41.

The cells of the apical plate, ap, and those from which the cerebral

G. A. DREW ON YOLDIA LIMATULA, SAY. 23

ganglia eg, arise, are separated externally by test-cells, but internally
they are, and continue to be, connected by cells that do not take part in
the formation of other organs.

The other internal cells have increased in number, and their
origin can no longer be determined. At a little later stage, Fig. 42,
the dorsal internal cells are found to be arranged in a layer, sg, that
represents the beginning of the formation of the shell-gland. Other cells
soon arrange themselves over the remainder of the surface inside the
test and, together with the shell-gland, form a new ectoderm. This
ectoderm probably is formed from cells that were originally surface cells.
If this view is true, the original surface cells give rise to at least three
groups : those that remain at the surface, acquire cilia, and become
differentiated as test-cells ; those that form the cerebral ganglia and
the apical plate ; and those that form the new ectoderm inside the test.

In the stage of which Fig, 42 represents a section, a space seems
always to be present in about the same position. This space probably
represents the beginning of the lumen of the mid-gut.

The embryo continues to elongate, Fig. 32, becomes cylindrical, and
the test-cells flatten and become less distinct from one another.
Between about the 32nd hour and about the 70th hour, there is no
appreciable change in the external appearance of the embryo and, as
the test-cells are quite opaque, few internal changes can be followed in
living specimens.

The test, which is now fully formed, is composed of large vacuolated
cells, Fig. 44, t, the nuclei of which are almost, if not quite, in contact
with their inner walls. Just beneath each band of cilia the protoplasm
stains very deeply. Just inside the test, between it and the new
ectoderm, there are, frequently, a few scattered nuclei lying in a very
thin film of protoplasm. Their significance is not known.

The shell-gland becomes more definite, the lumen of the mid-gut,
Fig. 43, becomes surrounded by a definite wall of rather large cells, and
an invagination, std, extends into the mass of cells from the ventral
side of the blastopore. This invagination is the beginning of what I
have called the ventral tube,* but, as it has since been found that it
is formed as an ectodermal invagination that does not at first
communicate with the mid-gut, a special name is unnecessary, and it
will hereafter be referred to as the stomodasum.

* Johns Hopkins University Circulars, No. 132, 1897.

24 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

The stomodseum elongates and soon establishes communication
with the mid-gut, Figs. 44 and 48. It lies along the ventral side of the
embryo, just inside the test, and is joined, but is not inclosed, by the
layer of ectoderm, Fig. 49, std. The shell-gland becomes slightly
invaginated, Figs. 44 and 45, sg, but it very soon arches upward
preparatory to protruding laterally to form the mantle, Fig. 49. The
mid-gut becomes prolonged posteriorly, and the mass of cells from which
the cerebral ganglia are formed arch inward and form two pouches that
open at the surface. The pouches lie side by side, Fig. 46, and generally
open to the exterior together. Sometimes, however, they open between
different test-cells, Fig. 47, and seem to be only slightly connected with
each other. In other cases only one of the pouches opens to the exterior
and this opening may be small. In some cases there may be no opening.
The walls of the pouches are rather thick but each is composed of a
single layer of cells.

The shell-gland folds out at the edges and most of the dorsal cells
assume the character of mantle cells. Only those near the edges remain
especially large and glandular. The formation of the mantle is accom-
panied by the formation of the pedal ganglia, Fig. 57, pg, as ectodermal
thickenings, and the formation of the otocysts, Fig. 56, as ectodermal
invaginations. The next stage shows the otocysts apparently completely
closed off, lying very close to the clefts that separate the mantle from
the body wall.

The formation of the mantle and shell is accompanied by a slight
lateral compression and a corresponding dorso-ventral widening of the
embryo. The test-cells become very much thinner and the vacuoles
nearly or quite disappear, but the cilia remain as described and the
movements of swimming are not altered. Several new organs make
their appearance.

Cleared whole mounts show some of the internal organs, but
throughout it is necessary to resort to reconstructions from the study of
serial sections to get the relationship of the different parts.

The body of the embryo now lies between the valves of the shell and
these are inclosed in the test. The foot, Figs. 33 and 55, f, is formed
by a rapid increase of cells between the mid-gut and the stomodaeum.
The anterior adductor muscle, Figs. 33, 55 and 58, aa, is formed at a
corresponding time (about 85 hours) and consists of a few muscle fibres
which stretch from one shell-valve to the other. It is left, by the
growth of the embryo, separated from the rest of the body, except at its

a. A. DREW ON YOLDIA LIMATULA, SAY. 25

ends, which penetrate the mantle, and are attached to the shell. The
posterior adductor muscle, Figs. 33 and 55, pa, does not appear until
some hours later. It is always surrounded by other tissue. The
digestive glands, Figs. 33, 62 and 63, are formed as evaginations from
the anterior end of the mid-gut, which now extends posteriorly
nearly to the blastopore. The visceral ganglia appear as groups of
nuclei, whose position suggests ectodermal origin. The cerebral
pouches have moved some distance dorsally and posteriorly into the
interior of the embryo, and are situated at the end of a single pouch,
Figs. 33, r, which opens to the exterior between the test-cells at the
point where the pouches were originally formed. The paired pouches
thus appear as diverticula of a single elongated pouch. As the
pouches are carried into the interior, the cells forming their walls
push in and fill up their cavities, and a commissure is formed con-
necting the two masses of cells. The relation of the unpaired pouch
to the cerebral ganglia has been shown by three succeeding transverse
sections of the same embryo. Fig. 58 shows the external opening of the
unpaired pouch, Fig. 59 shows the anterior portion of the cerebral
ganglia, eg, with the laterally compressed unpaired pouch, r, lying
beneath them, and Fig. 60 shows the commissure connecting the
ganglia.

Having described the formation of the various organs attention will
be given to the embryo's appearance.

Further than the lateral compression and the corresponding dorso-
ventral widening, there has been little change in the embryo's external
appearance, but the internal appearance is greatly altered.

Inside the test a shell has been formed, the two valves of which
gape ventrally. The anterior adductor muscle, Fig. 33, aa, is quite large
and, being isolated from the general mass of the body, is very con-
spicuous. The posterior adductor muscle, pa, as yet consists of very
few fibres and is inclosed in other tissue. It is not easily seen in whole
mounts. The apical plate, ap, is connected with the ventral end of the
group of cells that form the wall of the unpaired pouch, r. This group
of cells forms a distinct stalk which runs dorsally and posteriorly to
connect with the anterior part of the body. The mid-gut, mg, and the
stomod^eum, std, are widely separated by the foot, f. The stomodaeum
opens to the exterior through the blastopore, bl, and the posterior end of
the mid-gut almost reaches, and finally breaks through into, it.

The stomod^um is often attached to the anterior, and to part of

26 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

the ventral margins of the foot, but its posterior end, Fig. 61, std,
lies free between the lobes of the mantle, sometimes almost, if not
quite, in contact with the test, sometimes raised well above it. The
anterior end of the mid-gut is enlarged to form the stomach, which
communicates with the pouches of the digestive glands. Most of the
tissue dorsal to the intestine has been used in forming the mantle,
which is now separated from it by a large space. This space extends
to and around the anterior adductor muscle anteriorly, and nearly to
the margin of the shell posteriorly. The cerebral, pedal, and visceral
ganglia, with their commissures, have been formed. The otocysts,
Fig. 61, ot, have apparently been entirely closed off, and each contains
an otolith, which stains deeply with hjematoxylin, and, a little later,
plainly shows concentric structure. As the otocysts have never been
open to the exterior, development having taken place inside a closed
test, the otoliths cannot be foreign particles.

The only places where the test is attached to the body of the
embryo, lying inside it, are around the blastopore, along the sides of
the apical plate, and around the opening of the unpaired pouch, Fig.
33, r. The embryo still swims rapidly, but has periods of rest.

This condition of the embryo is reached in from 90 to 120 hours.
It now stops swimming, settles to the bottom, the apical cilia shrivel,
and the test-cells break apart and frequently go to pieces at once.
The process of casting the test generally begins with the posterior
row of cells and works forward. It generally takes but. a very few
minutes. Casting includes, besides the test, the stalk which extends
from it to the cerebral ganglia, the apical plate and its connection,
and the stomodaeum from the blastopore to the position of the definitive
mouth. These parts, together with a few adhering test-cells that are
no longer stretched flat, but have rounded up, may not sever their
connection for a half hour or more, Fig. 34, but they finally break off
just below the cerebral ganglia, and casting is complete.

Until about the time that the test is cast, the valves of the shell
gape ventrally, and the tissues of the body extend completely across
from valve to valve. When the shell is finally closed, by the contraction
of the adductor muscles, there is no longer sufficient space for the
organs in their original position. The principal change of importance
due to this crowding takes place in the pouches of the digestive
glands, which are forced into the space dorsal to the alimentary
canal, and, in all of the specimens examined, seem to give rise to a

G. A. DEEW ON YOLDIA LIMATULA, SAY. 27

mass of large cells, Figs. 35, 64 and 65, z, that are not arranged in
any definite manner. It was at first thought that this appearance
was due to poor preservation, an explanation that very naturally
suggested itself, but later the large cells were found in the living as
well as in the preserved specimens. Reformation of the digestive glands
is accompanied by the disappearance of these cells.

The anterior adductor muscle soon becomes inclosed by the
forward growth of the body. For a few hours a space may remain
between it and the cerebral ganglia, but this soon disappears.

The foot, which at first is capable of only feeble movements,
grows quite rapidly, and is soon very active. Locomotion is chiefly
effected by the means of long and powerful cilia arranged along the
sides of the foot, Figs. 35 and 36, which are very active whenever
the foot is thrust out of the shell. These enable the animal to glide
over the surface on which it lies, but they do not enable it to rise. For
some hours after the test is cast, the foot shows no indication of the
modification which is later generally known as the sole, Fig. 52. The
first indication of this modification is a groove which appears along
the middle line of the ventral surface of the foot, Fig. 53. This groove
deepens, apparently by the growth of the parts on each side, Fig. 54,
and gives rise to a "sole" resembling that of the adult, Figs. 3 and 9.
The movements of the foot now closely resemble those of the adult, but
locomotion on smooth surfaces is still chiefly by means of the powerful
cilia along, its sides.

The formation of the " sole " is accompanied by the growth of the
postero-ventral portion of the foot into a more and more prominent
portion resembling a heel. The use of this outgrowth is not evident.
It remains in the adult as a small projection, just behind the opening
of the byssal gland ; an organ that makes its appearance about a
day after the test is cast, but does not seem to secrete threads.

About the time that the test is cast away, a thickening makes its
appearance on the inside of each lobe of the mantle, near its posterior
border, Figs. 34 and 35 g. This thickening, the rudiment of the gill,
acquires cilia and grows anteriorly so as to project over the surface of
the mantle. A constriction appears near the middle of each gill,
Fig. 36, g, the lobes become more prominent, and it is thus divided
into almost equal parts. Beneath its epithelium, following the curve
of its anterior margin, is a chitinous rod, the two ends of which lie
near the ends of the lobes, Fig. 36.

28 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Figs. 50 and 51 represent sections of a gill, taken in the planes
indicated in Fig. 36. The chitinous rod is cut in one, but the plane
of the other lies below it. These sections show the gill to be rather
evenly ciliated. Later stages, after the lobes have elongated, show
the cilia to be confined to the side next to the mantle. The third
lobe is formed from the ventral lobe, which broadens and constricts
in the manner detailed above. Other lobes are formed in a similar
manner.

The alimentary canal elongates, bends ventrally, and the stomach
enlarges. In from two to four days after the test is cast, the digestive
glands reappear. At first they do not have very definite walls, but
they gradually become well defined, and it is noticeable that the large
scattered cells disappear at a corresponding time. The two pouches
are no longer symmetrical, the left being prolonged posteriorly fills
up the space between the dorsal end of the stomach and the intestine,
Figs. 36 and 66. The greater development of the left digestive gland,
at this early date, possibly affects the future position of the intestine,
a loop of which comes to lie on the right side of the animal, very
near the shell.

About the 25th day, a ridge, corresponding in position to the outer
labial palps, appears as an outgrowth from the body wall. The palps
did not attain any considerable development in any of the embryos
reared from eggs.

The ganglia are all distinct in whole mounts as well as in sections.
At first the cerebral ganglia are much larger than either of the
others, Fig. 33, but the pedal ganglia grow rapidly, and soon come to
equal if not to exceed them in size, Fig. 36. No division into cerebral
and pleural ganglia has been observed.

GENERAL REMARKS.
ANATOMY.

Since Mitsukuri's8 paper appeared in October, 1881, much has been
written on the anatomy of members of this group. Pelseneer11 and
others have called attention to certain characters in support of the
supposition that the group is primitive.

Some of these characters do not seem to hold good with Yoldia.
Thus on page 274 in speaking of the primitive characters of the group
Pelseneer11 includes among others the following :

G. A. DREW ON YOLDIA LIMATULA, SAY. 29

"2°. Foot has a ventral 'creeping' disc (not existing in others
except Pectunculus)"

"5°. Absence of a posterior aorta."

"7°. Genital glands opening into the pericardial extremities of the
kidneys."

" 8°. Pleural ganglia still distinct from the cerebral, and pleuro-
pedal connectives arising by distinct origins from the cerebro-pedal."

Again on the same and succeeding pages :

"2°. With the Nuculidte (Leda, Yoldia, etc.), one sees again that the
retractors of the foot (columellaires) form an almost continuous series
(without specialization) between the two adductors."

"6°. With the Nuculidas the otocysts open freely to the outside."

I can do no better than to refer the reader to what has already
been said on these subjects (see pages 8, 16, 18 19, and 21). The foot
muscles of Yoldia are surely among the most highly developed of any
to be found in the Lamellibranchia, and the only reason they occupy
so much of the dorsal margin is because of their enormous size.

EMBRYOLOGY.

The most striking peculiarities in the development of Yoldia are
connected with the formation and disappearance of the test.

I find similar tests in the embryos of two species of Nucula (N.
proxima and N. delphinodonta), in both of which cases the tests are
cast. In the former, the eggs are thrown free in the water, and the
embryos closely resemble those of Yoldia, but are considerably smaller.
In the latter species, the embryos are much larger, and are somewhat
modified by being carried in a mucous case, which is attached to the
posterior end of the shell.

Outside of the group, so far as I have been able to learn, Dondersia
is the only other Mollusk whose embryo is known to be provided with
a similar test.

Comparing the embryo of Yoldia with PruvotV2 account of the
embryo of Dondersia, we find that the fully formed tests of both are
composed of five rows of cells, all of which bear cilia. Counting from
the anterior, the cilia on the third row of cells in the test of Dondersia,
and the second, third and fourth rows of cells in the test of Yoldia,
are long and collected into bands which surround the embryos. The
body of the embryo of Dondersia protrudes posteriorly during develop-
ment. No such protrusion takes place with Yoldia. Each is provided

30 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

with an apical plate, and, in either case, the test is finally cast off.
Internal changes in the embryos of Dondersia not being known, further
comparisons are denied us.

The young embryos of Dentalium as described and figured by
Lacaze-Duthiers7 and Kowalevsky6 bear a certain resemblance to those
of Dondersia and Yoldia. This is largely due -to three or more rows
of cells, each bearing a band of cilia. At this stage these ciliated cells
form the greater portion of the surface of the embryo. As the body
elongates posteriorly, these cells are crowded forward to form the velum,
but this does not seem to be cast off.

A somewhat similar resemblance is noticeable in the case of the
embryos of Patella as figured and described by Patten.9 In either
Dentalium or Patella, if we imagine the velum to be stretched
posteriorly over the shell-gland dorsally, and the foot ventrally, so as
to inclose the body, the oesophagus will be pulled out into a long narrow
tube ventral to the foot, and the position of the blastopore will corres-
pond to its position in Yoldia. Furthermore, the position of the foot
and shell-gland will correspond, and the alimentary canal will neces-
sarily be bent in the same way.

In a similar manner, if we imagine the cells of the velum of a
typical lamellibranch larva to be drawn posteriorly over the shell,
there will be the same relation of parts. The same seems to be true
of the typical larvae of gastropods if we take into account the twisting
that accompanies development in these forms.

Thus we may perhaps trace an homology between the test of
Yoldia and the velum of other forms. In this connection it is of
interest to notice that in a few forms the velum is known to be cast
away. This has been observed by Heath* for Chiton, Sigerfoos15 for
Teredo, and Mrs. Drewf, for Cardium. Wilson18 has described a similar
casting of the cells of the velum for Polygordius.

It seems quite possible that the embryo of Yoldia represents a
rather primitive form, and that the typical molluscan larvte may have
been developed from an ancestor approaching it in form and structure.
If this is true, the probability is that organs of Yoldia and its rela-
tives, which now show high specialization, have been developed from
the more primitive rather than from the highly complex.

* Dr. Heath has kindly allowed me to notice this point, although his observations
are not yet published.

f This work has not been published. The species upon which the observations
were made seems to be C. pinnulatum, Con.

G. A. DREW ON YOLDIA LIMATULA, SAY. 31

SUMMARY.
A brief statement of the chief points is as follows :

ANATOMY AND HABITS.

1. The mantle has two pairs of sense organs, Fig. 1, ae, and pe,
one long unpaired tentacle, st, and a fringe of marginal tentacles.
All of these organs are sensitive to mechanical stimulation. The
unpaired tentacle seems to be homologous to one of the ordinary
marginal tentacles.

2. The foot, Figs. 3 and 9, f, is a highly specialized burrowing
organ.

3. The palp-appendages, Figs. 3, 10 and 18, pap, are food-collectors.

4. The gills, Figs. 3, 19 and 20, are used in pumping water, for
which function they are well fitted by their structure.

5. Each genital duct, Fig. 23, gd, joins the outer, not the inner,
end of the corresponding excretory organ, very near its opening into
the branchial chamber.

6. No division into cerebral and pleural ganglia has been observed.
Neither do the cerebro-pedal commissures seem to have a double
origin.

7. The otocysts are not connected with the surface by canals.
Canals, apparently in process of degeneration, Figs. 21 and 22, oc, are
present in adults.

8. Both an anterior and a posterior aorta are present.

EMBRYOLOGY.

1. The original surface cells seem to give rise to the test, the
cerebral ganglia, the apical plate, and a new ectoderm inside the test.

2. The cerebral ganglia, Figs. 46 and 47, are formed as invagin-
ations from the surface.

3. The otocysts contain otoliths before the test is cast. Fig. 61.

4. The test, the apical plate, the stalk connecting the test with
the anterior portion of the body, and the stomodreum from the
blastopore to the position of the definitive mouth, are all cast away.

5. It seems that typical molluscan larvas may have been
developed from embryos resembling those of Yoldia, and that the
test of Yoldia and the velum of other forms are homologous.

32 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

NOTE.

Since the preparation of this paper, Dr. Walter Stempell's paper,
entitled Beitrage zur Kenntniss der Nuculiden (Fauna Chilensis, Heft
2. Zool. Jahrbucher Supplement IV, 1898), has appeared. While it
seems that Dr. Stempell has not worked directly upon Yoldia, his
paper treats in detail on the anatomy of related forms, and can easily
be made the basis of interesting comparisons. It is a point of much
interest that closely related forms of this group, even two species of
the same genus, may differ markedly in some not unessential details.
This probably accounts for most of the differences recorded in these
papers.

In only two points do our papers seem to conflict. On page 355
Dr. Stempell, in speaking of the hypothesis* that the unpaired tentacle
is homologous with the tentacles fringing the postero-ventral margins
of the mantle, objects on the ground that Leda sulcata, while possessing
the unpaired tentacle, has no marginal tentacles along the posterior
portion of the mantle. It seems to me that the evidence cited, begin-
ning with page 6 of the present paper, will outweigh this objection.

The other point of difference lies in the probability of the modified
foot of these forms acting as a creeping organ. Dr. Stempell has not
studied the animals alive and relies on published evidence that I
cannot fully accept. (See Stempell's paper, page 374, and the present
paper, foot-note, page 9.)

* Brooks and Drew. Notes on the anatomy of Yoldia I. Johns Hopkins Univ.
Circulars, No. 126, 1896.

G. A. DREW ON YOLDIA LIMATULA, SAY. 33

LITERATURE CITED.

1. BROOKS. On an Organ of Special Sense in the Lamellibranchiate Genus
Yoldia. Pros. Amer. Asso. for the Adv. of Sci. 1874.

2. DORST. Ueber das Nervensy stem und die Sinnesepithelien der Herzmuschel
(Cardium edule, L.) nebst einigen Mittheilungen iiber den histologischen Bau ihres
Mantels und ihrer Siphonen. Morph. Jahrbuch., Bd. 12, 1887.

3. FLEMMING. Die Haaretragenden Sinneszellen in der Oberhaut der Mollusken.
Archiv fur Mikrs. Anat., Bd. 5, 1869.

4. FLEMMING. Untersuchungen iiber Sinnesepithalien der Mollusken. Archiv.
fur Mikro. Anat., Bd. 6, 1870.

5. KELLOGG. A contribution to our knowledge of the Morphology of Lamelli-
branchiate Mollusks. Bui. U. S. Fish Com. Vol. 10, 1890.

6. KOWALEVSKY. Etude sur 1'embryogenie du Dentale. An. Musee d'his-
torie naturelle de Marseille, Zool. Tome 1, 1883.

7. LACAZE-DUTHIERS. Historic de 1'organisation et du developpment du Dentale.
An. des Sci. Nat. Ser. 4. Tome 7, 1857.

8. MITSUKURI. On the Structure and Significance of some Aberrant Forms of
Lamellibranchiate Gills. Quart. Jour. Micro. Sci., Vol. 21, 1881.

9. PATTEN. The Embryology of Patella. Arb. Zool. Institut Univ. Wein., Bd.
6, 1886.

10. PELSENEER. Report on the Anatomy of the Deep-sea Mollusca collected by
H. M. S. Challenger during the years 1872-1876. Challenger Report. Vol. 27, Pt. 74,
1888.

11. PELSENEER. Contribution a 1'etude des Lamellibranchs. Archives de
Biologic. Tome 11, 1890.

12. PRUVOT. Sur le developpment d'un Solenogastre. Comptes rend. Acad. Sci.
PC is. Tome III, 1890.

i. RAWITZ. Der Mautelrand der Acephalen— Jen. Zeit. fur Naturwiss., Bd. 22,

1888.

14. RICE. Die systematische Verwertbarkeit der Kiemen bei den Lamelli-
branchiaten— Jen. Zeit. fiir Naturwiss, Bd. 31, 1897.

15. SIGERFOOS. Notes on the Organization of the Larva and the Post-larval
Development of Ship-worms. Johns Hopkins Univ. Circulars, No. 126, 1896.

16. THIELE. Beitrage zur Kenntniss der Mollusken III Ueber Hautdriisen und
ihre Derivate. Zeit. fur wis Zool., Vol. 62, 1897.

17. VERRILL AND BUSH. Revision of the Genera of the Ledidae and Nuculidse of
the Atlantic Coast of the United States. Am. Jour, of Sci., Vol. 3, 1897.

18. WILSON, E. B. The origin of the Mesoblast-Bands in Annelids. Jour, of

Morph., Vol. 4, 1890.

34

JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

KEY TO LETTERS.

aa. Anterior adductor muscle. mp.

ac. Apical cilia. oc.

ae. Anterior projection of the mar- oe.

gin of the mantle. on.

afm. Anterior foot-muscle. ot.

ap. Apical plate. pa.

bg. Byssal gland. pap.

bl. Blastopore. pc.

bs. Blood-space. pe.

eg. Cerebral ganglia.

cr. Chitinous rods. pfm.

ec. Ectoderm. pg.

ep. Epithelium. pn.

es. Exhalent siphon. r.

f. Foot.

g. Gill. rl.
gd. Genital duct. s.
gs. Gill suspensory membrane. sg.

h. Heart. st.
ie. Inner end of excretory organ. std.
int. Intestine. sto.
is. Inhalent siphon. t.
Ibs. Lower blood-space. ti-
ll. Left lobe of liver. ulm.
Una. Lower longitudinal muscle. vg-
1m. Longitudinal muscle. x-
Ip. Labial palp.

m. Mantle. J-

mco. Common opening of the excre- z-
tory organ and genital duct
into the mantel chamber,

mg. Mid-gut.

Muscles of the gill plates.

Otocystic canal.

Outer end of the excretory organ.

Otocystic nerve.

Otocyst.

Posterior adductor muscle.

Palp-appendage.

Pericardial wall.

Posterior expansion of the mar-
gin of the mantle.

Posterior foot-muscle.

Pedal ganglia.

Palp nerve.

Pouch leading to the cerebral
ganglia.

Right lobe of liver.

Shell.

Shell-gland.

Siphonal tentacle.

Stomodseum.

Stomach.

Test.

Tentacle nerve.

Upper longitudinal muscle. ,

Visceral ganglia.

Depression at the point where
the cerebral ganglia are formed.

Cut wall of gill plate.

Large cells that appear when the
liver is forced into the space
dorsal to the intestine.

G. A. DREW ON YOLDIA LIMATULA, SAY. 35

EXPLANATION OF FIGURES.

Fig. 1. Expanded adult specimen, seen from the right side.

Fig. 2. Left shell-valve, seen obliquely from the dorsal margin.

Fig. 3. Adult specimen represented as seen from the right side, with the right
shell-valve and mantle-lobe removed. The siphons are shown cut longitudinally.

Fig. 4. Same as the above with the right palp and gill removed, figured to show
the relations of internal organs. The nerves leaving the ganglia are shown only on
the left side. (Partially reconstructed.)

Fig. 5. Specimen about 1 mm. long. Represented as a partially transparent
object, seen from the right side, after the right shell-valve and mantle-lobe have been
removed. (Partially reconstructed.)

Fig. 6. Tip of a retracted siphonal tentacle.

Fig. 7. Portion of an extended siphonal tentacle, with a swelling caused by the
pressure of the blood, seen as a transparent object.

Fig. 8. Transverse section of a siphonal tentacle.

Fig. 9. Anterior portion of the shell, and the reflected margins of the foot, just
before they are brought together and withdrawn into the shell.

Fig. 10. An animal as it appears while feeding.

Fig. 11. Posterior portion of a specimen about .5 mm. long.

Fig. 12. Transverse section of a specimen about 1 mm. long, showing the thicken-
ing of the ridges leading to the base of the siphons, and the pushing in of the ventral
wall of the exhalent siphon. Taken near the base of the siphons.

Fig. 13. Posterior portion of a specimen about 2 mm. long.

Fig. 14. Transverse section of a specimen about 1 cm. long, taken through the
bases of the siphons. One of the ridges is shown joining the inhalent siphon. The
section also includes a portion of the wall attaching the exhalent siphon to the
dorsal portion of the mantle.

Fig. 15. End of the exhalent siphon of a specimen .7 mm. in length.

Fig. 16. Projection of the margin of the mantle opposite the extremity of stripe

on the anterior end of the shell.

Fig. 17. Section of the anterior projection of the margin of the mantle.

Fig. 18. Transverse section of a palp-appendage.

Fig. 19. A pair of gill plates seen from their anterior surfaces. The suspensory
membrane, the walls of the plates joining the plates next in succession, the longitudi-
nal muscles, and the bases of the chitinous rods are shown in section. These are all
cut across in separating the pair of plates from the pair lying in front. (Drawn from
a study of sections.)

Fig. 20. Portion of a gill, seen from the ventral side.

Fig. 21. Otocyst with its canal and nerve, and a portion of the cerebro-peda
commissure. A nerve frequently leaves the commissure near the otocystic nerve and
is distributed to the muscles of the foot.

Fig. 22. Portion of the extremity of an otocystic canal, showing two closed

pouches.

36 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Fig. 23. Inner and outer ends of the left excretory organ and genital duct, seen
from the ventral side.

Fig. 24. Two celled stage.
Fig. 25. Four celled stage.

Fig. 26. Eight celled stage seen from the animal pole.
Fig. 27. Eight celled stage. The cells preparing for the next cleavage.
Figs. 28 and 29. Later cleavage stages.
Fig. 30. Ciliated embryo about 14 hours old.
Fig. 31. External appearance of an embryo about 22 hours old.
Fig. 32. External appearance of an embryo about 70 hours old.
Fig. 33. Reconstruction of an embryo about 90 hours old. The outline and test-
cells were obtained from a whole mount. The cilia are indicated only at the
margins. The anus breaks through later.

Fig. 34. Reconstruction of an embryo that has just cast the test, represented
with the right shell-valve and mantle-lobe removed. The stomodfeum and the stalk
extending from the cerebral ganglia, together with adhering test-cells, are still
attached.

Fig. 35. Reconstruction of an embryo about 10 hours after casting. Represented
with the left shell-valve and mantle-lobe removed.

Fig. 36. Reconstruction of an embryo about 12 days old. Represented with the
right shell-valve and mantle-lobe removed.

Fig. 37. Section of an embryo about 10 hours old. The plane of the section lies
through the blastopore.

Figs. 38, 39 and 40. Successive sections of an embryo about 13 hours old. Some
of the surface cells are beginning to assume the character of test-cells. The group
of cells eg., figures 38 and 39, are connected with the cells ap., figure 40, by cells of
the same character.

Fig. 41. Median vertical section of an embryo about 17 hours old.
Fig. 42. Median vertical section of an embryo about 22 hours old.
Fig. 43. Median vertical section of an embryo about 30 hours old.
Fig. 44. Median vertical section of an embryo about 36 hours old.
Fig. 45. Transverse section of an embryo about 35 hours old, taken near the
third band of cilia.

Fig. 46. Transverse section of an embryo about 45 hours old, taken through
the cerebral pouches.

Fig. 47. A section similar to the preceding showing the pouches separate.
Fig. 48. Transverse section of an embryo about 45 hours old, taken where the
stomodseum joins the mid-gut.

Fig. 49. Transverse section of the same embryo as the preceding, a little further
posterior.

Fig. 50. Section of the gill rudiment in the plane indicated by the line 50,
Fig. 36.

Fig. 51. Section of the same gill rudiment as the preceding in the plane indi-
cated by the line 51, Fig. 36.

Figs. 52, 53 and 54. Transverse sections showing the formation of the " sole" of
the foot.

G. A. DREW ON YOLDIA LIMATULA, SAY. 37

Fig. 55. Median vertical section of an embryo about 90 hours old. (A portion of
the stomodreum was drawn from the next section.)

Fig. 56. Transverse section of an embryo about 75 hours old, showing the in-
vaginations that form the otocysts.

Fig. 57. Transverse section of an embryo about 70 hours old showing the
forming pedal ganglia.

Figs. 58, 59 and 60. Successive transverse sections through the anterior portion
of an embryo about 95 hours old.

Fig. 61. Transverse section of an embryo about 95 hours old, taken through the
otocysts.

Fig. 62. Transverse section through the liver pouches of an embryo about 95
hours old.

Fig. 63. Horizontal section through the liver pouches of an embryo about 95
hours old.

Fig. 64. Median vertical section of an embryo about 10 hours after casting.

Fig. 65. Transverse section of an embryo of the same age as the preceding.

Fig. 66. Horizontal section through the dorsal portion of an embryo about 12
days old.

G. A. DREW ON YOLDIA LIMATULA, SAY. 39

LIFE.

*

The writer of this dissertation was born in Jasper County, Iowa,
November 15, 1868. He received his early educational training in the
public schools of that State, and in the year 1890 was granted the
degree of Bachelor of Science at the State University of Iowa.

Three years have since been spent in teaching, and five years have
been devoted to study in this University.

In the year 1897 he was awarded a University scholarship, and he
has since held a fellowship.

YOLDIA LIMATULA, SAY.

PLATE

Fiii.li >.

YOLDIA LIMATULA, SAY.

PLATE II

s<

1-ifiJI.

.
;
.

>-

Kq.20.

Ka.14.

w,

I'if/J.i

YOLDIA LIMATULA, SAY.

PLATE III.

YOLDIA LIMAIULA. bAY.

KLAI L IV

YOLDIA LIMATULA. SAY.

PLATE V.

>• «M

\f-r&
«\ %««.,••

•

• \x

.* -^-: «.
., -i «••',

^

PHYSIOLOGY AND HISTOLOGY OF THE CUBOMEDUS/E

INCLUDING

DR. F. S. CONANT'S NOTES ON THE PHYSIOLOGY

A DISSERTATION PRESENTED TO THE BOARD OF UNIVERSITY STUDIES OF THE JOHNS HOPKINS
UNIVERSITY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

BY

E. W. BERGER

BALTIMORE

THE JOHNS HOPKINS PRESS

1900

PRINTED BY

(gafftwore (preee

THE FRIEDENWALD COMPANY

BALTIMORE, MD., U.S.A.

This Memoir is a continuation of the work upon the Cubomedusee
which was begun by the late Dr. FRANKLIN STORY CONANT, and it
contains his notes of physiological experiments, as well as new results
which have been obtained by Dr. E. W. BERGER from the study of
material which had been collected by Dr. CONANT, who had hoped to
make it the object of further study.

In order that this work may be made public as a continuation of
Dr. CONANT'S researches, his sister, GRACE WILBUR CONANT, has, with the
cooperation of other members of his family, made an adequate and
generous provision for its publication.

For this gift, which is at once a contribution to science and a
memorial of an able and promising investigator, lately student and
fellow in this institution, the Johns Hopkins University returns its
grateful acknowledgments.

DANIEL C. GILMAN, President.
W. K. BROOKS, Professor of Zoology.

CONTENTS.

INTRODUCTION. PAGE

History 1

Epitome of Anatomy 2

PHYSIOLOGICAL.
CHARYBDEA.

Light and Darkness 5

Concretions 8

Sensory Clubs 9

Velarium and Frenula 11

Pedalia, Interradial Ganglia, Tentacles 12

Stomach, Suspensoria, Proboscis, Subumbrella 13

Margin, Radial Ganglia, Nerve 15

Stimulation 17

Activity of Charybdea 17

Temperature 17

Food and Feeding 18

Occurrence of Charybdea 18

AURELIA AND POLYCLONIA (Cassiopcea) J9

SUMMARY 22

DR. CONANT'S NOTES.
CHARYBDEA.

Light and Darkness 24

Sensory Clubs 26

Nerve 29

Side, Subumbrella 30

Pedalia, Velarium, Ganglia 31

Tentacles 32

Proboscis, Stomach, Phacelli 33

Temperature 33

Food and Feeding 33

Occurrence of Charybdea 33

Activity of Charybdea 34

AURELIA AND POLYCLONIA 35

CASSIOPCEA 39

AURELIA 39

vi CONTENTS.

HISTOLOGICAL. PAGE

Method 40

Anatomy 41

Distal Complex Eye-
General 41

Cornea 42

The Lens 42

The Capsule 44

The Retina 45

(a) The Prism Cells 46

(b) The Pyramid Cells 48

(c) The Long Pigment Cells 50

(d) Subretinal Nerve Tissue 53

(e) Discussion of Literature 53

(f) Function of the Retinal Cells, Patten's Theory, and further

Literature 56

The Proximal Complex Eye 60

The Simple Eyes 61

Lithocyst and Concretion 63

The Epithelium of the Clubs 64

Network and Multipolar Ganglion Cells 67

The Nerve Tissue 67

The Supporting Lamella 68

Epithelium of Ampulla and Floating Cells 68

The Endothelium of the Peduncle ' 73

The Tentacles and Pedalia—

The Ectoderm 74

(a) Thread Cells 74

(b) Muscle Fibers 74

(c) Ganglion Cell 75

The Endoderm 75

SUMMARY 77

LITERATURE 78

REFERENCE LETTERS 80

DESCRIPTION OF FIGURES . 81

INTRODUCTION.

This paper may be regarded as a continuation of the Cubomedusan
studies pursued by Dr. F. S. Conant while in Jamaica, in 1896 and 1897,
with the Johns Hopkins Marine Laboratory. His systematic and
anatomical results have since been published as his Dissertation ("The
Cubomedusae") by this University. Conant described this paper as
Part I, hoping soon to add a second part on the physiology and the
embryology, for which he had some notes and material at hand.
Returning, however, to Jamaica with the laboratory, in 1897, he
continued his physiological experiments, and preserved material for
histological purposes. Upon the untimely death of Conant, his material
and notes were placed in my hands by Professor Brooks, to whom I
here take the opportunity of expressing my appreciation and sincere
thanks for the honor thus conferred and for the many favors received.

In this paper I shall note at some length Conant's physiological
results and append his notes. I shall also add my results on the
histology of the eyes and the sensory clubs in general, with some few
facts on the histology of the tentacles. The embryology will be
reserved for a future paper.

The forms used in the physiological experiments were Charybdea
Xaymacana, one of the two species (see Literature V, a and b)
first found and described by Conant; Aurelia aurita; Polyclonia and
Cassiopcea. The greater number of Conant's notes are on Charybdea,
and were left by him just as taken at the time of experimenting.
Many of these notes are highly interesting and in the main fit in
well with Romanes'1 and Eimer'sIV results.

Dr. Conant's work on Charybdea, in 1897, was wholly done at
Port Antonio, Jamaica. At first Conant had only varying success in
obtaining Charybdea, scouring the harbor and neighboring water at
all hours, only to obtain but few specimens. It was on the forenoon
of August 7th, while we were dredging at the head of East Harbor
with a steam launch, that many Charybdeae were brought up in the
dredge. This gave Conant a clue to their whereabouts and to the
means of obtaining them, and from that time on he was able to
l

2 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

obtain them in abundance. His first physiological experiments were
begun on August 4th and continued thereafter at intervals of several
days until his departure from Jamaica on September 6th.

Dr. Conant usually performed his experiments during the second
half of the forenoon, after the animals had stood for a few hours in
the laboratory.

The building that was rented at Port Antonio for a laboratory
had, in the basement, a photographer's dark-room, which was of great
service to Conant in his experiments.

The experiments on Aurelia, in 1897, were also performed at
Port Antonio, between August 6th and 9th. The experiments on
Cassiopoea were probably made at Port Antonio, where specimens
were occasionally obtained.

The notes on Aurelia and Polyclonia, in 1896, were taken at Port
Henderson, between May 12th and June 27th.

In his notes Conant speaks of Polyclonia and Cassiopoea. It is
at present undetermined whether he really had both forms or whether
he uses the two names for the same form. It seems likely that in
1896 he thought the form to be Polyclonia, while for some reason,
in 1897, he supposed it to be Cassiopoea. I have examined several
specimens of these medusae brought from Port Antonio and find that
they all have twelve marginal bodies and twenty-four radial canals,
according to which (V, Haeckel's System), they should be Polyclonia.
Conant, however, speaks of removing sixteen marginal bodies, which
seems to indicate that he had Cassiopoea. A careful classification of
this form of medusa found about Jamaica seems to be a desideratum.
I suppose, however, that for our purpose in this paper it will make
little difference which name is used, the two forms being so similar
in form and structure. I have, therefore, decided to retain both the
names used by Conant.

For the complete anatomy of Charybdea the reader is referred
to Dr. Conant's dissertation, "The Cubomedusre" (8b), or the Johns
Hopkins University Circulars (8a), both published by the Johns
Hopkins Press. But, for the convenience of those who may be less
familiar with Cubomedusan anatomy, the following brief summary
of the anatomy of Charybdea is given :

The Cubomedusse, as the name implies, approximate cubes, with
their tentacles (four in Charybdea) arranged at the four corners of
the lower face of the cube. These tentacles are said to lie in the

E. W. BERGER ON THE UUBOMKDUS/U. o

interradii. Half way between any two points of attachment of the
pedalia (the basal portions of the tentacles) and a little above the
margin of the bell (cube), in a niche, hang the sensory clubs, one on
each side, four in all. Each sensory club hangs in a niche of the
exumbrella and is attached by a small peduncle whose axial canal
is in connection with one of the four stomach-pockets and in the
club proper forms an ampulla-like enlargement.

Each club is said to lie in a perradius, and, like the tentacles,
belongs to the subumbrella. This is shown by the course of the
vascular lamellae, bands of cells that, stretching through the jelly
from the endoderm to the ectoderm all around the margin, form the
line of division between sub- and exumbrella.

Each club has six eyes. Two of these on the middle line of the
club facing inwards are called the proximal and distal complex eyes,
to distinguish them from the four simple eyes that are disposed
laterally, two on each side of the line of the two complex eyes. All of
these eyes look inwards into the bell cavity through a thin transparent
membrane of the subumbrella. Besides the eyes and the ampulla
already mentioned, a concretion fills the lowermost part of the club,
and a group of large cells, having a network-like structure and called
network cells by Conant, fill the uppermost part of the club between
the proximal complex eye and the attachment of the club to its
peduncle (Plate II, Fig. 13). What is evidently nerve tissue, fibers and
ganglion cells, fills the rest of the club, with two groups of large
ganglion cells disposed laterally from the network cells. A sensory
(flagellate) epithelium covers the club.

Most Cubomedusse, among them Charybdea, have a velarium
(comparable to the velum of the -Hydromedusae), a membrane of
tissue that extends inwards at right angles all around the margin.
This velarium, like a velum, has a central opening through which
the water is expelled from the bell-cavity when the animal pulsates.
In the perradii and in the angle between the velarium and the body
wall, are the frenula, which give support to the velarium much like
brackets support a shelf, except that here the brackets are above the
shelf instead of below.

In the upper part of the bell is the stomach, with the phacelli in
its interradii, and continued ventrally into the manubrium, or the
proboscis. The cavity of the stomach is continued in the perradii
through the four gastric ostia into the four stomach pockets, which

4 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

occupy the sides of the bell and extend to the margin. Immediately
below the gastric ostia, and in the bell cavity, are the suspensoria, one
in each perradius. These support the floor of the stomach much as the
frenula support the velarium, except that the suspensoria are placed
under the shelf (to continue Conant's figure) and not above it as are
the frenula.

A nerve ring, underneath the epithelium of the subumbrella,
passes from near the origin of each pedalium at the margin to the
origin of the peduncles of the sensory clubs, a little above the margin,
giving off a branch to each club. Eight ganglia are found in the
course of this nerve. The four pedal ganglia lie near the bases of
the pedalia, and are hence interradial ; the four radial ganglia lie
near the bases of the peduncles of the clubs, and are perradial. A
small nerve, radial nerve, can be traced a short distance upwards
from each radial ganglion. Underlying the epithelium of the frenula
and the suspensoria are ganglion cells and nerve fibers in larger
numbers than elsewhere (excepting the ganglia mentioned) in the
subumbrella. Otherwise, ganglion cells and nerve fibers underlie the
epithelium of the subumbrella, including the inner surface of the
velarium, as also do muscle fibers, except in the perradii and in the
region of the nerve, where the latter become interrupted.

PHYSIOLOGICAL.

CHARYBDEA.

Light and Darkness— Experiments 1-9, 10, 33, 34.— As already stated
in the Introduction, a part of Conant's experiments were performed
in a photographer's dark-room, with the animals in a deep glass jar.
In the dark a fair proportion of the animals became nearly quiescent
on the bottom, but upon lighting a lamp many started up immediately,
while others took a longer time to come to the surface and swim.
These experiments were tried a number of times and on different
occasions with very similar results. Some medusas, however, tried
immediately after being brought in, seemed not to react so well upon
being placed in the dark-room, nor would they become quiescent.
This, probably, was due to the fact that the animals had not yet
recovered from the effects of being caught and placed in new
surroundings. (Experiments 1, 2, 3.)

Other experiments (4-8, 33, 34) were tried by carrying the jar
with the animals from the weaker light of a room into the more
intense light of outdoors or into direct sunlight. The usual result
was an inhibition of pulsation and a settling to the bottom, while
the medusae immediately became active again upon returning with
them to the room. These results were so marked that no doubts can
be entertained as to their cause, though some exceptions occurred in
which animals placed in the sun continued to swim on the surface
or soon recovered pulsation. In some experiments, too, no animals
responded to the inhibitory stimulus of the brighter light or all very
soon recovered. (See, however, Temperature.)

Reducing the light by placing a coat over the jar produced
the same effect in some experiments (8, 9, 10) as did reducing the
light in other ways, while removing the coat produced the same
effect as exposure to brighter light. In these instances it appears
to be the transition from weaker to stronger light that inhibits
pulsation, rather than the actual intensity of the light ; and vice versa.
It must be noted, too, that when left for some time in any one place

6 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

the animals changed, some coming to the surface and others going
to the bottom.

These experiments show beyond doubt that Charybdea is sensitive
to light, and that it is moderate light that stimulates the animals to
activity, while darkness and strong light inhibit activity. While the
individual exceptions, as Conant himself suggests, are well explained
on the supposition of individual diversity, yet it appears that other
conditions, such as the time of day, temperature, etc., may have been
responsible for some of the exceptional experiments in which no
animals responded as expected.

While light of any intensity seems to have stimulated Romanes11
Sarsia and Tiaropsis (HydromedusaB) to activity, we note that it
is moderate light that stimulates Charybdea. This fact is evidently
correlated with the circumstance that Charybdea usually lives upon
or near the bottom.

It may further be added in regard to Romanes1 Tiaropsis
polydiademata, that when it was suddenly exposed to light it went
into a spasm preceded by a long latent period during which there
was a "summation of stimulating influence" in the ganglia. Sarsiae
would congregate toward the source of light and in general were
more active in light than in the dark, while sudden darkness often
inhibited a swimming bout. Romanes proves for Sarsia that
the marginal bodies are the seat of luminous stimulation and
that it is the light rays and not heat rays that stimulate. He
also remarks that he has obtained similar results on the covered-eyed
(Scyphomedus*) medusa, namely, that they respond to luminous
stimulation.

It may here be of interest to note a few observations made by
myself at Wood's Holl, Mass., on a beautiful Olindiad, which is
abundant in the Eelpond at the above place. I found that in a
room, in the ordinary light of evening, the animals swam actively ;
but the moment the electric light was turned on they stopped swim-
ming and settled to the bottom or attached themselves to a branch
of some weed or stem suspended in the water. This was the result
in every trial. It is found, further, to be little active during the
brighter parts of the day, when one must dip quite deep with a net
in order to obtain it. A similar observation is also made by
Murbach", who further states that this medusa may be deceived
into laying its eggs by placing it in the dark.

E. W. BERGEK ON THE CUBOMEDUS.E. 7

One cannot help but remark how analogous is the behavior of
medusas, in respect to light and darkness, to the behavior of many
of the higher animals,— and medusae are among the most lowly
organized of the animal creation.

Were one to conclude from the behavior of Charybdea in light
and darkness in the laboratory, that it remained on or near the
bottom in the daytime but became more active near or at the
surface evenings, nights and early mornings, one would probably not
be far from the truth. Dr. Conant, while towing near the bottom
with a weighted net, in water four to five feet (1.2-1.5 m.) deep not far
from shore and deeper farther out, found Charybdea in abundance
mornings and afternoons, but very few in the evening. In the
evening some few were usually taken in the surface tow. (See Intro-
duction, Occurrence and Activity.)

Again, who knows but that Charybdea is active during the day,
on the bottom where it was dredged (the light there would only be
moderate), and quiet at night. This supposition would seem to be
true, at least, for those forms of Cubomedusse that live in deep
water. We can hardly suppose that they should regularly rise to
the surface from great depths and become active. This much we do
know that bright light inhibits Charybdea's activities, while it
probably would not be active in perfect darkness.

I do not know just what interpretation to put upon Conant's
finding Charybdea at Port Henderson at the surface during the
early part of the forenoon, before the sea-breeze roughened the water
("Cubomedusffi" p. 7). This fact hardly fits in with my conclusions
above. Perhaps Charybdea's habits vary with its habitat.

Finally, while I find no experimental evidence in Conant's notes
about what parts of Charybdea are sensitive to light, yet it would
seem preposterous, from histological evidence and from Romanes'
results on Sarsia, to doubt that the eyes of the marginal bodies are
the seat of this stimulation.

Dr. Conant further experimented by cutting off certain organs
and parts from the Cubomedusan bell. These excisions consisted
chiefly in cutting out the concretions of the sensory clubs, cutting off
the whole club, eliminating a part or whole of the margin and the
velarium, cutting the bell into sectors, excising the stomach and
parts connected with it, and other parts.

8 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Concretions — Experiments 10, 11. — The four concretions were
removed from each of four animals. Two of these (Experiments 10,
and another (X), not appended, to save space) seemed to be little if
at all affected by the operation. One of the two (10) swam actively,
at first up and down more changeably than those intact, but later
mostly near the surface. The other one also swam actively and
showed nothing to indicate weakened sense-perception. The other
two (11) did not stand the operation well, as Conant remarks, and
immediately went to the bottom, where they remained, one swimming,
while eight hours later one was still in good condition.

Several attempts with stronger light by removing the coat from
the jar made no difference in the behavior of 10; it continued to
swim as heretofore. Upon a final trial, however, with removing the
coat, it went to the bottom, thus showing a possible reaction to light;
but when next seen it was keeping to the bottom.

That the concretions should function as organs of light sensation,
as the first of the above animals might seem to indicate, I believe is
out of the question.* The fact, too, that this same animal (10), together
with another (X), swam actively, immediately changing their course
upon coming to the surface, in reality behaving quite as normal
animals, hardly permits us to conclude from the behavior of the
other two (11) that the concretions function directly as organs of
equilibrium or space relations. May these concretions not function
simply as weights for keeping the sensory clubs with their eyes
properly suspended? Since these concretions lie at the lowermost part
of the clubs and in closed sacs and unsupported by cilia, it would
seem that the above suggestion as to their being weights is not
improbable. Direct observation (Experiment 20) by Conant shows,
furthermore, that the clubs always hang with a tendency for the
concretions to be lowermost, regardless of the position of the animal.

Again, while they may function as weights, as just explained,
the fact that the epithelium of the clubs is flagellated (a flagellum,
continued as a nerve fiber, to each cell — see Histology), the supposition
lies near that these flagella are the ones influenced by the concretions
as the clubs bear against one side of the sensory niche or the other.

* It was at one time supposed that the concretions in the marginal bodies of
medusae represented lenses and the surrounding nerve tissue the optic nerve, a
supposition so highly improbable that it never gained any acceptance. (Ib., p.
41, note.)

E. W. BERGER ON THE CUBOMEDUS^. 9

A somewhat similar view seems to be held by other observers and
is noted by Lang in his text-book ( " The outer epithelium of the
auditory body carries the auditory hairs"). It seems, then, that in
functioning as weights for suspending the clubs, they may also serve
at the same time for making the pressure of the club against the
niche greater than if they were absent, and thus in part serve in
equilibrium. On this supposition we should expect, furthermore, that
after the removal of the concretions the animal would be little,
if at all, affected, since the clubs themselves, without the concretions,
would still be of sufficient weight to be influenced by gravity and thus
to bear against the walls of the sensory niche. It must be noted,
however, that Conant's experiments upon equilibration in Charybdea are
negative. Also, that Charybdea has any auditory sense is negatived by
two attempts of Conant's with a violin — one attempt with the violin
near the animals, and another with it in contact with the dish. (From
an unpublished note.) Hence, some other word such as sensory or
equilibrating should perhaps be substituted for "auditory" in the
above quotation.

Removing the concretions from Aurelia gave negative results very
similar to those on Charybdea. (Experiment 42.)

Sensory Clubs — Experiments 12-19, 20, 24. — The entire sensory clubs
were removed from a number of animals. A paralysis of pulsation
followed by a rapid recovery was the usual result. In some instances,
however, there was no paralysis, while in others no recovery followed
paralysis. This is true in a general way whether one club only or all
were removed. While no permanent paralysis followed the removal
of one or two clubs, yet permanent paralysis did occur after the
removal of a third club, as, of course, also after the removal of a
fourth. It is evident, too, that as the removal of the clubs progressed
recovery seemed to be weaker after each cutting, except in one case
when pulsation seemed to be quickened after the removal of a second
club. The pulsations after recovery seemed to be not so strong and
regular, often quite feeble, and in one instance in groups. Pieces of
tissue with a club attached and pulsating regularly, ceased pulsating
after removal of the club, in one instance, however, still giving
occasional contractions.

These results are quite the same as those of Romanes1 on
Aurelia, Cyan^a, etc., and of EimerIV on Aurelia, Rhizostoma,

10 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Cotylorhyza, etc.* In these forms Romanes sometimes obtained
complete paralysis after the removal of the sensory clubs only, as also
after the removal of the whole margin, though this was not marked
in Aurelia. In Cyanaea and other forms motor centers seemed to
be more abundant than in Aurelia, so that paralysis was oftener
followed by recovery. He concludes that while the principal motor
centers reside in the lithocysts, other centers doubtless exist that
may function vicariously, but that the centers of the margin are
more definitely limited to the marginal bodies in the Scyphomedusae
than in the Hydromedusse, in which the whole margin seems to be
replete with centers. He feels positive, furthermore, that no motor
centers exist in Aurelia's margin outside of the marginal bodies
(lithocysts). Eimer's results are essentially the same as Romanes',
so that for a more detailed comparison of the two, Romanes1 works
should be consulted.

Romanes' conclusion for the Hydromedusaa is that the motor
centers are not so definitely localized in the marginal bodies, but
in the margin generally, the excision of the marginal bodies alone
producing only partial paralysis, as would also the removal of the
margin from between the marginal bodies, but not so marked.
For the Hydromedusae he concludes, then, that all the centers of
spontaneity are definitely localized in the margin, but not limited
to the marginal bodies. To this he mentions one exception, namely,
Staurophora laciniata, in which another center is found near the
margin and two others in two opposite arms of the proboscis.

I made the remark in an abstract (VI) on Conant's notes that
Romanes did not obtain recovery of pulsation after removal of all
the lithocysts in Aurelia. As noted above, he did obtain recovery, so
that Conant's results on Charybdea and also Aurelia (see Polyclonia
and Aurelia) are quite in agreement with Romanes.

The paralysis following the removal of the clubs in Charybdea is
evidently, primarily, the result of a loss of a part of its nervous
mechanism (motor centers), and, secondarily, of nervous shock, and
points to the existence of a definite nervous mechanism in the
clubs. The histological evidence is here, as usual, corroborative of the
physiological.

Another interesting phenomenon observed after the removal of

* Eimer's results I get from Romanes and Hesse"1.

E. W. BERGEE ON THE CUBOMEDUS^. 11

one or all of the clubs was the strange behavior of the proboscis.
This would reach from side to side, expanding and contracting its
lips as if trying to grasp something. This behavior is very similar to
that of the proboscis of Tiaropsis indicans when Romanes stimulated
any part of its subumbrella, or of Limnocodium sorbii, a little fresh-
water medusa, when he stimulated its margin -or the region of the
radial canals. (Ib., p. 242.)

I may add that I observed a very similar movement of the
proboscis of the Olindiad, before mentioned. When I pulled off pieces
of its gonads by means of quick jerks, with a small forceps, it would
continually reach toward the injured part of its subumbrella. This
medusa is generally quite active with its proboscis and can occasion-
ally be seen to reach with it.

Romanes states in one place that the proboscis is not affected by the
excision of the margin. This is evidently not the case in Charybdea,
in which excision of the sensory clubs (which really belong to the
margin — see "Cubomedusas") decidedly stimulated the proboscis to
active movements. This, furthermore, points to the marginal bodies
as being organs of considerable importance in giving information in
the life of Charybdea. In Romanes' Sarsia and other medusas, however,
the proboscis did respond to the stimulation of the tentacles and the
marginal bodies, as also would the bell respond to a stimulation of
the proboscis (manubrium), thus showing a reflex nervous connection
between these regions of the bell, similar to that described for
Charybdea.

Velarium and Frenula— Experiments 18, 29, 30, 41c.— "The power
of originating contractions" to use Conant's own words, "evidently
resides in the velarium or in ganglion cells of the frenula, just
as it does in the proboscis and the floor of the stomach." Isolated
pieces of the velarium contracted by themselves as did the whole
velarium when all other tissue had been removed. An isolated velarium
with the margin and the pedalia attached gave irregular contractions.
When the pedalia with the interradial ganglia were removed it still
contracted; and when all the other tissue was cut off contractions
continued.

Cutting the velarium caused the pedalia to be strongly contracted
inwards so that the tentacles were brought inside the bell. Cutting
away the velarium did not interfere with the pulsations of the bell,
but progress was much retarded.

12 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGEAPHS.

Cutting the frenula caused the pedalia to contract but seemed
not to affect the ability to swim. Comparing the velarium of the
CubomedusaB with the velum of the Hydromedusas, I recall no
observations similar to the ones here noted, though it seems that the
two may have quite similar functions. It seems somewhat probable
that the velum, and also the velarium, may function in obtaining
food, — and this besides their function in swimming. Their probable
function in swimming, as is well known, is evidently to narrow the
mouth of the bell and thus to cause the water to be forced out in a
smaller but more rapid stream, giving the animal a steady and more
prolonged movement through the water at every contraction of the
bell. In regard to taking food, I observed that a small crustacean, in
the process of being swallowed by an Olindiad, seemed to be held by
the velum being firmly contracted about it while the proboscis was
working itself over the crustacean. It would seem, furthermore, that
my supposition is supported for Charybdea by the fact that the
pedalia and tentacles were contracted so as to be brought inside the
bell when the velarium was cut. The stimulus of cutting the velarium
may be comparable to a stimulus from some object touching it, and
thus cause the pedalia and tentacles to come reflexly to aid in
capturing or holding the object, a fish, crustacean, or such, to be
captured.

Pedalia, Interradial Ganglia, Tentacles — Experiments 15, 23, 27-31,
41b. — When the pedalia were removed, the power of the animal to guide
itself was completely gone. When one pedalium was cut the others
contracted, while stroking the outer edge of the pedalia, touching the
sensory clubs, or sharply pricking the subumbrella, often produced the
same result. (See also Nerve.) The upper part of the subumbrella
seemed not so sensitive and more seldom produced the reflex of the
pedalia, while the base of the stomach did not give it at all. Stroking
the outer edge of the pedalia of TripedaUa cystophora, the second of the
two species of Cubomedusse described by Conant, also caused the pedalia
to be contracted inwards. I may note here that the muscle fibers under
the ectoderm of the pedalia are specially well developed at and near the
inner and outer edges, both in Charybdea and Tripedalia. On the
flattened sides of the pedalia the muscle fibers are fewer.

When the pedalia were cut off far enough up to remove the
interradial ganglia, coordination was not affected and the animal

E. W. BERGER ON THE CUBOMEDUS^. 13

could pulsate well enough but with little progress. (See above under
Velarium and Frenula.)

An isolated tentacle is capable of squirming contractions, and
when stimulated at either end, it would contract wholly or in part
only.

The pedalia, then, it would seem, serve also as a steering apparatus,
for which they are admirably fitted, considering their blade-like
thinness.

Considering, now, the reflexes noted under this head and the
preceding one, we find that there is an intimate nervous connection
between the velarium and frenula, subumbrella, sensory clubs, nerve,
and a single pedalium, on the one hand, and the pedalia on the other
hand. This is born out fully, furthermore, by the histological evidence
—(See Introduction and "Cubomedusae"). Considering the subumbral
plexus of ganglion cells and fibers, including the velarium and the
frenula, which is in connection with the nerve ring and this again
with the sensory clubs and the interradial ganglia at the bases of
the pedalia, we have a basis for these reflexes. While Con ant failed
to demonstrate nerves ("Cubomedusas") from the interradial ganglia
to the pedalia, yet, that a nervous connection exists between the
pedalia and the bell is well shown by his physiological experiments.
I have, furthermore, demonstrated ganglion cells under the ectoderm
of the tentacles (see Histology).

Romanes obtained quite similar results in the Hydromedusa?. He
found that when a tentacle of Sarsia was slightly stimulated, it alone
would contract, but when it was more strongly stimulated the other
tentacles also would respond as also the manubrium. I find no evidence
in Conant's notes of any such response of the manubrium of Charybdea,
except when the clubs were cut off.

The reflex obtained on stimulating the subumbrella of Charybdea,
when the pedalia would contract, is somewhat different from that
obtained by Romanes, who found that the most sensitive part of the
subumbrella in producing a reflex of the margin was at the junction
of the manubrium to the bell and that the subumbrella below this
point did not give the reflex.

Stomach, Suspensoria, Proboscis, Subumbrella— Experiments 12, 18,
19, 24-26, 29, 31.— The proboscis and the stomach with the phacelli
when cut out, contracted with or without the lips removed. The
isolated lips also contracted (twitched).

14 JOHNS HOPKINS UNIVEKSITY MORPHOLOGICAL MONOGRAPHS.

Pieces of the sides connected only with the stomach and suspen-
soria, or with the margin (Experiment 47 (?)) twitched spontaneously,
but seldom did so when these were removed. In one instance the
whole side was cut out so as to exclude the radial ganglion but still
connected with a portion of the suspensorium. This pulsated, or
contracted, but on being halved transversely, the lower half ceased to
contract while the upper half connected with the suspensorium,
continued to contract.

Cutting off the whole stomach end of the animal excited to very
rapid pulsations of the remaining part, with the stream of water
stronger out the aboral end than past the velarium.

Conant says, "It seems I get no good evidence of the subumbrella
without connection with special nerve centers being able to contract
by itself." The piece in which he did get contractions he suspects
may have been intimately associated with some part of the frenula
or the suspensoria. In Polyclonia no such doubt exists, for small
pieces of subumbrella were seen to contract. A small piece of
subumbrella of Charybdea with a sensory club attached could contract
by itself.

From the above it would seem that a center capable of inciting
to contractions resided in the suspensoria as well as in the sensory
clubs, and this may be one of* the centers that becomes potent upon
the removal of the clubs. This is further supported by Conant's
observation (Introduction and "Cubomedusae") that an extra large
number of ganglion cells is found under the epithelium of the
suspensoria. A somewhat similarly located center of spontaneity
described by Romanes for Staurophora laciniata (Hydromedusa) has
already been noted.

As to the rapid pulsations of the bell after cutting out the
stomach end, this also is similar to Romanes1 results on Aurelia and
other Scyphomedusse, when he cut off parts of the manubrium or an
aboral ring out of the bell. In these instances, however, Romanes
soon obtained a slackening of the rhythm following the temporary
acceleration. The temporary acceleration he attributes to the stimulus
of cutting, and the slackening to a lack of some afferent stimulus
from the removed tissue. Conant obtained the same results on
Polyclonia by removing the oral arms (see Polyclonia) but says
nothing about a slackening of the rhythm in Charybdea. I believe
the increased rhythm in Charybdea was in part due to the decreased

E. W. BERGER ON THE CUBOMEDUS^. 15

amount of labor necessary to force the water out of two openings
instead of one, namely, past the velarium. Just how much this
observation bears upon Romanes' theory of rhythmic contraction,
that the rhythm is due to an alternate exhaustion and recovery of
the contractile tissue, as opposed to the gangiionic theory of rhythm
of physiologists, one does not wish to speculate much. Yet, I feel
that the observation rather supports this theory. The tissue having
to do less work, would become less exhausted at each contraction and
require less time for recovery and hence have a more rapid rhythm.
I here sum up Romanes' theory in a few words. The ganglia
liberate a constant and comparatively weak stimulus, one perhaps
about minimal. This stimulus sets off the contractile tissue ; but as
the tissue contracts and becomes exhausted the constant stimulus
becomes, in relation to it, sub-minimal, and it does not contract
again until it has recovered and the stimulus is again strong enough
to set it off. The gangiionic theory of rhythmic contraction supposes
that the ganglia liberate stimuli to the contractile tissue at successive
intervals. Romanes had this theory suggested to him by the rhythmic
contractions he succeeded in obtaining by subjecting deganglionated
bells to a continuous but weak faradic stimulus, or by placing them
into weakly acidulated water, or into 5 per cent, glycerine. Romanes
claims that his theory better explains muscular tonus and the
contraction of involuntary muscle. He does not, however, hold this
theory to the exclusion of the gangiionic theory, since only too often
does he speak in terms of the latter. He further brings in his
support the fact that the frog's tongue, in which no ganglia have
been demonstrated, can be made to contract rhythmically when
subjected to a weak and continuous stimulus. He also calls attention
to the rhythmic contractions seen in the Protozoa, the snail's heart,
etc. Finally, physiologists are much inclined to explain the rhythmic
contraction of the heart and other involuntary muscles, in part, at
least, as due to a property of the contractile tissue.

Margin, Radial Ganglia, Nerve — Experiments 18, 21-23, 30.—
Complete removal of the margin did not stop pulsation ; but the
removal of the radial ganglia stopped it permanently. While this
experiment seems to have been tried only once, yet, taking into
consideration the results of other operations, it would seem that the
principal centers of spontaneity reside in these ganglia. (It should

16 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

here be remembered that the interradial ganglia were probably
removed at the removing of the margin.)

Cutting the nerve in the eight adradii caused the pedalia to bend
inwards at right angles to their normal position but did not in the
least affect the coordination of the sides. When, however, the sides
were cut in the eight adradii to the base of the stomach, coordina-
tion for the main part ceased, and ecfch side pulsated in its own
rhythm.

I have said that the principal centers of spontaneity reside in the
radial ganglia. Upon further thought this hardly seems warranted.
No doubt, among the principal motor centers must be placed the
ganglionic masses of the clubs, and the radial ganglia, together
with the homologous interradial ganglia, represent centers of equal
value. I speak of these two sets of ganglia as homologous, since
strictly speaking, they both belong to the margin, and the clubs at
whose bases they lie probably represent modified tentacles. Conant's
experiments leave us in the dark as to the function of these ganglia.
Next in order, it would seem, are the ganglion cells in the suspensoria,
as is suggested by the contractions of an isolated side with a portion
of a suspensorium attached. (See previous head.) While we have
seen that the frenula and the velarium can contract by them-
selves, yet, I find no evidence that these can impart their con-
tractions to any adjacent tissue.

Conant's results on cutting the nerve eight times and then
continuing the cuts to the base of the stomach are quite the same as
Romanes and Eimer obtained upon Aurelia. Romanes, however,
concludes that in his Sarsia, Tiaropsis, etc., coordination was broken
when only short incisions were made in the margin. Charybdea
appears, then, to agree with Aurelia rather than with the Hydrome-
dusa?. Yet, since Romanes at first obtained similar results to those
of Charybdea on Sarsia, but on further experimenting concluded that
coordination had really been destroyed at the first cutting, we cannot
speak with certainty that coordination had not been destroyed in
Charybdea before the cuts had been continued to the base of the
stomach. I say not with certainty, because the injury to the bell
being slight, coordination may have been maintained on the principle
of a simultaneously (simultaneous for the octants) alternate exhaustion
and recovery of the contractile tissue on the principle of Romanes'
theory.

E. W. BERGER ON THE CUBOMEDUS^E. 17

Stimulation.— Romanes found when he stimulated a deganglionated
bell of a Hydromedusa, that it responded by a single contraction,
while that of a Scyphomedusa responded with several quite rhythmic
contractions. Charybdea in this respect agrees with the Scyphome-
dusa?. Romanes' results were also verified on Aurelia. (Experiments
12c, 15, 50, 51.)

Activity of Charybdea. — In speaking of the activity of Charybdea,
I cannot do better than refer the reader to the notes. (Experiment
41.) Conant remarks in his dissertation what an active swimmer
Charybdea is, and this is further borne out by his later observations.

Temperature. — Ice in the water seemed to have no effect, except
when held against an animal, when a slowing of pulsation followed
in a few instances. On some pulsating actively in the sun the
temperature of the water was found to be 92° F. (Experiments 33-35.)

Conant does not tell as how cold the water became when he
placed ice in it, but judging from his results, it seems that he might
have obtained a decided slowing of pulsation if the water in which
the medusaB swam had been permitted to approach anywhere near
the freezing point, say 35-40° F. Romanes obtained decided slowing
of pulsation, and even complete inhibition, on a bell of Aurelia, as
also a lengthening of the latent period on some strips cut from a
bell of Aurelia, by lowering the temperature of the water. Replacing
Aurelia in warmer water had the effect of immediate recovery and
increased rhythm. In Aurelia, raising the temperature increased the
rhythm but diminished it when the temperature of the water became
70-80° F. After a slowing of pulsation due to such a rise of temper-
ature, it would not quicken again when the animal was placed in
water of its normal temperature. Romanes explains this by supposing
that the tissue of the medusa had been permanently injured by the
abnormally high temperature. It would be interesting to observe
how the tropical Aurelia behaved under such treatment, seeing that
Charybdea pulsated actively and without apparent injury in water at
92° F. Limnocodium, noted by Romanes, and probably a tropical
species, lived happily in water at 85° F. in the lily house of the Royal
Botanical Society. The temperature of the water could be raised to
100° F. before it proved fatal to this medusa. Such facts point to a
decided difference in the constitution of the protoplasm of tropical and
2

18 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

temperate medusae. Romanes' Sarsia became frantic when placed in
milk-warm water.

While writing the above, I was led to wonder whether the
temperature of the water may not have been the stimulating
influence in those experiments on light (previously noted) in which
the medusae continued to swim actively in the sunlight.

Food and Feeding. — See Experiment 36.

I again make note of a few observations made by myself on
the Olindiad. A crustacean became entangled in the tentacles of a
medusa ; apparently this wished to retain it, for the proboscis reached
in the direction of the crustacean, which, however, got away. I then
placed, by means of a needle, another small crustacean against one of
the tentacles. This was seized but not retained, for the animal
pulsated and it was washed away by the water. Twice I saw a good-
sized crustacean in the proboscis. In one instance the velum appeared
to hold the part of the crustacean not yet in the proboscis. I noticed
another with a crustacean wholly in the proboscis, which was much
lengthened out, the upper part of the crustacean being in the stomach.
The next morning the crustacean was wholly in the stomach and the
proboscis normal. At 5.30 p. M. the crustacean was ejected, nothing
but the shell and some rubbish remaining.

These medusae seem to pay no attention to being touched by one
of their kind, except to give a pulsation or two.

The proboscis appears very "intelligent" in its actions.* First,
some of the tentacles can be seen to contract and to bend inwards,
then the side next the tentacles contracts and the proboscis is seen
to reach in that direction. I could not see, however, what the irritant
was.

Occurrence of Charybdea— Experiments 37-40. — Dr. Conant's remarks
("Cubomedusae") on the occurrence of Charybdea at the surface of
quite shallow water and near the shore (which is quite at variance
with former observations, that the Cubomedusae are essentially deep-
sea forms) are further borne out by his observations at Port Antonio.
As already noted in the Introduction, Charybdea was here found in
abundance in quite shallow water and near shore, but on the

* By no means do I wish to attribute intelligence to these animals.

E. W. BERGER ON THE CUBOMEDUS^E. 1!)

bottom instead of at the surface as at Port Henderson. It is possible
that the animals had been active near the surface earlier in the
morning and that some unknown conditions determined their settling
to the bottom earlier in the former place than in the latter.

Conant's conjecture, " whether these were their natural conditions,
or whether the two forms," Charybdea and Tripedalia, "were driven
by some chance from the deep ocean into the harbor and there
found their surroundings secondarily congenial, so to speak,"
seems to be borne out in favor of the former supposition (for
Charybdea at least),— that these are their natural conditions and that
Charybdea Xaymacana is essentially a shore form.

AURELIA AND POLYCLONIA (CASSIOPCEA)

Experiments 42-53.

Many of the observations on these forms relate to the rate of
pulsation. In an Aurelia, following the removal of a lithocyst, there
was a pause followed by pulsations. In about two minutes rhythmic
pulsations were renewed. Four minutes after the operation there
were nineteen pulsations to the half minute, while twenty minutes
after there were only nine, and these in groups of six and three.
The normal rate of pulsation was twenty-five to the half minute.

Polyclonia behaved much in the same manner as Aurelia. Upon
the removal of lithocyst pulsations continued, but in groups with
short pauses. The normal rate of pulsation was twenty-seven to the
half minute, while three minutes after the operation it was
seventeen, and eleven minutes after, fifteen to the half minute.
The tissue connected with a removed lithocyst gave contractions.
Placing a Polyclonia in fresh sea-water more than doubled the
rate of pulsation, which, however, soon fell to the normal rate, and
lower in one instance. In small individuals the rhythm is decidedly
more rapid than in those of larger size. The few observations on
this point would seem to show that it is in inverse proportion to the
squares of the diameters of the bells.

The removal of a single oral arm or of the whole eight, in
Polyclonia, had much the same effect as the removal of a lithocyst:
there was a decided slowing of the rate of pulsation, while the
immediate effect of cutting was an acceleration or a return to near
the normal rate. About a day later this same animal had quite

20 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

regained its normal rate of pulsation and continued to live over two
weeks. A long latent period followed the cutting of an arm, before
the stimulation of cutting manifested itself.

An Aurelia, with all its lithocysts removed, still gave spontaneous
and coordinated contractions after allowing time for recovery from
the operation. This was the result in one instance, while in several
others only a few contractions were observed. Removal of the
sixteen marginal bodies (lithocysts) in a Cassiopoea produced paralysis
for a time but recovery soon followed. A Polyclonia with its entire
margin removed was paralyzed but had so far recovered in a day
as to be able, at intervals, to give spontaneous pulsations.

The removed margin of a Polyclonia pulsated vigorously. This
margin was then split so as to make a ring within a ring but
connected at one point by a small bridge of tissue. The waves of
contraction, which always originated on the ring with the lithocysts,
passed the bridge to the inner ring quite as Romanes experienced.
The outer ring was next split so as to separate the exumbral
portion from the subumbral, when it was found that the contractions
always originated from the latter. Seven days after its removal,
this same margin was still alive and pulsating vigorously, and
broken-off pieces of the subumbral portion were pulsating by
themselves. Fifteen of the ganglia were removed. It was then
found that while most of the pulsations originated at the remaining
ganglion, now and then contractions originated in other parts where
no ganglion remained. Two days later this margin was still alive
with contractions originating as often from other parts as from the
ganglion. A similar observation was made on a margin of Cassiopoea.

A Polyclonia with the eight lithocysts of one side removed, to
compare with a normal one, gave no evidence of affected coordination.

An oral lobe from an Aurelia could give contractions some
minutes after removal.

In another Aurelia a circular cut was made about the base of
the oral lobes through the epithelium of the subumbrella. The
animal could pulsate well enough but coordination seemed a little
affected, while in another one with a like cut but semicircular, no
effect was noticed.

These results on the removal of the lithocysts (and margin in
Polyclonia) in Aurelia, Polyclonia and Cassiopoea agree quite with
those on Charybdea and, of course, also with Romanes1 and Eimer's

E. W. BERGER ON THE CUBOMEDUSiE. 21

results as to paralysis and recovery following the removal of the
lithocysts, or margin, in Aurelia, Cyanea, etc. I recall no similar
observations, however, on removing a single lithocyst, and the
question of an explanation for the slowing of the rhythm thus
brought about arises. Romanes gives as an explanation for the
slowing of the rhythm (Aurelia, Cyanea, etc.) following the tem-
porary acceleration upon removing the manubrium or a portion
from the center of the bell, as due to a lack of an afferent stimulating
influence upon the ganglia from the excised tissue. May a similar
explanation not serve to explain the slowing following the removal
of a single lithocyst, above noted? The removed lithocyst could no
longer give its efferent stimulus to the remaining ganglia nor to
the tissue, so that the former would have a weaker stimulating-
influence, in consequence of which the latter (the contractile tissue)
would be deprived of a part of the original stimulus of the
remaining ganglia as also of that of the removed ganglion. The
whole would thus result in giving to the contractile tissue a weaker
stimulus, which, again, would require longer and greater recovery on
the part of the tissue in order to be set off by the stimulus at
hand. This explanation is given on the basis of Romanes' theory
of rhythmic contraction previously explained.

Of course, it may be suggested that the musculature had lost
tonus, due to the lack of influence of the removed ganglion (lithocyst),
in consequence of which there was a lowering of irritability on the
part of the contractile tissue. This would require a greater sum-
mation of stimulating influence (Ganglionic theory of contraction)
on the part of the remaining ganglia to set it off. Again, the loss of
irritability on the part of the contractile tissue may have been due
to a lack of nutritive influence from the removed ganglion.

Romanes1 explanation, that the slowing of the rhythm following
the removal of the manubrium and central parts of the bell in
Aurelia and Cyanea is due to a lack of an afferent stimulus on the
ganglia from the removed tissue, likewise explains the similar results
obtained by Conant by removing the oral arms from Polyclonia.

The fact that a margin of Cassiopoea and also of Polyclonia,
connected with but one ganglion, often originated contractions in
other parts as well as from the ganglion, seems to show that
motor centers resided in the margin outside of the ganglia. This
would be somewhat at variance with Romanes' conclusion, that no

22 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

such centers existed in the Scyphomedusse. Conant does not state
whether the Polyclonia margin in question was kept in fresh sea-
water or whether the water was not changed during the seven days.
If the latter is the case, then some poisonous compounds may have
been formed that acted as a stimulus much as weakly acidulated "
water served Romanes in producing rhythmic contractions in
deganglionated bells.

Again, while it is true that no ganglia are known to exist in the
margins of the Scyphomedusa3 outside of the ganglia in the marginal
bodies, yet, ganglion cells and nerve fibers are found in the sub-
umbral part of the margin as well as in the rest of the umbrella.
And as I know no reason why scattered ganglion cells may not
function as ganglia, it is possible that the contractions in question
were spontaneous.

Finally, is it possible that the remaining ganglion originated
the contractions in different parts of the margin, thus acting
at a distance from the points at which contractions originated?
Romanes gives an instance in which he believed to have evidence
that this was the case. Upon a final consideration I am inclined
to this latter explanation.

SUMMARY.

Summing up for Chary bdea, we have seen that it is very sensitive to
light, strong light as also darkness inhibiting pulsations, while
moderate light stimulates it to activity. Also, a sudden change from
weaker to stronger light, or vice versa, may inhibit or stimulate to
activity respectively. This behavior of Charybdea seems to be
correlated with its habit of life on the bottom. We have no
reason to doubt but that the eyes of the sensory clubs are the seat
of light sensation.

The experiments on equilibration are negative, giving us no
certain light on the function of the concretions, though it appears
that they may serve, in part at least, for keeping the sensory clubs
properly suspended. Their function in giving the animal sensations
of space relations is not, however, excluded.

Excision of the sensory clubs demonstrates that they are the seat
of important ganglionic centers, the removal of which results in
temporary paralysis and weakness. That they also are the seat of
organs (eyes, network-cells, concretions) that are of importance in

E. W. BERGEE ON THE CUBOMEDUS^E. 23

giving information in the life of Charybdea, is evident from the
reaching motions of the proboscis after the removal of the sensory
clubs. Other centers of spontaneity in their order of importance
probably are : the radial ganglia (one experiment) ; the interradial
ganglia (?) ; the suspensoria, as shown by their supplying stimuli to
isolated pieces of the sides connected with them ; the frenula and the
velarium, the latter of which gave contractions when removed with
the frenula or in pieces only. No evidence is given that the frenula or
the velarium can impart their contractions to other tissue, though this
seems probable for the former. The proboscis can also contract of itself.

Reflexes between the velarium, frenula, subumbrella, sensory clubs,
nerve, and any one pedalium, on the one hand, and the pedalia on
the other hand, are very common, and point to the pedalia with the
tentacles as organs of defense and offense. The pedalia serve also as
rudders in swimming.

Finally, as judged by the results in this paper, Charybdea seems
to occupy, physiologically, a position intermediate between the
Hydromeclusas and the Scyphomedusa3. In its great activity as a
swimmer, in its response to light, and in its reflexes it is Hydrome-
dusan, while in the paralysis and recovery following the removal of
its marginal bodies, as also in its response with several pulsations
instead of one, when a deganglionated bell is stimulated, it is Scypho-
medusan.

The observations on the Discomedusa?, Aurelia, Polyclonia, Cassi-
opoea, demonstrate the existence of motor nerve centers in the
marginal bodies; but that other centers are present is shown by the
recovery of pulsation following the removal of the marginal bodies
or the margin. These results are mainly confirmatory of those of
Romanes and Eimer. They differ from these in the fact that margins
of Polyclonia and Cassiopoea, with only one ganglion attached,
originated contractions distant from the ganglion. Removing of a
single lithocyst resulted in a slowing of pulsation, as did also the
removal of the oral lobes, though the immediate effect in the latter
case was an acceleration. Isolated pieces of the subumbrella could
contract.

DR. CONANT'S NOTES.

Below follow Dr. Conant's notes. They are printed about as
Conant left them. Their order of succession, however, has been

24 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

changed to bring similar experiments together, while useless and
often repeated ones have been omitted, and short elliptical sentences
completed. Where the present writer wished to add any explanation,
the same has been placed in brackets.

CHARYBDEA.

Light and Darkness. — 1. Eight medusae, in a deep glass jar and
covered by a black coat, except one inch around the top, were
placed in the dark-room.

a. When light from a lamp was thrown on the surface (one
inch) layer, the animals were active near the surface ; when the
light was withdrawn, one or two were on the bottom and not moving
but were probably pulsating.

b. After four or five minutes in the dark, three or four besides a
feeble one are on the bottom. It took about two minutes to get them
all to swim [by the lamp]. Of the three on the bottom, one, at any
rate, was not pulsating. [Three other attempts like a and b were
made, with very similar results.]

2. Experiment No. 1 was repeated several weeks later. Four in
a large round glass dish were placed in the dark-room. A lamp
being held to the dish all but one were found to be on the bottom.
That one quickly went to the bottom, while two of those on the
bottom quickly came to the top. In two or three minutes the one
that had gone to the bottom began to pulsate and at about the
same time the other one that had remained on the bottom also
began to pulsate, while the two that had gone to the top stayed
there swimming very actively. [Repeated with like results.]

3. Fresh ones did not show the reaction to light after darkness
so well as did those in the experiments previously recorded. They
were experimented with about nine A. M., while usually they were
tried later in the day. I had rather suspected from previous work
that they would not react so well when fresh.

4. a. In walking with the jar (1) of jelly-fish of experiment 1
from the dark-room to the back porch of the laboratory (fifty steps),
in the bright sun and a cool breeze, all were found upon entering
the laboratory door to have settled to the bottom and most of them
to have ceased active swimming. In five minutes two or three were
swimming somewhat, and in five minutes more all but one or two
(eight in all) were swimming.

E. W. BERGER ON THE CUBOMEDUS^}. 25

Walking with the jar about the laboratory did not suffice to
make any change in their swimming, nor did blowing on the surface
make any appreciable change.

b. Upon taking the jar to the back porch and placing it on the
stone or cement flags, in the shade and a cool breeze, in four
minutes time all were on the bottom not even pulsating.

Upon replacing them on the laboratory table all began to swim
about at once. [Repeated.]

c. The jar (1) was placed on the back porch again ; in fifteen
seconds three were on the bottom ; in one-half minute all but one.
In three or four minutes all were on the bottom, but two were
swimming lively and the others pulsating. In another minute all
were swimming.

d. The jar (1) was tried again, not resting it on the flags but
holding it by my hands on the sides. The effect was just as quick ;
they stopped pulsating at once. By the time I had got back to my
table in the laboratory, one was at the surface and another arrived
just as the jar was set down.

[Several other experiments of an order similar to those just noted
were tried, with very similar results.]

5. Two buckets stood side by side in the laboratory. One bucket
(1) had more Charybdeas in it than the other bucket (2), and also
had more since brought in (about an hour). The water of one (I)
was also more discolored and with more organic matter (sea weed,
etc.). In the laboratory the animals were active on the surface of
both buckets. Placed in the sunlight on the porch, no breeze, the
sun slanting so that one side of the water in the buckets was
bright while the other side was shaded, the jelly-fish in (1) went
mostly to the bottom, while those in (2) seemed unaffected though
some showed a tendency to go to the bottom after a longer
exposure. The experiment with (1) was repeated and it took some
five minutes for them all to go to the bottom. In a few minutes
after replacing them in the laboratory several were active again on
the surface.

6. Jar (a) with five large ones stood on my table ; they were quite
active. Placed in the sun (no breeze), on the porch, one or two sank
to the bottom at once and the others seemed to slow their activities
somewhat but not very markedly. In a few minutes all were swim-
ming, apparently more actively than before, in the bright sunlight.

3

26 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

[In other experiments Conant shows that it is not the stimulus
of walking that causes them to swim when carried into the room,
for they would not swim when he walked with them on the porch.
Also, he shows how they may change, some swimming, others not,
when left for some time in any one place.]

7. In a tumbler were two pulsating very vigorously. Placed in
the bright sunlight, very little breeze now and then, they showed no
change whatever.

8. Some in a jar were covered with a black coat. The coat was
taken off, and almost immediately they stopped pulsating, or pulsated
but feebly, and sank to the bottom. The coat was put on again with
one part near the bottom of the jar exposed. Almost at once, the
animals, which were quite motionless, pulsating but little, resumed
pulsation, which became more and more vigorous, and quickly swam
to the top again. It seems plainly to be a reaction to light. [Such
experiments as this were repeated at different times with very like
results.]

9. A bucket with several bobbing actively on the surface was set
out in a smart shower, and the animals continued bobbing on the
surface as before. I could not see that they made the slightest
attempt to go below.

There can be no doubt but that there is an individual difference
in sensitiveness to the reaction of light after darkness. E. g., I just
removed the coat from a dish with four in it ; one went to the bottom
at once, another presently, a third remained active at the surface,
the fourth when noticed was on the bottom.

There is also a difference in the length of time they stay on the
bottom as well as in the quickness in the response to light. Some
recover very quickly, should say in less than a minute, and at once
become very active. Some stay for a long time and only resume
activity upon the coat being placed over them. Perhaps this explains
some of the observations in Experiment 1.

Sensory Clubs. — 10. All four concretions were removed and the
animal stood the operation well. It swam more restlessly, however,
than others did in the same surroundings. It seemed at first to show
a trace of loss of sense-perception. It swam up, and down again,
more changeable than those intact, which stay rather more constantly
either on the bottom or at the surface. This may, however, have been

E. W. BERGER ON THE CUBOMEDUS.E. 27

due solely to the restlessness of the animal after the operation. Later
it swam actively for by far the most part on the surface only, which
points to the truth of the preceding statement.

It showed no reaction to light. A coat placed over the jar was
removed, when it was found to be on the surface and it remained
there. This was twice repeated. I noticed specially that on pushing
the bell above the surface of the water it at once turned and went
deeper as the normal animal does. Finally, given another trial with
removing the coat from the jar, it went to the bottom as the normal
animal usually does. After this, when next seen, it was keeping to
the bottom. [This experiment was repeated on another occasion with
almost identical results, no loss of sense-perception being noticeable.]

Sometimes it seemed as if access of light at removing the coat
acted as a stimulus to one or more of those that were quiescent on
the bottom. This was noticed again on the following day.

11. Two more were operated upon. These did not stand the
operation well and stayed on the bottom, one swimming, while eight
hours later one was in better condition (pulsating) than two left in
the same dish for comparison.

12. a. Three clubs were cut off leaving only the stalks. A
temporary paralysis of the power to swim was the immediate effect.
Later it partially recovered this power. The proboscis, which was
previously quiet, now showed convulsive twitchings and movements.
It continued for some time to move to one side and then the other
(after short pauses of varied length) as if to grasp some object.
The lips of the proboscis were also moving and at times expanding.
Often the movements were towards the side on which the club was
uninjured.

b. The fourth club was next removed. A temporary paralysis
as before resulted, followed by a quick recovery of pulsation; but
the animal was now much weakened. The movement of the
proboscis continued— shortening, lips expanding, moving to this side
or that. The pulsations of the bell were kept up even when too
weak to swim.

c. The sensory niches of this same animal were treated with 2.5
per cent, acetic acid by means of a pipette. The stalks of all four
clubs showed white. Pulsations ceased. The velarium showed feeble
local contractions. The movements of the proboscis and suspensoria
drawing down the stomach continued. Upon stirring the animal it

28 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

gave rather feeble, somewhat convulsive pulsations with local
(fibrillar) contractions ;. the pulsations in some cases were pretty well
coordinated, but were more on the twitching kind.

13. Three clubs were removed. The animal pulsated well, only
a little less strongly, perhaps. After a minute or two the fourth
club was removed. It pulsated almost immediately, perhaps thirty
seconds after the operation. It swam very well and pulsated feebly
five hours after the operation.

14. One from jar (a) (Experiment 6) was operated upon.
When the first club was cut off there was a paralysis of pulsation
followed by a quick recovery. Cutting off the second club seemed
to stimulate pulsation, the third to diminish it; after cutting off
the fourth club it still pulsated. When placed in a large jar it
pulsated on the bottom, but not strong enough to swim. The
pulsations were fairly regular and sometimes seemed to occur in
groups of two, but these groups were not well marked.

15. Another one from jar (a) was taken. One club was cut out,
upon which there was a very temporary paralysis followed by good
pulsations afterwards. The proboscis, as in all cases noticed, gave
active movements to this side and that side. These movements of
the proboscis were often very quick and definitely directed as if a
well defined stimulus were given. After the operation one pedalium
contracted so as to be at a right angle to the main axis of the bell ;
shortly a second pedalium also contracted. Placed in a small round
dish the animal swam actively.

A second club was removed, and it swam as well as before. After
fifteen minutes it was not swimming but pulsating against the jar.
Upon stirring it a little it swam vigorously ten to fifteen strokes
and then stopped. It seemed weak and its movements appeared
not so definite, though this might be due to weakness.

A third club was removed. The only change seemed to be
rather greater weakness.

After about five minutes the fourth club was removed. Para-
lysis of pulsation followed. It had the power to contract its pedalia
when these were rather vigorously stimulated with a needle. It
also gave one feeble pulsation when so stimulated.

16. The sensory clubs were removed from another. After removal
of the third one it still pulsated actively, but stopped completely and
apparently for good after the removal of the fourth club. Another

E. W. BERGER ON THE CUBOMEDUS^E. 29

one stopped pulsating apparently for good upon removing the third
club.

17. All four sensory clubs were removed from one, cutting as
high up as possible so as to remove the endodermal tract of nerve
fibers of the peduncle. It pulsated afterwards apparently the same
as if the stalks had been left intact.

18. A small piece surrounding a sensory club and including the
margin can contract by itself. The piece observed pulsated with
quick pulsations and rhythmically but intermittently. After a fresh
cutting away of such a piece, the portion of the velarium attached
was seen to contract rhythmically, while the rest of the subumbrella
was not so seen. The part of the subumbrella above the radial
ganglion that was cut off did not contract by itself. The same
portion of the velarium cut off did give contractions.

19. A sensory club with the surrounding region cut out pulsated
rhythmically ; when the club was cut from the end of its stalk
pulsation stopped. This observation was repeated on another, and
contractions were seen after the removal of the club. A piece of
the subumbrella wall from the same animal gave contractions now
and then even after an hour.

20. The normal position of a sensory club seems to be with
the concretion almost at the lowermost end ; often with it certainly
lowermost, but probably oftener with the perpendicular passing
through the center of the attachment of the club to its peduncle
and just by the inner edge of the concretion. The eyes point inwards.

When the animal is on its side the concretions are always quite
lowermost. When the animal was inverted the tendency was for
the concretions to be lowermost. In this position the eyes may
point in several directions. In one instance those of one club pointed
rather outwards, while of two other clubs they pointed more in the
plane of the body wall. (See also Experiments 24, 29.)

Nerve. — 21. Cutting the nerve eight times, once on each side of
each sensory club, produced no loss of coordination in pulsating.
The animal was weakened, however, by the operation, which was made
drastic to insure cutting the nerve; but it was still able to swim.
This experiment was repeated four times.

22. That coordination was continued after the nerve was cut
was proved beyond doubt by cutting from the edge up (eight times)

30 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

so as to perfectly separate the sensory clubs and the pedalia.
Pulsations continued synchronously in all four sides — not the
slightest evidence that one side contracted out of time with the
others.

23. The eight cuts were made as in the preceding experiment
with no loss of coordination noted. When the cuts were carried up
to the base of the stomach, however, coordination ceased. The four
side pieces seemed to contract each in its own time. Only two sides
could be observed at one time, and they at any rate did not contract
synchronously. One side often gave two contractions while the
other side rested or gave one.

Yet, a little later, three of the sides at any rate showed a
pretty good coordination. The fourth was smaller and did not seem
to get into the game much — it went more on its own schedule.
The four pieces were then cut apart and placed together under a
dissecting microscope. No coordination at all could be made out.
No evidence, therefore, of any definite rate of pulsation inherent in
the sensory clubs.

Cutting the nerve causes the pedalia to forcibly contract inwards.

Side, Subumbrella. — 24. A whole side was cut out, the transverse
cut being above the sensory organ so as to take off [leave off] the
radial ganglion also. This pulsated, or rather contracted, nicely.
The upper end had been cut just through the suspensorium. It
especially gave twitchings like the twitchings of the stomach. The
piece was then halved transversely, when the upper part containing
the portion of the suspensorium twitched as before while the lower
part was not seen to contract again. This was repeated with the same
result, except that a portion of the lower part gave a slight contraction
several times. The part that contracted was at the upper end of the
piece, i. e., nearest the suspensorium. The contractions were also more
longitudinal than transverse, as the regular contractions would be.

The piece connected with the sensory clubs of course pulsated
nicely. Upon cutting off the sensory club from the stalk, pulsation
ceased, but twitching of the velarium continued. This was repeated
with the same effect.

In the same animal, in cutting off the sides, the stomach was
left, the cut being through the gastric ostium. The floor of the
stomach was now cut off by cutting out the four interradial points of

E. W. BERGER ON THE CUBOMEDUSJE. 31

attachment. The stomach and the proboscis gave vigorous contrac-
tions and tied themselves all up so that I could not cut off the
proboscis.

The four pieces of the floor of the stomach left on the interradii
gave contractions nicely. The phacelli continued their squirming
movements.

25. Cutting off the whole aboral end of the animal excites to
very rapid pulsations of the remaining part. The stream, as shown
by particles in the water, is apparently stronger out the aboral end
than past the velarium.

It seems that I get no good evidence that the subumbrella is able
to contract of itself without connection with special nerve centers.
In the one case noted (Experiment 31) I could not be sure but that
the part that contracted was intimately associated with the suspen-
sorium or frenulum.

26. A piece of the subumbrella cut off and having, so far as I
could determine, no connection with ganglia, frenula, or suspensoria,
gave contractions. Another piece was not seen to contract.

A small piece of the subumbrella connected with a club can contract.
The proboscis can give contractions of itself when cut off with the
base of the stomach. Even a cut-off lip can twitch by itself. A
portion of the subumbrella by itself also showed twitchings. (See also
Experiments 18, 19, 25, 26, 29, 47, 49.)

Pedalia, Velarium, Radial and Interradial Ganglia. — 27. The pedalia
with their tentacles were cut off at their bases to insure cutting out
the interradial ganglia. The animal could pulsate well enough, but
intermittently and without much progress (the velarium, of course,
being injured). Cutting one pedalium caused the others to contract.

28. When the pedalia were cut off from one, the power of direct
motion was entirely gone. It swam in circles, turned summersaults,
changed its course continually, the oral end getting ahead of the aboral
end, or trying to do so. The whole power of balancing was gone. It
seemed excited by the operation and swam continually. [Repeated.]

29. The pedalia can be made to contract inwards by stroking their
outer edge with a needle. This was noted last year and has been
seen several times this year. Their inner edge is not so sensitive.

Touching a sensory club caused the pedalia to contract inwards in
two cases.

32 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

The pedalia could be made to contract by giving the subumbrella
a prick,— generally a rather severe one was necessary. The upper
part of the subumbrella seems not so sensitive as the lower part and
the proboscis, and the base of the stomach did not give any reflex
at all (two specimens). One of the two could be made to give the
reflex only with much difficulty. This was a very lively one. It
would even stand severe pricks on the nerve, or even through the
region of the sensory clubs, without contracting the pedalia or stop-
ping pulsations.

Cutting the frenula seemed not to affect the ability to swim well.
Cutting in this region brings about the reflex of the pedalia.

In the preceding fish the velarium was cut away wholly in some
places, in other places it was left only as ragged strips. The pedalia

•

became very strongly contracted and the tentacles were brought inside
the bell. Pulsations that seemed strong produced much less progress
than with the velarium intact. [Repeated.]

30. One with the whole margin cut off still gave pulsations. Upon
the removal of the region of the radial ganglia, however, pulsations
were seen no more.

The velarium in the above continued to give twitchings. The four
pedalia were cut off with plenty of the tissue at their bases to insure
the removal of interradial ganglia, and twitchings of the velarium with
irregular contractions continued. No full contraction all around the
velarium was noticed. When all the tissue was trimmed off as nearly
as possible down to the velarium, the latter still gave twitchings and
irregular contractions as before, — even more so as if excited by the
operation. The power of originating contractions evidently resides in
the velarium or in the ganglion cells of the frenula just as it does in
the proboscis and the floor of the stomach.

Small pieces cut from between the pedalium corners and the
frenula, so as to have tissue on them from neither, could contract
by themselves. (See also for Pedalia, Experiments 15, 23, 41 b ;
Velarium 18, 41c.)

Tentacles. — 31. A cut-off tentacle can contract by itself, sometimes
with squirming contractions. A prick at either end can produce a
forcible contraction. A slight prick at the distal end may produce a
local contraction. The proximal end is more sensitive, but this differ-
ence is not very marked. One with only the tentacles removed
seemed to be a little less able to guide itself well.

E. W. BERGEE ON THE CUBOMEDUS^. 33

Proboscis, Stomach, Phacelli. — 32. The lips of the proboscis are
highly contractile by themselves. The movement of the stomach and
the phacelli goes on, after the lips are cut off, with increased vigor,
due to the stimulus of shock. The vigor and frequency of their
contractions, however, diminish quicker than that of the cut-off lips.
(See for Proboscis, 12, 15, 18, 26, 29 ; Stomach, 18, 24, 29, 31 ; Phacelli,
18, 24, 31.)

Temperature. — 33. Temperature does not seem to have much effect.
Some placed in a tumbler half full of water, in the bright sunlight,
swam vigorously over three-fourths of an hour. The water was quite
warm to the hand.

34. The above experiment was . repeated with the same results.
A thermometer placed in the water with them showed 92° F. ; hung
in the sun near by, it showed 94° F.

Ice in the water did not stop their pulsating temporarily or
permanently, except that it did for a short time after being held
against one. Even then it took some time (fifteen to twenty
pulsations) before it produced any effect.

35. Ice placed in the water again showed no marked effect. They
swam as lively as ever. Some, after pulsating against the ice for a
little while, seemed to be less vigorous, but quickly recovered in
another part of the jar. Others did not seem to be the least bit
affected by contact with the ice.

Food and Feeding. — 36. I tried to feed one. A red and a white
copepod were put into the subumbrella cavity. No attempt to eat it
was observed in either case, though the copepods remained in the
subumbrella cavity for some time.

Animals found in the stomach of Charybdea : small fish were
most frequently seen; at another time a small stomatopod; again, a
small polychaste; small shrimps; amphipod.

Those taken on August 16th (3 to 4 p. M.) seemed to have, for the
most part, food in the stomach, and this more so than those taken in
the morning.

Occurrence of Charybdea.— 37 . In the first tow on the bottom
(with a net made of mosquito-netting and weighted with rocks in
order to sink it) the haul was forty. I do not think that we could
have been towing more than four or five minutes. The time was

34 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

about seven A. M. A light breeze was blowing and there had been a
heavy shower a half -hour previous.

38. The usual time of towing was about 6.30 to 7.30 A. M. The
water was four to five feet (1.2 to 1.5 m.) nearest shore but deeper
farther out. At this time of day one could count on getting plenty
of the larger sized (15 to 20 mm.), many small ones, but very few of the
smallest. This was the experience of several mornings.

On August 12th I towed about nine A. M., and got but few of the
larger sized, many small, ones, and very many of the smallest.

The next day (7.00 to 7.45 A. M.) those obtained were mostly of
the larger size. On the same day (3 P. M.) others of the party towed
at the same place and obtained but few.

On another day I towed in the afternoon (3 to 4 p. M.) and
obtained great numbers as I usually did in the morning.

39. We towed about 7.30 to 8.30 at night. Very few Charybde* were
taken. On this evening we towed five times in the same locality,
and obtained but seven or eight specimens. Towing with the same
net on our way home, it was filled with Aureliae and five or six
Charybdeae. It seems as if Charybdea came to the surface at night.
Those towed in the evening were dead the next morning.

The next morning Richard, our colored attendant, towed from
5.30 to 6.30. There were heavy showers. The usual find of large and
medium ones was obtained. There were only two with planulae.

40. The material of September 2nd was obtained about six A. M.
They were mostly of large size. In all only fifteen or twenty were
taken. Richard explained the small number by saying that the
bottom had changed in the locality where we usually towed and that
he got no weeds in his net, but mud.

The next day more were brought in by Richard (6.30 A. M.)
There were rather more than yesterday but the quality was the same.
There were three with planulae.

On another morning Richard brought in a great many, about a
hundred. Among these there were three with planulae (light-colored
and budding); on a previous day there was one with the reddish-
brown kind and with a mouth.

Activity of Charybdea. — 41. a. About five o'clock in the morning a
Charybdea was taken in the tow. It was in good condition
swimming incessantly round and round without change of direction,

E. W. BERGER ON THE CUBOMEDUS^E. 35

in a jar of about twenty centimeters in diameter. It came to the
surface every now and then, after eight to fifteen pulsations. The
tentacles and the phacelli were of a lilac shade. If a pencil was
placed in its way it would pulsate against it repeatedly without any
effort to dodge around it.

6.58 A. M., 124 pulsations were counted to the minute.

8.00 " 124 " "

9.25 " 136

10.15 " 131
11.00 "

At 10.15 it went around the dish in eight seconds, taking eighteen
or nineteen pulsations. If a bright platinum spatula or a black
pencil was placed in its circuit it would repeatedly butt against
it each time it came around. After the second or third pulsation
against it, however, it seemed to have some sense to change its
direction.

b. The pedalia have no perceptible action of their own. They
move inwards slightly toward the axis at each pulsation, but
scarcely as much as one would suppose from their attachment to
the pulsating margin. It seems as if they were for "winging" the
moving animal more than for anything else.

c. The velarium is loose and it flaps. It seems to take part in
swimming something more than the passive diaphragm function,—
i. e., it straightens out during the recovery after each contraction of
the bell.

AURELIA AND POLYCLONIA.

[The following experiments were performed at Port Henderson,
Jamaica, in 1896.]

42. May 12th. An Aurelia was pulsating normally at the rate
of twenty-five or twenty-six pulsations to the half-minute. One
lithocyst was cut out, when a few contractions, evidently caused by
the stimulus of cutting, followed; then, rest. In the first minute
there were only about five pulsations. In two or three minutes
rhythmic pulsations were resumed. Four minutes after the cutting
there were nineteen pulsations to the half-minute. About twenty
minutes after there were nine to the half-minute, in groups of six
and three.

36 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

A Polyclonia, about four and one-half inches (115 mm.) in diameter,
gave twenty-six or twenty-seven regular pulsations to the half-
minute. After one otocyst was removed, pulsations continued, but in
groups with intervals of pause : e. g., thirteen, pause ; ten, pause ; six.
Three minutes after the removal of the lithocyst there were 5, 3, 1,
3, 5, or seventeen pulsations to the half-minute. Eleven minutes
after the operation there were fifteen to the half-minute. The
removed lithocyst and surrounding tissue gave contractions.

43. May 13th. The Aurelia was in rather poor condition but
would pulsate upon being stirred. The other seven lithocysts were
removed when only a few contractions originated thereafter.

The Polyclonia was in good condition, but was pulsating only
intermittently when first seen in the morning. "When the remaining
seven lithocysts were cut out and no more pulsations were observed,
the oral arms could still move.

May 14th. Both were found dead upon returning in the evening.

44. May 15th. An Aurelia and a Polyclonia were taken in the
morning. The Aurelia was two and one-half to three inches (62.5-75
mm.) in diameter, with three tufts of phacelli, three oral arms and
seven lithocysts. The Polyclonia was normal and seven or eight
inches (175-200 mm.) in diameter.

In the Aurelia all the lithocysts were removed. Spontaneous and
coordinated contractions could still occur after time had been allowed
for the shock from the operation to pass away. The next day the
animal was still alive and pulsating, but ragged, and the next day
following was quite dead.

In the Polyclonia the normal rhythm was fourteen pulsations to
the minute. Some pulsations were apparently quicker than others
and the intervals were not the same. Thirteen, ten, and twelve
pulsations were also counted. After putting the animal into fresh
sea-water, it pulsated thirty-three to the minute. Six minutes later it
was still pulsating at the same rate, while in four minutes more
eleven pulsations, many of which were in groups of two, were noted.
In five minutes more it pulsated eleven times to the minute with
only one double pulsation. One oral arm was then cut off and the
rhythm counted about one minute afterward — fourteen pulsations,
then a pause of fifteen seconds, then two pulsations, in all sixteen to
the minute were counted. About ten minutes later there were eight
pulsations, two or three minutes later only three, while in two or three

E. W. BEEGER ON THE CUBOMEDUS^E. 37

minutes more only three. There was a long latent period — two or
three seconds — before the stimulation of cutting off the arm made
itself evident in the rhythm.

A second oral lobe was removed. Then there followed twenty-four
pulsations, a pause of two seconds, and two pulsations, in all twenty-
six pulsations to a minute. The rate of pulsation soon fell to the
previously abnormal low rate.

Third lobe removed: 21 pulsations in first half minute and then
16, or 37 per minute.

Fourth lobe removed: 17 pulsations in first half -minute plus 13
gives 30 for the minute.

No difference in the coordination of the animal was shown as a
result of the removal of one-half the number of oral arms.

Fifth lobe removed : 17 pulsations plus 15 equals 32 to the minute.

Sixth lobe removed : 17 in first half -minute plus 4 in the second
half-minute gives 21 pulsations for the minute.

Seventh lobe removed : 17 plus 9, or 26 per minute.

In all these instances the rhythm in the second half of the first
minute was irregular and intermittent.

Seventeen and then seven pulsations were provoked after the
animal had become quiescent, or nearly so, by merely handling it.

45. Eighth oral lobe was removed and pulsations stopped. The next
day the animal was in good condition. The pulsations counted in the
evening were 12, 14, 14, 11, per minute. The rhythm was not
regular ; there was a tendency to groups of twos, threes, or more, but
no prolonged intervals of rest were observed. When placed into
fresh sea- water, the pulsations were fourteen to the half -minute or
twenty-six to the minute ; seventeen to the half-minute, and thirty-
three to the minute were also counted. This specimen gave spon-
taneous contractions during two weeks, after which it was thrown
out, the aboral end being eaten through and little or no regeneration
having taken place.

46. Two more were operated upon : A. Its rhythm was 18, 14, 17.
Its entire margin was cut off. The separate pieces of the margin
pulsated, 6, 7, 4, 6, 7, 9. The animal seemed paralyzed by the
operation; it responded by a contraction now and then to stimu-
lation but gave no spontaneous pulsations. B. Its rhythm was 17,
15, 12, 12. All its oral arms were removed. Its rhythm was only
raised to seventeen and not perfect. In twenty-five minutes it had
fallen to eleven, in four hours to ten pulsations [per minute].

38 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

May 22nd. A and B are living as also the pieces of the margin
of A; all are giving spontaneous pulsations now and then at compar-
atively long intervals — even A, with its margin removed.

May 26th. Everything is still living. The one with the margin
cut (A) counted sixteen and nineteen pulsations per minute, though
this was not kept up all the time.

June 2nd. A and B and pieces are still living and contracting
spontaneously. It is now two weeks, and they were thrown out eaten
through at the aboral end with little or no regeneration.

47. The margin was cut off another one (C) and it was then
paralyzed. The margin contracted vigorously by itself. The margin
was next split, but a connection of about one-half an inch wide was left
between the two rings. Over this bridge the contractions passed
from the outer and inner ring. The inner ring did not originate
any contractions. Both rings were then cut near their connecting
bridge of tissue and the larger ring with the marginal bodies was
split longitudinally so as to separate the exumbral from the sub-
umbral portion. It was found that the contractions started only from
the subumbral portion while the exumbral portion did not contract
at all.

June 5th. Five of the eight small pieces of C were not seen to
contract either to-day or yesterday. A slow rotary motion was
observed in some of the pieces suggesting ciliation, but no cilia or
currents pointing to ciliation were seen with a low power. C was
seen to pulsate spontaneously. Possibly it did yesterday but it was
not watched closely. A piece of the subumbral surface of C broken
off (not from the margin) was found to contract spontaneously.

48. June 6th. In a fresh one (D) from Port Royal, the eight litho-
cysts of one side were removed in order to compare its movements
with an intact one. Coordination was apparently unaffected.

June 9th. The margin of C is still pulsating vigorously. Parts
of the subumbrella broken loose from the strip pulsated by them-
selves now and then. Fifteen lithocysts were removed, leaving only
one at the end of the strip. It was found that with this single
ganglion (lithocyst) left, and originating most of the contractions, now
and then a contraction would originate at another part of the strip
where there was no ganglion. Three days later contractions
originated as often from other parts as from the ganglion.

E. W. BERGER ON THE CUBOMEDUS^E. 39

CASSIOPEIA.

[The remaining experiments were all performed in 1897, at Port
Antonio.]

49. Removal of the sixteen marginal bodies caused paralysis for
a time ; then recovery followed.

Contraction was limited to the subumbrella.

A portion of the subumbrella not from the margin can contract
by itself as well as a portion of the margin with the marginal bodies
(lithocysts).

In the margin cut off as a strip with only one marginal body
attached at one end, contractions sometimes started from the opposite
end.

AURELIA.

50. Size, seventeen or eighteen millimeters. Pulsations, thirty-two.
Lithocysts, nine. The operation consisted in the removal of the
concretions with as little injury to the pigmented parts of the
marginal bodies as possible. One whole marginal body, however, was
removed in the operation. Soon after the operation the pulsations
were 28, 26, 20, 20, per minute.

Another one; size fifteen millimeters. Pulsations were forty per
minute. The operation consisted in the removal of the concretions
and pigmented parts of the marginal bodies with as little injury to
the adjoining parts as possible. After the operation it seemed as if
the intervals between the pulsations were irregular,— not a series at
regular intervals. An hour or so after the operation the pulsations
were very intermittent. During the afternoon it was not seen to
pulsate except when it was stirred up, when six or seven vigorous
pulsations followed. These, however, were rather aimless.

51. One sensory club (marginal body) was cut out, including its
basal part also. In one or two other cases more or less injury was
done to adjoining parts also. Pulsations ceased upon the removal of
the last club, but upon placing it in an aquarium and allowing it to
come to rest for two or three minutes, pulsations were now and then
seen. In the evening, this one and another did not pulsate except
when stirred, when they pulsated with good progress.

52. A circular cut, about two inches in diameter, was made
through the epithelium of the subumbrella around the base of the

40 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

oral lobes. The animal pulsated well enough, but the contractions
seemed not so simultaneous in all parts of the margin as normally.
After a few days it had partly regenerated but died. One of the oral
lobes cut off had some power of contraction, and this some time after
the operation. A similar cut, but semicircular, made no difference
between the contractions of the two halves.

53. The whole region of the sensory clubs was cut out when the
animal was not seen to pulsate again, except in the evening, when
pulsations were observed. The oral lobes also moved.

HISTOLOGICAL.

Method. — The following results on the histology of the sensory
clubs, their eyes, and the tentacles, as already noted, were obtained
from some of Dr. Conant's preserved material. These results relate
almost wholly to Chary bdea, with only a few references to Tripedalia,
noted in their proper place.

A portion of this material was killed after keeping the animals
in the dark for some time, for the purpose of discovering any
changes in the pigment of the eyes. I believe that a retraction
of the pigment of the long pigment cells that project between the
prisms and pyramids of the vitreous body in the retina of the distal
complex eye is very evident in eyes killed in the dark. (But more
on this below.)

I obtained my best results from the material preserved in
saturated corrosive sublimate, to which had been added (5 to 10 per
cent.) acetic acid. This- also was Conant's experience in his previous
work on Charybdea and Tripedalia.

My best sections were obtained by embedding the sensory clubs
in celoidin, passing the little blocks of celoidin with the sensory
clubs into chloroform until perfectly transparent, and then into
paraffine. I then cut sections as we ordinarily cut paraffine sections,
mounted and stained them on the slide. My purpose in using this
method was to avoid the displacement of the vitreous bodies of the
eyes during embedding and cutting. This object was fully realized
and more besides. Since the sections cut by the celoidin-paramne
method gave me so decidedly the best differentiation of the axial
fibers of the retinal cells, as also of the cilia, basal bodies, etc., I am
inclined to believe that the celoidin was in part responsible for this
differentiation.

E. W. BERGER ON THE CUBOMEDUS^E. 41

Most of my series were cut 4 (i in thickness. All in all I cut
sixty-five clubs besides making some maceration preparations from
material preserved for that purpose. These sixty-five series represent
material from fourteen bottles. As a whole, my material was good,
but the material from one bottle was decidedly superior for showing
the axial fibers of the prisms and pyramids of the retinal cells.
This shows the advantage of plenty of material. It will be evident
that I had plenty of material.

I found iron-haematoxylin the most satisfactory stain. I stained
for a shorter or a longer time — one-half to several hours and longer—
and then washed out the sections until under a low power of mag-
nification they appeared quite unstained, the nuclei and a few other
parts only appearing darkly stained.

Depigmentation I practiced but little. I obtained many of my
series almost wholly unpigmented, especially those I cut last. Others,
of course, were very heavily pigmented. I am not certain but that
alcohol slowly dissolves out the pigment after a long period of
preservation. Slight variations in the technique of killing and pre-
serving may also, perhaps, determine the stability or solubility of the
pigment, as, of course, also the condition of the pigment at the time
of killing.

Anatomy. — For a short epitome of the anatomy of a Cubomedusa
and of a Cubomedusan sensory club see p. 2 of the Introduction.

The Distal Complex Eye— General. — The distal (larger) complex eye
(Fig. 7) and the proximal (smaller) complex eye (Fig. 13) are so
named to distinguish them from the lateral simple eyes of the clubs.
The distal complex eye consists of the following parts : a cellular
cornea, continuous with the epithelium of the sensory club ; a cellular
lens (externally cellular and internally often quite homogeneous)
immediately beneath the cornea ; a homogeneous capsule just internal
from the lens, and evidently a secretion from the lens cells; a
vitreous body composed primarily of prisms and pyramids just
beneath the capsule ; and a retina of pigmented cells, with sub-
retinal nerve tissue, ganglion cells and fibers. To my knowledge
all observers (except Carriere, who missed the capsule) are quite
agreed on the anatomical structure of the distal complex eye as also
4

42 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

on the proximal complex eye and the lateral simple eyes.* It is on
the histological structure of some of the various parts that differences
exist.

Cornea. — Little need be said on the cornea except that it consists
of flattened cells applied to the outer surface of the lens. It is
continuous with the epithelium of the club and evidently a modified
portion of this epithelium (Fig. 7). All observers conform to this
statement.

The Lens. — The lens is of cellular origin, but in its interior the
cells are often so changed — absence of nuclei, cell walls, and pro-
toplasmic structure — as to make a mass quite homogeneous and
structureless. While this internal mass sometimes shows practically
no structure, yet at other times it is found broken up into masses
much the size and shape of cells but without nuclei, while again,
cells with nuclei may be quite evident. This occasional breaking
up of this mass is evidently predetermined by its original cell
structure. Iron-ha3matoxylin stains this inner mass very dark and
it is difficult to wash out the stain. Borax carmine and Lyons
blue give the best results on the lenses. In figure 7 the lens of
the distal complex eye is shown as quite homogeneous internally,
while in figure 13 (proximal complex eye) it is drawn cellular. In
this latter lens the inner cells are quite round and nucleated as
they may also appear in the distal eye. What I have said applies
equally to the lenses of both complex eyes, though the cellular
nature of the inside of the lens is more readily demonstrated in
the proximal eye.

It appears that it is in younger specimens that the central mass
of the lens shows the cellular structure best, and that as the animal
grows older this structure is more and more lost until no trace

* Haake2 says that in the adult Charybdea Rostonii the vitreous bodies of
the complex eyes are absent but present in the young. It is difficult to
explain this observation except on grounds of imperfect preservation of the
adult material, for in all observations on other forms a vitreous body is
described. Haake evidently did not use sections, and for this reason his
results must be regarded as of doubtful accuracy. Haake also says that the
simple lateral eyes of the clubs are absent in the adult, but present in the
young.

E. W. BERGER ON THE CUBOMEDUS.E. 43

of it remains. As concerns most of my series I could not well
determine which were from younger and which from older
individuals, yet, several series of quite small (5 mm.) and therefore
young animals, in which the eyes were so small that the lenses were
compassed into less than half a dozen sections, the cellular structure
of the lens was very evident.

The external cells of the lens form a spherical shell (both complex
eyes) which, in section, shows as a hollow ring (Figs. 7, 13). The
thicker ends of these cells lie at the inner (toward the capsule) half
of the sphere and the cells taper toward the corneal surface, dove-
tailing laterally with their immediate neighbors as also distally with
those from the opposite side of the sphere. The thicker inner ends
of the cells contain the large nuclei with nucleoli. At a point (*Figs. 7
and 13) on the inner (next the capsule) surface of the lens the cells only
approximate each other and thus leave a place which is easily
broken through, as is shown by portions (drops, probably representing
cells or portions of cells) of the mass within the lens becoming
squeezed out into the substance of the capsule and the vitreous body,
and found occasionally also among the cells of the retina. A
considerable portion of the inside of the lens may be found thus
squeezed out, and its path can often be traced. This phenomenon is
evidently brought about by a contraction of the shell of the lens
during fixation and before the inside of the lens has become
hardened.

In origin the lens is evidently ectodermal, originating from an
ectodermal invagination which becomes pinched off as a hollow
sphere, the outer (i. e. next the cornea) half of which becomes the
lens, the inner half the retina (i. e. vitreous body plus the so called
retina). (See Retina.) The transition from retinal to lens cells is
quite readily made out at the lower side of Fig. 7, but the corres-
ponding structure on the upper left side is not so manifest. It is
further evident that the lens is again an invagination into this
sphere, and the point at which the lens cells approximate (where the
central mass of the lens may be squeezed out as above described)
represents the place of pinching off of the original lens-retina sphere.
It appears, then, that the lens is formed in the lens-retina sphere in
the following manner: The cells of the secondary invagination
going to form the lens begin to lengthen distally (i. e. toward the
cornea) during their invagination to form a hollow sphere, at the

44 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

same time dovetailing with each other and budding off cells to form
the inside of the lens (Figs. 7, 13).

At the lower side of the lens, near the margin of the retina, the
cells of the lens are slightly indented or pushed inwards (Fig. 7, ind.),
I believe this to be due to the weight of the lens in the normal
position of the club, when the lens rests against the margin of the
retina and the capsule and adjacent tissue.

Anticipating the description of the retina, it may here be added,
that the retina is formed from the inner half of the lens-retina sphere.
The cells of this portion of the sphere become differentiated into
prism cells, pyramid cells, and long pigment cells, while laterally,
beyond the margin of the vitreous body, they are differentiated into
pigmented iris cells (Figs. 7, 6a).

Above are my results on the lens. Haake2 speaks of the lens as
consisting of a cellular " Kern " with a covering of lamellated cells.
Carriere describes it as cellular and filled internally with a " Gerinsel,"
or coagulation. Carriere and Haake are each in part right. Glaus
describes it as wholly cellular. Schewiakoff regards the lens as wholly
cellular, and like Glaus has not noted that internally this cell
structure may be quite obliterated. Schewiakoff regards the lens and
retina as formed from an invaginated sphere, and shows the
transition from the lens cells into retinal cells as I have figured.
Conant also gives the structure of the lens for the complex eyes as
cellular but missed the change of structure that the interior of the
lens may undergo.

The Capsule. — The capsule of the lens (Figs. 4, 7) lies immediately
below (inward from) the lens. In structure it is homogeneous, except
for certain fibers from the long pigment cells of the retina that
traverse it, while sometimes also other fibers can be seen which,
possibly, are branches from the fibers just mentioned or continuations
from the fine fibers of the prism cells of the retina soon to be
described. I have, however, no evidence that the fibers from the
prism cells extend beyond the prisms in whose axis they lie. The
capsule lies very closely applied to the lens, never becoming separated
from it in sections, and is, hence, regarded as a secretion from the
lens cells. Just what its function may be is difficult to surmise.
The proximal complex eye possesses no capsule. I have thought,
however, that if the lens should be adjustable, the capsule might

E. W. BERGER ON THE CUBOMEDUSJE. 45

•

serve as a protection to the prisms of the vitreous portion of the
retina during the adjusting movements of the lens. (But more on this
below.) To my knowledge all previous observers are quite agreed
on the structure of the capsule. Carriere and Haake, however,
missed it altogether,

Retina. — While I have enumerated (following previous observers)
the vitreous body and the so-called retina as distinct parts, yet, as
the sequel will show, they are, histologically, different parts of the
same thing — namely the sensorium proper of the eye— and I propose
to use the term retina for both taken together, while I retain the
expression vitreous body (as hitherto used) for the vitreous portion
of the retina. This simplifies matters ; and using a word that is
already used for analogous structures of other eyes (vertebrates,
anthropods, molluscs) is conducive to clearness. I have been tempted,
furthermore, to use the words rods and cones for the prisms and
pyramids that I find in the vitreous bodies of the retinas of the complex
eyes. But since the prisms in reality approximate prisms and the
pyramids pyramids, in their shape, I have decided to retain the
words prism and pyramid for these structures. The former of these
terms (prism) was first used by Conant in his description of the
complex eyes.

What I shall call the retina, then, in the distal and proximal
complex eyes of Charybdea, consists of three kinds of elements :
the prism cells, the pyramid cells, and the long pigment cells. (Figs.
4, 7, 22, prc, pyre, Ip.) We may also describe the retina as com-
posed of three zones: the vitreous zone (vitreous body of authors),
the pigmented zone, and the nuclear zone. (Figs. 4, 7, 22, vb, pz, nz.)

The cells composing the retina form a single layer in the shape
of a hollow cup, into which cup the lens with its capsule fits. (Fig. 7.)
This single layer of cells takes in the thickness of the vitreous zone,
the pigmented zone, and the nuclear zone. Indeed, the distinctions
vitreous zone ( vitreous body), pigmented zone, and nuclear zone
characterize three topographical regions of the retinal cells.

That the retina is made up of three kinds of cells is most
readily demonstrated in transverse sections through the vitreous
body. Fig. 1 is such a section, taken quite near the pigmented
zone (at about the level x, Fig. 4). Three different kinds of areas
are readily made out in such a section. The more numerous areas

46 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

•

(pr) are transverse sections of the distal prisms of the prism cells,
the less numerous and lighter areas (pyr) are transverse sections
of the pyramids of the pyramid cells, and the large oval heavily
pigmented areas (Ip) are the transverse sections of the long pigment
cells. The dots within the two first named areas represent fine fibers
in the axes of the prism and pyramid cells, to be described below.
The presence of three kinds of cells can again be readily seen in
such Figs, as 4 and 7, in which the elements of the retina are cut
parallel to their long axis. (Fig. 22.) Again, a transverse section
through the most distal part of the pigmented zone of a slightly
pigmented retina (Fig. 2) also shows us the presence of three kinds
of elements. The larger and more heavily pigmented areas (Ip) are
the long pigment cells; the smaller, lighter areas (pyre) with a
central dot are the pyramid cells, and the more numerous dots, with
no definite polygonal areas outlined about them (prc), belong to the
prism cells. Thus, I believe, we have conclusive evidence of the
existence of three kinds of cells in the retina of the distal complex
eye.

(a) The prism cells are the more numerous, and, as the name
implies, end distally in a vitreous polygonal prism (Figs. 4, 7, 22, pr).
The prismatic structure of the vitreous body is also shown in Figs.
10 and 11, which are drawn from a macerated preparation of Conant's.
(See the descriptions of these figures.)

In Figs. 4 and 7 the prism cells correspond to the cells with
the darker nuclei (npr); in Fig. 2 they are represented by the dots
without defined polygonal areas about them (prc), and in Fig. 1 by
the most numerous areas (pr). These cells, then, consist of a centrad
portion with nucleus, a pigmented portion with granules of a dark-
br'own pigment, distal from the nucleus, and a distal vitreous prism
which extends to the capsule of the lens.

In the axis of each prism is a fine darkly-staining fibril extending
the entire length of the prism. I found no good evidence that this
fiber extends into the capsule. Centrad this fiber is continued
through the pigmented part of its cell and approaches to or near
the nucleus (Fig. 2, dots without defined polygonal areas; Fig. 7,
part of retina left unpigmented). In some instances I could trace
this fiber quite to the nucleus, while in others it ended before reaching
the nucleus or a little to one side of it. I am inclined to believe,
however, that it extends past the nucleus and is continued as a nerve

E. W. BERGER ON THE CUBOMEDUS^. 47

fiber. I believe this to be so because the fiber is evidently sensory,
and a priori we should expect it to be so continued. Further, I
find decided evidence in sections of the simple eyes to show that the
fibers there extend past the nucleus into the subretinal tissue where
I could not trace them farther. (Fig. 16.) Again, that the flagella
of the epithelial cells of the club are also continued into the cells,
in some instances could be traced past the nuclei (Figs. 12 and 26),
and the fact, too, that the retinal cups of the eyes represent
invaginated epithelium (the axial fibers of the prisms are hence
cilia?)— all this leads me to believe that the axial fibers of the prism-
cells extend centrad past the nuclei through their cells and are
continued as nerve-fibers. (See below under pyramid-cells and under
epithelium). Immediately upon entering the pigmented part of its
cell the axial fiber of a prism-cell has a dumbbell-shaped enlarge-
ment which lies quite at the distal edge of the pigmented part
of the cell (Fig. 7, uiipigmented part of figure). This, of course, can
be seen only in unpigrnented retinas. This dumbbell-shaped body,
(Basalkorperchen of Apathy), which name I give it, since it evidently
is homologous to the basal bodies described by others for the cilia of
epithelia, can be most beautifully seen as two minute spheres lying
close together and in line with the nucleus. These two little spheres
of the basal bodies put to the test the highest powers of the
microscope; but, when, after a prolonged and careful study, one
satisfies himself of their existence and exact shape, the very difficulty
with which they are resolved adds a zest to be appreciated. The
length of a basal body is about one-fifth to one-fourth that of the
nuclei of the prism-cells.

The structure of the nuclei of the prism-cells is that of a dense
network (Figs. 4, 7, npr) which stains dark with heematoxylin. A
nucleolus can often be seen in these nuclei. In some few series,
again, these nuclei did not show a network-like structure, but the
chromatin was arranged in masses (Figs. 5, npr). These nuclei can
usually be distinguished from those of the other cells of the retina
by their denser, darker-staining network (Figs. 4, 7, npr), or as
shown in Fig. 5 (npr). Their denser structure and staining capacity
are a distinguishing characteristic of the nuclei of the prism-cells.
I must add, however, that not in every series is this apparent.

That portion of a prism-cell that contains the nucleus rarely
contains any pigment; and when pigment is present, I believe that

48 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

it has been dissolved in from the pigmented zone. The nucleus,
again, lies a little centrad from the pigmented part of its cell, so that
an unpigmented zone is seen in .the retina between the pigmented
zone and the row of nuclei (Figs. 4, 7, 22).

Centrad the prism-cells are continued as a single process (Figs.
6, b, 'c, d, and 8a, b, c, d). In some sections I thought I could trace
these processes to the basement membrane, but I could not satisfy
myself that such appearances were not due to artificial splitting in
the tissue. Schewiakoff makes a similar remark about his supporting
cells, which cells I believe are the same as my long pigment cells,
but these do not extend to the supporting lamella.

At the margin of the retina the cells do not develop prisms but
remain pigmented and form an iris (Fig. 7), which was so named by
Claus and also described by Schewiakoff. These cells also assume a
somewhat different shape (Fig. 6a). This cell (Fig. 6a) is seen from
its broader side with which it is applied to the capsule or the lens.
Schewiakoff figures similar cells. That the cells of the iris are prism
cells without the prisms does not necessarily follow. They simply
represent cells of the retinal cup that have become differentiated to
serve as an iris.

As to the exact origin of the prisms, and pyramids (to be
described below), it is difficult to say anything definite. If the
so-called basal bodies of the axial fibers are really homologous with
the basal bodies of flagella, then it would seem that they (the prisms
and pyramids) are secretions comparable to cuticular secretions.

(b) The pyramid-cells, like the prism-cells, are differentiated
into three regions: a distal vitreous pyramid, a pigmented part, and
a centrad part with nucleus. The pyramids are seen in transverse
section in Fig. 1 (pyr) and in longitudinal section in Figs. 4 and
7 (pyr)-*

Each pyramid extends between the bases of the prism-cells
about one-third to one-half the depth of the vitreous body (Figs.
4, 7, 12 (pyr). The pyramids are also a shade lighter than the prisms,

•

* In the series from which Fig. 3 is taken the pyramid-cells are not so
readily demonstrated. Indeed, I missed them altogether at first in this and
some other series and supposed that there were only two kinds of cells (19),
but upon a careful re-examination I could demonstrate them to my satisfaction.
They did not show, however, in the particular section of Fig. 3, so that they
are not indicated in this figure.

E. W. BERGER ON THE CUBOMEDUS.E. 49

which fact is characteristic. In the axis of each pyramid is a
darkly-staining fiber quite like the one described for the prism-
cells (Figs. 1, 4, 7, 22). That this fiber extends distally beyond the
limits of the pyramids I could not determine, but I do not think
that it does. Centrad this fiber extends into the pigmented portion
of its cell quite to or near the nucleus as was described for the
fibers of the prism-cells (Figs. 7, 22). Whether or not these fibers
extend past the nucleus and become continued as nerve fibers, the
same course of reasoning holds as was given for the fibers of the
prism-cells. Each of these fibers possesses a basal body just on its
entrance into the pigmented part of the cell (Fig. 7), but I could
not determine that it was dumbbell-shape. In form it represents
an enlargement of the fiber itself, which gradually tapers again to
its normal size. The continuations of these fibers within the pig-
mented parts of the pyramid-cells, as also the basal bodies, could
only be demonstrated in unpigmented series.

Patten5 describes axial fibers extending centrad through the rods
(vitreous portions) of retinal cells (" retinophora") into the region
of the nucleus and past the nucleus (arthropods and molluscs). My
retinal cells (prism and pyramid cells) evidently correspond to Patten's
retinophora, but I find no evidence that one of my retinal cells
represents more than a single cell, while Patten gives evidence that
his retinophora are made up of two cells closely applied to each other
as twin cells. If this were also true for the retinal cells that I have
described, I believe my macerated preparations would have shown
it. Schreiner12" and Hesse13 also figure and describe axial fibers for
the rods of the visual cells in polycha3tous annelids, and Schreiner12a
also for molluscs. Neither of these observers finds the fibers to extend
distally beyond the rods nor centrad toward the nucleus as Patten
and myself show. Neither Schreiner nor Hesse figures these cells as
twin cells as Patten does, so that to my knowing Patten stands
alone in this respect. Andrews11 describes and figures rods for the
visual cells of polychaste annelids but no axial fibers. He was the
first to describe these rods in annelids.

The pigmented zone of the pyramid cells, in heavily pigmented
series, is filled throughout with dark-brown pigment granules, and is
quite like that of the prism cells (Figs. 4, 7). In transverse sections,
however, through the most distal part of the pigmented zone, of
unpigmented series (Fig. 2), lighter areas with central dots could
5

50 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

occasionally be demonstrated, which areas are the pyramid cells. In
Fig. 2, the more definite polygonal outline as well as the lighter shade
of these areas was a distinguishing feature. The difference in shade
was not wholly due to a difference in pigmentation but to a
structural difference.

The nuclei of these cells are usually a little larger than those of
the prism cells and are filled with a finer and less dense network
(Figs. 4 and 7, npyr), in consequence of which they present a lighter
appearance in sections when examined with a high power. It will be
seen in the figures (4, 7) with what regularity these lighter nuclei
lie opposite the pyramids. Some few exceptions occur. These are
probably due to the fact that a nucleus or pyramid was not differenti-
ated by the technique. If this opposition between the pyramids and
the lighter nuclei were all, I believe it would be sufficient evidence
for associating these lighter nuclei with the .pyramid cells.*

(c) The long pigment cells are about as numerous as the pyramid
cells. In these cells, as in the prism and pyramid cells, three regions
can be distinguished : the region of the nucleus, a pigmented region
(the distal half of which extends between elements of the vitreous
body), and a distal rod-like portion, or fiber, which is continued
between the prisms into the capsule of the lens (Figs. 4, 7, 9). The
pigmented portion is about twice the length of that described for the
other cells, and also often of greater diameter, so that in transverse
sections (Figs. 1, 2, 3) these cell-areas are larger than those of the
other cells. As nearly as I could determine, these cells are pigmented
just like the other retinal cells described. In quite unpigmented series,
however, they often contain more pigment than the other cells do

* I go into this at some length because the cell-walls in the series that
showed the nuclei best differentiated as lighter and darker ones did not show
well, and there might be some doubt that these lighter nuclei belonged to the
pyramid cells. I could, however, in many instances, trace the axial fibers of
the pyramids through the pigmented zone to these lighter nuclei (as already
noted) which fact can leave no doubt but that some of these nuclei belong to
the pyramid cells. (Similar nuclei, however, are found to belong to the long
pigment cells, to be described below.) Centrad these pyramid cells are continued
into a single process just as the prism cells were shown to be (Fig. 7). Figures
6, 8, 9, and 21 show samples of all the pigmented cells found in macerated
preparations, and none of these (except Fig. 9, long pigment cells) show more
than a single centrad process. Hence, I conclude that centrad both the pyramid
cells and prism cells are continued as a single prolongation.

E. W. BERGER ON THE CUBOMEDUS^E. 51

(Fig. 2). Distally, the pigmented part becomes narrowed to a strong
pigmentless fiber (Figs. 3, 4, 7). This fiber stains quite dark with
iron-haematoxylin and appears homogeneous. It passes between the
prisms into the capsule, where it usually bends in a direction toward
the margin of the capsule (Fig. 7) and passes diagonally across this
to the lens. In sections, a space is often seen about these fibers in the
vitreous body, which I regard as a shrinkage space (Figs. 3, 4), since
it is not evident in all series (Fig. 1). In Fig. 7, I have assumed that
these spaces are due to shrinkage and have not indicated them. Also,
in this same figure I have assumed that the spiral appearance of the
fibers (Fig. 4) is due to a shortening of the prisms during fixation,
and have drawn them straight. At the lens these fibers seem to
end. In a few instances they were seen to branch upon reaching
the capsule (Fig. 4). In Fig. 9, also, which shows some of these cells
from a macerated preparation by Conant, the rods show evidence of
branching at their distal terminations. In the same preparation I
thought I could see that a fiber became expanded into a membrane
spreading over one of the lens-cells. I could not satisfy myself,
however, that this was the actual condition of things. Judging from
Fig. 9, one might conclude that all the fibers are branched distally;
yet, if such were the case I should have seen more of it in sections,
but branching as seen in Fig. 4 is the exception. Hence, if all these
fibers do branch, I am inclined to believe that it must be among the
bases of the lens-cells. Or, if the fibers do expand into membranes to
cover the lens-cells (I could not explain purpose), the evidence in
Fig. 9 may be nothing more than fragments of this membrane left
attached to the ends of the fibers. As is seen in Fig. 7, most of these
rods end opposite the cells of the lens, and not usually between two
adjacent cells as Schewiakoff has described for Chary bdea marsupialis.
The nuclei of these cells are like the nuclei of the pyramid cells (Figs.
4, 5, 7, 9) and often have a nucleolus.* Centrad these cells are
continued into a number of processes as is seen in Figs. 5, 7 and 9.
How far the several centrad processes extend and where they end I
cannot say; but, as seen in Fig. 5, they soon taper to a thin end
which I suppose may be continuous with a nerve fiber. I believe
Schewiakoff was mistaken when he stated that these cells extend to
the basement membrane.

* I have been able to demonstrate nucleoli in all the different nuclei of the
cells of the sensory clubs.

52 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

I have found no evidence in these cells of the existence of an
axial fiber such as I have described for the prism and pyramid cells.
I find no definite arrangement of the nuclei of the retina into definite
layers, but the nuclei of the three kinds of cells lie quite mixed,
sometimes one kind lying deeper than the other as can be seen in
the figures. Again, they may lie quite at the same level. (This
point will be referred to later.)

It is these long pigment cells that I believe retract their
pigmented part from between the prisms and pyramids when the
medusa? are placed in the dark, protruding with their pigment
when placed in the light. Fig. 5 is a section from a slightly
pigmented retina killed in the dark. The parts of the cells pro-
jecting beyond the pigmented zone, and which would lie between
the prisms and pyramids (here not shown) of the vitreous body are
seen to be narrower than in sections from retinas killed in the light
(Figs. 1, 3, 4, 7) and the cells themselves appear in a condition of
retraction as is shown by their large centrad portions with the nuclei,
which latter, also, here lie at quite a lower level than the other nuclei.
(The pyramid cells were not shown in this series.) I occasionally
found appearances like Fig. 5 in retinas killed in the dark (indeed,
in some the pigmented portions in the vitreous body were much
thinner and more retracted than in Fig. 5). Yet this appearance
was not of sufficiently general occurrence to leave no doubt as to
its significance. As positive evidence, however, I cannot give it any
other interpretation than the one given — that the cells retract
themselves with their pigment when in the dark. Again, it must
be added that the nuclei of these cells may occasionally lie quite
deep even in retinas killed in the light. Indeed, like structures in
different retinas may vary considerably in size and shape. None
of my darkness retinas, however, showed such a large proportion of
the pigmented parts of the long pigment cells projected between
the prisms and pyramids as did the light retinas. I examined
and tabulated all my series with respect to the extent the long
pigment cells were projected into the vitreous body, and I found
that those which showed these cells with their pigment least
projected between the prisms and pyramids to be those that had
been killed in the dark. I thus feel satisfied that the pigmented
parts of these cells become in part or quite completely retracted from
between the prisms and pyramids of the vitreous body when in the

E. W. BERGER ON THE CUBOMEDUSJE. 53

dark, but just how this is accomplished — whether the whole cell with
its nucleus takes up a deeper position, the cell substance at the same
time collecting in the region about the nucleus, as shown in Fig. 6
and the diagram (Fig. 22), I cannot with certainty state. It would
seem, too, as though the pigment became less in the cells exposed to
darkness, for I rarely, even in the most retracted heavily pigmented
series, saw the pigment to extend farther towards the nucleus than
commonly. The time of keeping in the dark, prior to fixing, varied
from three-fourths of an hour to one and one-half hours. I could
not bring the amount of retraction into relation with the time of
exposure, except that in general the retinas longest exposed showed
the greater amount of retraction.

(d) The tissue underlying the retina is described by former
observers (Claus, Schewiakoff, Conant) as composed of nerve-fibers
and ganglion cells. I cannot give it any other interpretation, but I
must add that the supposed ganglion cells are seen only as nuclei,
no cell bodies ever being demonstrable in any of my sections. Conant
also recognized no cell bodies. Occasionally, as in Fig. 7, long fibers
could be traced for some distance in this subretinal tissue, in some
instances quite to or from a visual cell. Pigment was not regularly
observed in this tissue, as Schewiakoff describes, and when present I
believe it has been dissolved in from the pigmented zone.

(e) Schewiakoff describes the retina (my pigmented and nuclear
regions) as composed of spindle-shaped visual cells (my pyramid
cells?) alternating with pigmented supporting cells (long pigment
cells), with the nuclei of the former lying more centrad than those of
the latter. The visual cells are pigmented only at their periphery, or
surface, leaving an unpigmented axis, while the supporting cells have
pigment throughout their whole substance within the pigmented
zone. Distally, the visual cells have hyaline rods, or fibers, which
extend into spaces in the vitreous body, and pass through this and
the capsule to the lens. The vitreous body is described as homogene-
ous, except the spaces for the visual rods, and a secretion from the
retinal cells.

It will thus be seen that my results are quite different from
those just described. I find the vitreous body to be composed of
prisms and pyramids with axial fibers, while the long pigment
cells (supporting cells of Schewiakoff) are continued into the
vitreous body, and becoming narrowed into a non-pigmented fiber,

54 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGEAPHS.

extend to the lens as described. The prisms and pyramids are,
further, the distal continuations of cells whose pigmented and
nuclear parts lie in the so-called retina, but which, together with the
vitreous body, I have named the retina proper. Conant has so sum-
marily disposed of Schewiakoff s distinction between retinal cells based
on pigmentation and location of nuclei, that I need not say more.
Schewiakoff 's Fig. 18 corresponds to my Fig. 1. In this figure he
shows the vitreous body as homogeneous with pigmented areas
(my long pigment cells) and with spaces with his visual rods. It is
quite evident that his spaces with the visual rods correspond to
my lighter areas with central dots ; i. e. my pyramids of the
vitreous body are the same as the spaces shown in his Fig. 18.
It is quite evident that Schewiakoff mistook the lighter areas for
spaces. That they are not spaces can readily be seen by comparing
them with real spaces. It is, of course, possible, too, that the reagents
had dissolved the pyramids, leaving only the axial fibers with a little
pyramid substance about them, and that this is what Schewiakoff
saw. I often found small circular spaces in the centers of the
pyramid areas, as also in the prism areas (Fig. 3), which might be
taken for hyaline visual rods, fibers, in transverse section, but in
such spaces I could usually see a small dot to one side of the space
that I take to be the rod (fiber) proper. Fig. 14 also shows such
small circular spaces that have very much the semblance of. hyaline
rods. This figure is a transverse section of the vitreous body of the
proximal complex eye, in which no long pigment cells or pyramid
cells are present, but it serves well to illustrate the point. The above
explanation also accounts for the large size of the visual rods (fibers)
in Schewiakoff's figures. That the fibers of the pyramid cells (visual
rods of Schewiakoff) do not extend to the lens is quite evident in my
Figs. 4 and 7.

Again, since the long pigment cells are often not seen to termi-
nate in a fiber, but a part of the fiber can often be seen in the
distal part of the vitreous body and in the capsule, it will be quite
readily seen how Schewiakoff should associate his visual rods, or
fibers, with these distal parts of the fibers of the long pigment cells
and suppose his visual rods to extend to the lens.

Again, since the long pigment cells sometimes cannot be seen to
terminate distally in a fiber, while the vitreous body at the same
time may be broken away from the pigmented zone (Fig. 4), it is

E. W. BERGER ON THE CUBOMEDUS^E. 55

quite evident how Schewiakoff should have interpreted the parts of
the long pigment cells in the vitreous body as conical pigmented caps
placed opposite his supporting cells (long pigment cells).

Finally, since Schewiakoff had only twelve marginal bodies to
study, and since this tissue is difficult to preserve properly, I do not
believe that I am doing Schewiakoff any injustice by explaining away
his results as I have done. This fact remains, that Conant and
myself agree in all points in which we differ from Schewiakoff.

To Conant belongs the credit of having first demonstrated the
prismatic structure of the vitreous body, and he also regarded the
prisms as a part of the retinal cells. H. V. Wilson 15>8b suggested,
however, some years prior to Conant, that the vitreous body might
be of a prismatic structure. Conant had evidence also of both the
prism and pyramid fibers, as is well shown in his figures of trans-
verse sections but he found his evidence too meager to make any
very definite statements. Indeed, Conant concludes that there are
three kinds of fibers in the vitreous body and complains of finding
but two kinds of cells in the so-called retina (pigmented and nuclear
zones) to which to refer them. He saw the pyramids with their
axial fibers as lighter areas in transverse sections of the vitreous
body (his Figs. 64 and 68, and my Figs. 1, 4 and 7), but suggests
that they may be the same as the long pigment cells, the cells
having only to project themselves or their pigment in order to
become long pigment cells. This suggested to him to preserve
material both in the light and in the dark. I do not think Conant's
supposition to be a fact, for I find the pyramids in specimens
preserved in the light as well as in the dark. It is, of course,
possible that the pyramid cells are in a stage of structural transition
to the long pigment cells, for, besides their pigmentation, they also
have like nuclei. Furthermore, I held for a long time with Conant
that there may be only two kinds of cells in the retina, but I soon
found the pyramids so definitely shown as to leave no doubt but
that they represented a third kind of cell. For me it remained to
first definitely see all the fibers in the vitreous body as also the
pyramids in sagittal sections.

Conant describes the long pigment cells with their fibers extending
between the prisms of the vitreous body quite as I have described,
and in this my work is only confirmatory of his. Conant does not,
however, describe the several centrad processes of these cells, nor is

56 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

he clear that their distad processes extend to the lens, though he
speaks of fibers within the capsule.

(f) What, now, is the function of these three varieties of cells
of the retina? Schewiakoff regards his visual cells (pyramid cells),
as the name implies, as having a visual function. That they have
such it seems reasonable to suppose, since they have an axial fiber
in their pyramids. If the pyramid cells are visual cells, it appears
that the prism cells also are such. Indeed, since these are the only
ones present in the proximal eye and the more numerous ones in the
distal eye, and like the pyramid cells have an axial fiber in their
prisms, it seems that they are the visual cells par excellence of the
Cubomedusan eye. Also, the analogy between the prisms and
pyramids on the one hand, and the rods and cones of the vertebrate
eye on the other hand, does not seem to be so far fetched. It may
be of interest, here, to briefly consider Patten's theory of color
vision.5*3

The gist of Patten's theory is this : In the eyes of certain
molluscs and arthropods, in the parts of the retinal cells corres-
ponding to my prisms and pyramids, he not only finds an axial
fiber (or fibers) but finer fibrils that extend at right angles from
these axial fibers to the surface of the rods (I shall here, for
convenience, call the prisms, pyramids, etc., rods) where they probably
become continuous with other fibrils in the surface of the rods.
These fibrils from the axial fibers are arranged in superimposed
planes, and if I understand rightly, an axial fiber with its radiating
fibrils may be compared to the axial wire with its radiating bristles
of a brush used for cleaning bottles, provided the bristles of such
a brush be arranged in superimposed planes. The lateral arrange-
ment of the fibrils will, of course, be modified according whether
a rod is circular, hexagonal, square, etc., in transverse section. It
will also be remembered (p. 49) that Patten describes the retinal
cells studied by him as composed of twin cells, and he gives the
name retinopkora to a pair. The system of fibers and fibrils in the
rods he names a retinidium. Centrad the axial fibers are continued
past the nucleus as a nerve fiber. The fibrils extending laterally in
superimposed planes from the axial fiber of a rod, Patten supposes
to be the ones stimulated by the incoming rays of light, the
retinophora being so arranged that the light rays entering them are
parallel to the axial fibers or perpendicular to the lateral fibrils of the

E. W. BERGER ON THE CUBOMEDUS^. 57

retinidium. Again, since the rods are usually the shape of truncated
pyramids or cones the lateral fibrils, which are perpendicular to the
axial fibers, are of different lengths accordingly as they are situated
at the larger or smaller end of a rod. Patten assumes similar fibrils
to exist in the rods and cones (particularly the cones) of the verte-
brate eye, and he thus makes a general application of his theory.
He supports himself in this rather sweeping generalization by the
claim to have demonstrated the twin-cell nature of the cones in
amphibia and fishes.

For illustration, Patten supposes that if red light only were
admitted to the retinophora this would stimulate the fibrils near the
broader end of the cone (but that all the fibrils of the retinidium
would be stimulated a little) and that we would thus have the
sensation of red light. Likewise, if violet light only were admitted,
the fibrils at the narrower end of the cone would be stimulated, and
we should have violet light. Similarly, if light including all the
different wave lengths of the spectrum were admitted, all the lateral
fibrils would be stimulated and the sensation of white light produced.
The method of stimulation need not be that of a vibration of the
fibrils.

Certain grave objections may be raised against such a theory,
the most serious, perhaps, being the fact that no such fibrils as
Patten has described have as yet been demonstrated for the eyes of
those animals that we know have color vision. Yet, as a whole, the
objections are perhaps no more serious than any that can be brought
against other theories of color vision. What Patten's theory does do,
—it gives us a definite mechanical basis to work from, and if these
fibrils should be demonstrated for the rods and cones of vertebrates,
physiologists would then have a mechanical basis for color vision
quite as they now have for hearing. As Patten says, the problem
is primarily a mechanical one. However, the theory cannot well
pass for more than a suggestion, a stimulus for future work, and in
this lies its present value.

It is quite evident that my results for the retinal cells of
Charybdea are, if any thing, a support to Patten's theory. While I
have not been able to demonstrate the fibrils that are the essential
to Patten's theory, yet I have demonstrated the axial fibers of the
rods, and if these fibers should be continued as a nerve fiber to some
central ganglion (as I believe is reasonable to suppose, see p. 47), I

58 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

do not see how we can avoid the conclusion that these axial fibers
of the prism and pyramid cells are somehow concerned in vision.
In Patten's theory these fibers would represent a conducting element,
the real sensory element (fibrils perpendicular to these axial fibers)
not having been demonstrated by me.

I have recently read in a short review of Patten's theory9 that
the evidence we at present have points to the tips of the cones
(vertebrate eye) as being the seat of the sensation of red. This would
be exactly the converse of what Patten's theory supposes. Whether
or not this objection is a real one, future investigation only can
determine.

Hesse13 regards the axial fibers that he describes for the rods in
worms as the primitive fibers of Apathy. In this I agree with him,
regarding the axial fibers I have described as " Primitivfibrillen."
Further, I believe, if I understand Apathy rightly, that the fibrils
described by Patten as extending laterally from the axial fibers
correspond to Apathy's " Elimentarfibrillen."

It is the long pigment cells that are the puzzling element. Since
there can be little doubt but that these cells can project and retract
their pigmented parts (as already described), it would seem that a
part of their function is to check the diffusion of light in the vitreous
body when exposed to strong light. This function would be quite
analogous to that of the pigmented cells of the vertebrate retina,
which in light become projected between the rods and cones. Similar
observations have also been made on the compound eyes of arthropods
by Herrick10 and by Parker7, who find that the distal retinula cells of
Palaemonetes project themselves distad in the dark, thus surrounding
the vitreous cones with a cylinder of pigment, while (Parker) the
pigment of the proximal retinula cells migrates centrad and the
accessory cells move distad; in light the reverse takes place. Other
observations of this kind are not wanting for Crustacea, insects and
arachnids. To my knowledge, the pigment changes that I have
described are the first of their kind for medusae.

I suggested while describing the capsule, that the lens might be
adjustable. That the fibers of the long pigment cells extend to the
lens is my principal reason for this. May these cells not represent
ganglion cells and their distad fibers nerve fibers? That they are not
sensory (i. e. are stimulated by light waves) seems to be suggested by
their not having any axial fiber and in having several centrad pro-

E. W. BERGER ON THE CUBOMEDUSJ2. 59

cesses. These facts suggest that they are not sensory but the center
of a reflex mechanism.* When the sensory cells proper are stimulated,
the impulses are conducted centrad into some nerve center (it may
be the nerve tissue underlying the retina, or other nerve centers such
as the two groups of ganglion cells in the upper part of the club, or
the radial ganglia) from which center, again, impulses return over
fibers leading to the long pigment cells causing them to project their
pigment, and conducting the impulse to the lens, to produce a change
in its adjustment. Since these cells are not so numerous as the
prism and pyramid cells taken together, but in turn have a number
of processes continued centrad (the sum of which processes approx-
imates the number of sensory cells, prism and pyramid cells) it
appears that these cells are admirably adapted to function in just
such a mechanism as I have described, — each long pigment cell
serving a number of its immediate neighbors.

Further, we may conceive each of the centrad processes of the
long pigment cells as receiving a fiber from one of the sensory
cells directly as well as indirectly, as just described. While I
have been able to demonstrate only a single centrad process for the
sensory cells (prism and pyramid cells), yet this does not exclude the
possibility of a nerve fibril passing out from such a centrad process
to one of the processes of the long pigment cells, and it seems
possible that this constitutes the reflex mechanism. That nerve fibrils
ramify in ganglion and sensory cells, and may even leave these cells
to join those of other cells, has been well demonstrated by Apathy,6
so that my finding only a single process of the visual cells leading
centrad without giving off lateral fibers cannot be a serious objection.
Again, fine nerve fibers coming off from the main centrad process of
sensory cells in medusae have been figured by other observers, among
whom I mention the Hertwigs. Careful macerations at the seashore
would probably demonstrate them for Charybdea.

Hesse thinks that the eyes of the Alciopidse are adjustable. He

*It may be objected that my criterion, the presence of axial fibers, is not
necessarily characteristic of visual cells. However, the great general occurrence
of such axial fibers (Patten,5 Grenacher,10 Schreiner,12 Hesse,13 myself, in simple
complex eye, see below, and perhaps others) in eyes in which the retina has
only one kind of cells, would seem to indicate that they are quite character-
istic of visual cells. Note again that in the proximal eye of Charybdea there
is only one kind of cells and with axial fibers.

60 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

describes what he supposes to be muscle fibers just exterior (distal)
to the lens, and believes that a contraction of these fibers would
have the effect of forcing the lens nearer the retina, or vice
versa. His supposition, like mine, needs experimental verification.
Hitherto the only instance known of accommodation in the eyes of
invertebrates was that described by Beer17 for Cephalopods.

The Proximal Complex Eye.— With four exceptions, the description
and discussion given for the distal complex eye also holds good for
the proximal complex eye (Fig. 13). The four exceptions are: the
absence of a capsule to the lens; the absence of the long pigment
cells; the absence of the pyramid cells; and the different relative
position of the lens and retina. This eye, then, has a cornea
continuous with the epithelium of the sensory club, a lens, in
structure and probable origin quite like that described for the distal
complex eye, and a retina of prism cells with axial fibers for the
prisms. Since Conant8b has described this eye quite fully, and
discussed Schewiakoffs conclusions at length, I shall be brief.
Suffice it to say, that Schewiakoff describes two kinds of cells
(supporting cells and spindle-shaped visual cells) for the retina of
this eye just as he described for the distal complex eye. The
vitreous body he likewise describes as being homogeneous and with
spaces for the visual rods (fibers) of the visual cells. It is evident
that Schewiakoff has interpreted the structure of this eye from
analogy with his results on the distal complex eye. Glaus likewise
has described two kinds of cells for the retina, and the vitreous
body as homogeneous. Conant and myself find only one kind of
cells in the retina of this eye. The pigmentation that Schewiakoff
describes for the vitreous body I believe to have been dissolved in
from the pigmented zone of the retina, for I find no regular
pigmentation in the vitreous body. Haake's observation, previously
noted (p. 42), applies also to the proximal complex eye.

Conant's evidence for the axial fibers of the prisms was clearly
insufficient, so that he did not in this respect complete his Fig. 69.
I republish this figure with the prism fibers drawn (Fig. 13).

Since the long pigment cells are absent my reasons for supposing
the lens of this eye to be adjustable vanish.

Finally, a word on the origin of the lens and the relative
position of the lens and retina. The lens and retina in this eye

E. W. BERGER ON THE CUBOMEDUS.E. 61

are evidently not developed from an outer and an inner half,
respectively, of the invaginated and pinched-off lens-retina sphere
(as is true for the distal complex eye) but from proximal and distal
halves respectively. It is also quite easy to understand the
connection of the lens in this eye with the supporting membrane.
Since the cells of the ectoderm of the club can in many instances
be seen to extend to the basement membrane, or supporting lamella,
the cells of the lens, which arise from the ectoderm, simply remain
in connection with the basement membrane, this becoming thickened
to form a support for the lens. That the lens of the distal complex
eye has lost its connection with the basement membrane is evidently
due to the fact that the lens is formed from the outer half of the
lens-retina sphere. The cells of the lens are by this so far separated
from the basement membrane as to lose their connection with it.
Schewiakoff also notes the fact that the lens and retina of the
proximal complex eye are developed from proximal and distal halves
of the lens-retina sphere. He further supposes that the portion of
the basement membrane that acts as a support to the lens takes the
place of the capsule in the distal complex eye. This latter supposition
I do not think probable, since the supporting lamella does not form
a distinct covering to the lens on its retinal side.

The Simple .Eyes.— Since the shape and position of these eyes
have already been described (Glaus, Schewiakoff, Conant), I shall not
tarry long in this respect. Speaking generally, these eyes are flask-
shaped (Fig. 12), the proximal pair quite so, while the distal pair are
drawn out in the transverse diameter of the club. These eyes are
invaginations of the surface epithelium and the shape of the cells
lining these invaginations is quite like that of the epithelial cells,
except that their distal portions (bordering the lumen of the invag-
ination) are heavily pigmented. The proximal walls (Fig. 12, left side)
of the distal pair are heavier pigmented than the distal walls and the
proximal pair of eyes. Schewiakoff calls attention to this point.
The pigmentation is, furthermore, not only heavier, but the pigmented
portion of each cell is much longer in the proximal walls of the
distal eyes (indeed, the cells are longer) than in the distal walls.
The significance of this I do not understand. Indeed, I am inclined
to believe that in life all these eyes are pigmented quite alike and
that it is the reagents used that alter or dissolve the pigment in

62 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

certain places. Yet, the fact that the cells of the proximal walls
of the distal eyes have their pigmented portions nearly double the
usual length, shows some deeper significance.

I also note here the small secondary, non-pigmented invagination
into the tissue of the clubs from each of the distal simple eyes.
Schewiakoff describes this invagination, and it extends in a proximal
and dorsal direction (dorsal-side of club opposite complex eye) from
the dorsal sides of the distal simple eyes. The cells of these
invaginations are not pigmented, but quite like the other pigmented
cells in shape, and like these with distal flagellate fibers. I do not
see the necessity of assuming, however, that these secondary invagi-
nations are the real sensitive parts of these eyes, while the pigmented
parts serve as an iris, as Schewiakoff does in his general discussion.

The histological structure of both pairs of simple eyes is the
same. Sections and macerations give me evidence of only one kind
of cells, all pigmented alike (except, of course, the non-pigmented
secondary invaginations just noted). The cells in these eyes are
very closely crowded so that their nuclei lie at several different
levels. That they all extend to the lumen of the eyes and are all
pigmented could be demonstrated with certainty in many sections,
when some of these cells whose nuclei lay most centrad could be
followed with the greatest nicety to the lumen (Fig. 12). Macerations
(Figs. 8, unlettered cells 21) also show cells with very long cell bodies
pigmented at their distal ends and occasionally with a distal process
or fiber. While there are, therefore, spindle-shaped cells found, yet
they are in every other respect alike, and their differences of shape
and position of nuclei are simply the result of crowding. There is,
therefore, no evidence of supporting (pigmented) cells and spindle-
shaped visual cells (pigmented only externally) as Glaus and
Schewiakoff have described and which Conant and myself cannot
corroborate.

Distally, the retinal cells of the simple eyes have each a fiber
(flagellum) that extends into the lumen (Figs. 12, 15, 16, 21). Each
flagellum has a dumbbell-shaped basal body just on its entrance into
its cell quite like the basal bodies described for the visual cells of
the complex eyes (Fig. 12, part left unpigmented). Each flagellum,
or fiber, can usually be seen to extend into the cell. In one series I
found appearances like Fig. 16, which is a drawing of a part of a
section through one of the proximal simple eyes. This section is

E. W. BERGEE ON THE CUBOMEDUS^. 63

quite in the angle between the proximal complex eye and the group
of network cells in the upper part of the club. In this series I
could very definitely trace the distal fibers of the retinal cells
centrad, past the nucleus and into the subretinal nerve-tissue.
These fibers could be so easily followed that no doubt can exist as
to the fact noted. It thus appears that the axial fibers just
described pass centrad through the cells and are continued as nerve
fibers. On the evidence of such sections as Fig. 16 I have indicated
these fibers as extending centrad through their cells. The lumen of
the simple eyes is filled with a homogeneous vitreous secretion.
This is often incomplete in some parts ; occasionally the secretion
shows a formation of globules, but all this I believe to be due to
the action of reagents. Indeed, I have found simple eyes in which
hardly any secretion was present, while others showed an almost
completely filled cavity. In that portion of the vitreous secretion
just outside the mouth of the distal eyes I occasionally found numbers
of very darkly staining granules. I suspect that these are either
bacterial or algal organisms.

As already noted, Glaus and Schewiakoff describe two kinds of
cells for the retinas of these eyes which neither Conant nor myself
can demonstrate. Further, I believe I have shown that only one kind
exists. If any doubt should still exist, a section like Fig. 25 (which
is from the epithelium of the club, but similar smaller areas with
central dots could often be demonstrated in transverse sections of the
retinal cells of the simple eyes) I believe should be convincing.
Schewiakoff further describes flagella for the retinal cells (his visual
cells) of the simple eyes quite as I have described them for all the
cells. The pigmentation that Schewiakoff mentions as occurring in
the secretions within the lumina of these eyes I believe to have
been dissolved in from the pigmented zones. I find no definite pig-
mentation in these vitreous secretions. These secretions are evidently
products of the retinal cells and have been so regarded by former
observers.

Lithocyst and Concretion. — The cavity filled by the concretion is
lined in places by a single layer of cells, two of which are shown in
Fig. 7. This fact has been noted by both H. V. Wilson and Conant.
Such cells are evidently remnants of the cells that formed the con-
cretion. The supporting lamella completely surrounds the cavity of
the concretion.

64 JOHNS HOPKINS UNIVERSITY MOEPHOLOGICAL MONOGRAPHS.

The concretion filling the lithocyst has the shape of a hemipro-
late spheroid cut in the plane of the axis of revolution. Whether
it is of endo- or of ectodermal origin, I believe developmental studies
only can determine. Tests made in the Chemical Laboratory show
the presence of calcium sulphate with perhaps a very small trace of
phosphate.* Nitric acid slowly dissolves these concretions, but I
believe Glaus was mistaken when he said that they dissolve with an
evolution of gas. I watched them dissolve under the microscope, and
never could see the least bit of gas formed. If Claus's observation is
correct, then the composition of the concretions of C. marsupialis is
different from that of the concretions of C. Xaymacana. The concre-
tions, further, were dissolved out of the material preserved in formaline
and in osmic acid solutions. For dissolving them in situ I used either
nitric or hydrochloric acid, or both. A slight husk remains after all
the lime is dissolved.

The Epithelium of the Clubs.— The epithelium is thickest on the
dorsal side of a club. The thickening here, as in several other
places, seems to be due to a crowding of the cells, in consequence of
which the nuclei come to lie at different levels, but I believe that all
the cells quite reach the surface. The cells with their nuclei nearest
the surface are pyramidal in shape, with the bases of the pyramids
toward the surface, while those cells whose nuclei lie deeper (where
several layers of nuclei occur) may be spindle-shaped (Figs. 12, 23,
24, 26). Centrad these cells are continued into a single process, which
often seems to extend to the basement membrane (Figs. 7, 12, 13, 23,
24). Where the epithelium covers the region of the concretion, the
cells become flattened and with the long axis of their nuclei parallel
with the surface of the club (Fig. 7). The same holds true for the
corneal epithelium (Figs. 7, 13).

It is a significant fact that in many places the nuclei form only
a single layer, and in such places one cannot speak of spindle-shaped
cells. I cannot find any evidence of sensory and supporting cells as
Schewiakoff describes. The fact that spindle-shaped cells may exist
is simply a physical consequence of their being closely crowded.
Conant arrived at the same conclusion.

But I have another and better reason for supposing the existence

* Mr. J. C. Olsen, of the Chemical Laboratory, kindly made these tests for me.

E. W. BERGER ON THE CUBOMEDUS^E. 65

of only one kind of cells in the epithelium. In a tangential section
taken just through the tips of the epithelial cells (Fig. 25) I find
polygonal areas with a central dot. This section does not at all agree
with Schewiakoff s Fig. 8, in which he figures two kinds of cells. In
Fig. 25 there can be no evidence of two kinds of cells, unless both
kinds have like flagella, for these dots are the transverse sections of
flagella continued within the cells (Fig. 26).

The epithelium, then, is flagellate, a flagellum to a cell. Whether
there are flagella on the epithelium covering the region of the con-
cretion, I could not determine. But I believe that in all other parts,
excepting, of course, the corneas, it is flagellated. The fibers (flagella)
of the simple eyes are evidently the flagella of the invaginated
epithelium. Each flagellum has a basal body, and I could in many
instances determine that it was dumbbell-shaped (Fig. 12). This fact
was not always evident, however, and it was only occasionally that I
felt sure of it. Often the flagella showed only a general thickening
within the cells (Fig. 26) while, again, the thickening (basal body)
might be quite localized near the surface of the cell. Each flagellum
extends into its cell, and occasionally I could trace one clear past the
nucleus into the subepithelial nerve-tissue (Fig. 26), just as I did for
the axial fibers of the retinal cells of the simple eyes. In those
instances in which I could do this, the fibers could so clearly be
traced that little if any doubt can exist. I have thus made bold
and have drawn the flagella as continued through their cells into the
subepithelial nerve-tissue for all the cells of the epithelium of Fig. 12.

A word on the epithelium covering the network cells of Fig. 13.
Conant and Schewiakoff here describe fibers from the supporting
lamellas that pass in bundles in among the network cells. These
fibers are supposed to be a part of the supporting lamella which
reaches out to be a support for the epithelial cells. (Schewiakoff also
describes similar fibers for other parts of the epithelium.) Now, as
Conant himself shows in Fig. 13, these coarse fibers are not of the
same consistency and staining capacity as the supporting lamella. I
found them to stain just like the intracellular parts of the flagella or
like the central continuations of the axial fibers of the cells of the
simple eyes. I could, also, occasionally trace them to the surface of
the epithelium, and beyond, when they became continued as short
blunt processes or flagella (Fig. 13). I, therefore, conclude that they
are sensory fibers like those I have described for the other epithelial
6

66 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

cells. Yet, that they pass to the supporting lamella, just as Conant
shows in Fig. 13, would seem to indicate that they are fibers from
the supporting lamella or processes of the epithelial cells. While this
stands as an objection to their being sensory fibers, yet I cannot
explain away their being continued distally as a flagellum, except I
assume this continuation to be an artefact. This does not seem
probable. Perhaps they serve both purposes ; namely, that the cell
body with its axial fiber is continued to the supporting lamella, the
cell proper ending there, while the axial fiber is continued as a nerve
fiber. I believe this to be the proper explanation.

The epithelium of the peduncle is quite like the epithelium of the
club just described. Sections through the tips of the epithelial cells
of the peduncle and also sections sagittal to the axis of these cells
give sections like Figs. 25 and 26. I, therefore, conclude that this
epithelium is a sensory flagellate epithelium like that of the clubs.
Nerve tissue and unstriped muscle fibers underly the epithelium of
the peduncles. Glaus and Conant also describe a small ventral endo-
dermal tract of nerve tissue, which according to Conant is connected
with the endodermal nerve tissue found in the region of the radial
ganglia.

To sum up, the epithelium of the club and the peduncle is a
flagellate sensory epithelium whose flagella are continued through
the cells as nerve fibers into the nerve tissue below. A priori,
judging from the mass of nerve tissue underlying the epithelium,
we should expect the epithelium to be one strictly sensory. What
sense it serves is difficult to surmise. In the physiological part of
this paper I suggested that it might be tactile, serving in connection
with the lithocysts in giving the animal sensations of space relations.

Glaus mentions having seen patches of flagella on the epithelium
of the clubs. Schewiakoff supposes that his spindle-shaped sensory
cells have only a single flagellum, while his supporting cells have
many cilia. In the latter supposition he was evidently mistaken.
Conant (from an unpublished note) saw the flagella of the epithelium
on the living object and does not think that there could be more
than a single one to each cell. He also concludes from living speci-
mens squeezed out under a cover-glass, that there is only one kind
of cells in the ectoderm.

Cilia and flagella extending into the cells to which they are
attached are described by a number of observers.

E. W. BERGEE ON THE CUBOMEDUS.E. 67

I shall not endeavor to discuss the subject further, but shall
append the literature on the subject that has come to my notice.
(See Literature). Some of these observers ascribe a nervous function
to these centrad continuations. I am inclined to believe that they
represent the primitive fibrils of Apathy, whether the cilia or flagella
are motile or sensory. I should mention, however, that Apathy has
traced the " Primitivfibrillen " to be continuous with cilia, and also
traces them into the sensory rods of the sensory cells in the sense
organs of leeches. Eimer also describes cilia as continued centrad.

The Network Cells and the Multipolar Ganglion Cells. — Conant is
the first to accurately describe the true structure of the network
cells (Fig. 13) that fill the upper part of the club between the
proximal complex eye and the attachment of the peduncle. I cannot
add anything to Conant\s description. As their name implies, they
are filled with a coarse network-like structure with a central nucleus
and nucleolus. Schewiakoff erroneously described them as ganglion
cells and Glaus as supporting cells. I have sometimes thought that
they are not made up of a network, but of a vesicular structure ;
i. e. the network we see is really produced by the sections of
planes that intersect to form little polyhedral cavities. I could not,
however, satisfy myself on this point. I further saw similar but
smaller cells, with a finer network, disposed in small groups laterally
and distally from the attachment of the peduncle to the club.

What the function of these network cells is can only be guessed.
In size and shape they somewhat resemble some of the cells found
in luminous organs. Conant, however, nowhere mentions that
Charybdea is luminous.

Lateral to the larger group of network cells lie two groups of
large multipolar ganglion cells (a group on each side). Glaus
describes these cells, but Schewiakoff does not specially note them,
and evidently considered them a part of the network cells, which
he erroneously described as ganglion cells.

The Nerve Tissue. — I cannot add anything new on this. It
consists of fine fibers and ganglion cells, quite as described by Glaus,
Schewiakoff, and Conant, and fills the club between the ampulla
and the epithelium, except the spaces occupied by the eyes, lithocyst,
and network cells. It is likewise present under the ectoderm of the

68 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

peduncle, where also a small tract is found under the endoderm.
(See preceding head, or Glaus3, and Conant8b). As already noted,
under the distal complex eye, I find only large nuclei to represent
the ganglion cells. By saying this, however, I do not wish to dispute
their ganglionic nature. The large multipolar ganglion cells I have
noted under the preceding topic.

The Supporting Lamella.— The supporting lamella is a continua-
tion, through the peduncle, of the jelly of the bell. It completely
surrounds the ampulla and the lithocyst, and also forms a partition
between them, so that, as already noted, the lithocyst becomes
completely surrounded by it. It also sends a partition ventrally
between the complex eyes (Figs. 7, 13). Its thickening to form a
support for the lens of the proximal complex eye has already been
noticed. I shall limit myself in the discussion of the supporting
lamella to the above short resume, since Schewiakoff gives further
detail.

The Endothelium of the Ampulla and the " Floating Cells." — The
ampulla is lined by a secreting epithelium. This is shown by the
large masses of a secretion within the bases of the cells, and by
smaller masses scattered in the central and more distal parts
(Figs. 7, and 27, lower half). The section of the cells is such in Fig.
7, that the bases of some (those nearest the supporting lamella) are
taken, the central nuclear region of others, and the tips of those
farthest from the supporting lamella. The section may be said to be
taken diagonally through the bases and central parts of some of the
cells, but owing to the curvature of the ampulla wall, through the
tips of others. The secretion is a colloid substance, staining yellowish
gray with iron-haematoxylin, blue with Lyons blue, and reddish
with borax-carmine. Sometimes darkly staining rods and fibers of
unknown origin could be seen within the larger masses of the
secretion (Fig. 7). These rods and fibers could also be seen in
spaces within the cells, from which the secretion had evidently been
dissolved. I think there can be no question but that the masses
described are a secretion. Many series, however, do not show it;
indeed, an examination of Conant's slides gave me little evidence
of a secreting function, though I could demonstrate it in his sections
both within the endothelium and also the floating bodies. The

E. W. BERGER ON THE CUBOMEDUS^E. 69

presence or absence of this secretion is evidently correlated with the
feeding habits of the animals, or else it would be more generally
present.

The endothelium is thickest (the cells are longest) in the upper
part of the ampulla where the supporting lamella approaches the
lens of the proximal complex eye, and in the lower portion of the
ampulla (Fig. 7), in the angle between the concretion cavity and
the region of the distal complex eye. In general, the cells are
longest in the upper part of the ampulla, while in the lower part,
especially where they cover the concretion cavity and the dorsal
wall, they may be quite cubical instead of columnar. Often they
present a vacuolated appearance at their bases (Fig. 27). Glaus and
Schewiakoff describe and figure this endothelium, but not in detail.
No one, to my knowledge, has described this secretory function.

The nuclei of these cells are peculiar. They may contain a
network with a nucleus (Fig. 27). Again, they may show evidence
of amitotic division (Fig. 20, h, i, j). Indeed, Remak's scheme (Wil-
son18 "The Cell," p. 4(3) can be quite readily demonstrated. It is,
however, such dumbbell-shaped, elliptical, or ringed nuclei as seen
in Figs. 7 and 20 that are of special interest.

I have spoken of some of these nuclei as dumbbell-shaped,
elliptical, or ringed. This is so, however, only in sections. They are
really flattened spheres with a rod of tissue, of the same structure
as the nuclear wall, stretching between the poles. One may con-
veniently compare the shape of these nuclei with that of an apple,
the core of the apple representing the rod connecting the two opposite
flattened or slightly hollowed poles of the nucleus. For convenience
I shall call the rod connecting the two poles the axis of the
nucleus. The dumbbell or elliptical shape would be obtained by a
meridional section through the axis (Figs. 20, a, b, c, e, g, k, 1, m, n,
o, 7). Likewise a ringed appearance with a central dot would be
obtained by a section parallel with the flattened surfaces or perpen-
dicular to the axis (Figs. 20, d, 7). In a section not strictly meridional
the axis would be cut as in Fig. 29, a, or not show at all. As nearly
as I could determine, the inside of these nuclei is a vacuole, which
the axis penetrates.

The walls and axis of these nuclei have the structure of a very
fine and dense network that stains very dark with iron-haematoxylin.
It stains quite like the reticulum of any nucleus, but is very dense,

70 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

as though all the reticulum of the nucleus had been crowded together
at the surface. Judging from appearances like p (Fig. 20), the
hollowing out, so to speak, of these nuclei, would seem to be a
process of vacuolation, the reticulum becoming crowded aside to the
surface. But how, on this view, to account for the formation of the
axis, I do not know. Perhaps the axis is formed by a pushing in of
two opposite poles of a nucleus, the two invaginations meeting and
fusing. On this supposition one might expect the axis to be hollow
(cylindrical), but I could not determine that it was. Perhaps the
centrosphere (or spheres) (see the next paragraph) has something to
do with the formation of the axis (Fig. 20, b, g, e, etc.).

In the nuclei of Fig. 20 with the dark outlines, and of Fig. 7
a small reticular body is seen just opposite one end of the axis, or
opposite both ends in g. In d (Fig. 20) this body is seen next the
axis just below (outside) the hollow cup represented by the hollow
ring. In this instance a central granule is seen in the reticular
body, as also in c. I take this reticular body to be the centrosphere,
and the central granule in c and d the centrosome. In k, 1, m, n,
and o (Fig. 20), which are from another series, in which the walls
of the nuclei did not stain so dark as in the other nuclei of the
same figure, a nucleolus could be definitely seen, indeed, sometimes
quite perched upon the wall of the nucleus (k, 1). In several
instances I could see two nuclei, as in o. But besides these nucleoli,
I could in several instances see quite definitely a reticular body
(centrosphere) opposite the axis (m, n, o) quite as I described for the
nuclei with the dark outlines. In a, b, c, d, e and g the nuclei could
not be so readily demonstrated, but I could occasionally see a darker
stained body as in a, c and g, that I have no doubt is the nucleolus,
which here, again, is perched quite upon the surface of the nucleus.
This position of the nucleolus is perhaps due to its having been
crowded to one side by the nucleus becoming hollow. It is no
uncommon thing, either, to find several nuclei in a single cell,
sometimes in process of division or just divided as o and e (Fig. 20),
also h, i and j. The whole nuclear phenomenon that I have described
seems to be one of division. Perhaps it is somehow associated with
the giving off of the secretion of the cells, for these nuclei seem to
be found in greatest abundance in those cells in which the secretion
is most abundant. In Conant's sections I found but little evidence of
these nuclear phenomena as also little secretion, which all goes to

E. W. BERGER ON THE CUBOMEDUS^. 71

show the association of the nuclear phenomenon with the secretion.
I have failed to find any descriptions in the literature of nuclei to
which I could refer my observations.

The endotheliurn of the ampulla is flagellated (Figs. 7, 17, 27).
It will be seen that there are two slender flagella to a cell. Each
pair of flagella has a pair of basal bodies that are longer than thick,
and which are continued as a thin fiber towards the nucleus of the
cell. That these centrad continuations of the basal bodies extend to
or past the nucleus I could not determine. Sometimes the basal
bodies with the centrad continuations are pushed quite to one side
of the cell (Fig. 27), while in other cells they are applied quite to the
distal surface (Figs. 7, 17, 27). Fig. 17, and the part of Fig. 7 that
shows these points, are taken just through the tips of the cells. The
darker lines within the polygonal areas are the intracellular basal
bodies with their centrad continuations, while the thinner lines are
the flagella, and are supposed to lie in the plane just below the
plane of the figure. In those instances in which the centrad con-
tinuations are applied to the distal surface of the cells they could
occasionally be seen to bend centrad (Fig. 27b). While these cilia with
their basal bodies and centrad continuations are usually separate, as
shown in the figures, yet they are at times applied quite closely to
each other so that the double nature of the basal bodies and their
centrad continuations is not evident. When the intracellular con-
tinuations of the cilia become pushed to one side or applied to the
distal surface of the cells, I believe this to be due to the turgor of
the cells consequent upon the deposition of large masses of secretion
within them. But I must add that this explanation is not alto-
gether satisfactory, since in the endoderm cells of the pedalia of
both Charybdea and Tripedalia I found like conditions with no evi-
dence of a secreting function. (See below, under tentacles.) No one,
to my knowledge, has described the flagellation in detail, although
both Claus and Schewiakoff state that the endoderm is ciliated.

The "floating cells" in the stomach pockets and in the ampulla,
described by Conant, I believe are in part derived from the endo-
thelial cells of the ampulla. That a portion of them may arise from
the ovary, as Conant explains, I do not doubt ; I have, further, found a
mass of floating cells in a small Charybdea quite as Conant describes
for Tripedalia (his Fig. 71). In this Charybdea, however, I could find
no traces of any ovary. Conant speaks of larger and smaller floating

72 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

cells, and that the smaller ones are also found in the males. This
latter fact agrees with what I have suggested, that some of the
floating cells arise in the ampulla. My chief reasons for my suppo-
sition, however, are the following: I find globules of the secretion
of the ampulla cells in some of the floating cells and also scattered
loosely among them (Fig. 19). These globules in and among the
floating cells have the same general appearance and a similar
staining capacity as the secretion in the ampulla cells. Again, in
spaces within some of the ampulla cells I find bodies resembling the
floating cells with lumps of the secretion within them (Fig. 18).
The conclusion, therefore, lies near that some of the floating cells
originate within the cells of the ampulla, engulf within them some
of the secretion, and are then expelled into the lumen of the ampulla.
Better said, perhaps, they represent portions of the ampulla cells
with some of the secretion. I also found several instances in which
a floating cell had the appearance of being expelled from an ampulla
cell. Conant suggests for a similar observation that the cells were
about to be swallowed by the ampulla cells. I believe, however, that
my finding a secretion similar to that within the cells of the
ampulla, in some of the floating cells, as also bodies very much
like them and filled with secretion within the ampulla cells,
together with ConamVs finding floating cells in males, and finally
the observation that the floating cells are usually quite dilapidated,
never showing a healthy cell structure— all this leads me to conclude
that some of the floating cells originate from the ampulla cells, and
that they have a nutrient function in distributing the secretion.
This is quite the reverse of what Conant supposed, — that they were
taken in as nourishment by the ampulla cells. I also find what
appears to be a secretion in the endoderm of the tentacles of both
Charybdea and Tripedalia, and believe this is another source of the
floating cells. (See below, under tentacles.)

I also found other very darkly staining bodies (Fig. 19) both
within the floating cells and free in the ampulla cavity, and more
numerous in the ampulla cells themselves. This again goes to show
that floating cells take their origin from the ampulla cells. What
these darkly staining bodies are, I cannot say. Perhaps they are
something akin to the " Chromatoider Nebenkorper '' described by
Lenhossek (L), or they represent another kind of secretion. If these
floating cells are derived from the cells of the ampulla, the active

E. W. BERGER ON THE CUBOMEDUSJE. 73

nuclear division within these also receives an explanation. Some
nuclear matter can usually be observed in the floating cells.

The Endothelium of the Peduncle.— The endothelium of the peduncle
consists of flagellate columnar cells (Fig. 27, upper half). The cells
are vacuolated at their bases like some of the cells of the ampulla,
and contain a comparatively large nucleus with nucleolus. The
flagella are long and slender, quite like those described for the cells
of the ampulla, except that there is only one to each cell. The basal
bodies of the flagella are of a peculiar shape. They may be described
as a bent spindle, continuous at their distad ends with the cilia and
at their centrad ends with a fiber that can be traced quite to the
neighborhood of the nucleus. I could not trace these fibers into the
basal parts of the cells, except in one instance, and I could not be
sure of that (Fig. 27a).

Another interesting observation in connection with the basal
bodies is that they are bent in one direction on one side of the canal
and in an opposite direction on the other side. In Fig. 27, which
represents a longitudinal section of the endoderm and the supporting-
lamella of the dorsal (i. e. farthest from the eyes) side of the peduncle,
the distal ends of the basal bodies are bent towards the ampulla,
while on the ventral side they would be bent away from the ampulla.
This seems to suggest that the flagella move the contents of the
canal in one direction on the dorsal side of the canal and in an
opposite direction on the ventral side. Conant observed in living
material that bodies in the ampulla and the canal were moving
about, and that bodies within the tentacles were moving in opposite
directions at the same time. This last observation and the histological
facts just described, I believe, are mutually corroborative. Again, a
priori, we should expect some such mechanism as the one described
to bring about an exchange between the contents of the ampulla and
that of the stomach pockets. I have not as yet been able to demonstrate
a similar flagellate mechanism in the tentacles. Flagella and basal
bodies are present in the tentacles, but I could not determine that
the basal bodies had any definite arrangement like that shown in
Fig. 27. (See under tentacles.) I may add, yet, that the cells in the
canal of the manubrium have cilia, similar to the ones just described,
with large basal bodies, and with centrad continuations. Finally, I
am not certain but that these cells form buds at their ends quite

74 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

like those I describe for the endothelial cells of the tentacles (see
below), and that they aid in the formation of the floating cells. I
thought I saw such buds just at the entrance of the lumen of the
peduncle into the ampulla, but could not find conclusive evidence.

The Tentacles and the Pedalia. — My observations on the tentacles
were begun with the object of demonstrating a flagellate mechanism
similar to the one described above for the endothelium of the
peduncle. While I have failed to demonstrate such a mechanism for
the tentacles, yet several interesting points came to my notice. It
will be remembered that the tentacles of the Cubomedusae are not
directly attached to the bell, but that a blade-like portion, the
pedaliurn, intervenes between the tentacles and the bell. For figures
of the pedalia and the tentacles the works of Haake, Glaus, Conant
and Maas'22 may be consulted.

The Ectoderm. — The ectoderm of the tentacles is the seat of a
number of differentiations. It is quite thick, as the figures (28 and
29) show, and in this respect is very different from the pedalia, on
which the ectoderm cells are quite cubical. I found evidence of cilia
here and there, but I can add nothing definite about them. Neither
can I add any definite statements regarding the ectoderm cells proper,
but what I have to say relates to their differentiations.

(a) The thread cells are of two kinds, larger ones and smaller
ones. This is well shown in Fig. 29, which is part of a transverse
section of a tentacle of Tripedalia. Two kinds of nettle-cells are also
present in the tentacles of Charybdea, but they were specially well
shown in Tripedalia. The structure of these thread-cells seems to be
typical, and I have little more to say about them. I wish, however,
to call attention to the five or six unstriped muscle-fibers that are
attached to their basal lateral parts, and which connect them with
the basement membrane (Figs. 28, 29). Claus describes these muscle-
fibers and mentions that Fr. Miiller has described them before him,
but I have not found them mentioned elsewhere in the literature of
nettle-cells. Professor Brooks tells me, however, that he has often
found them. It would appear from Fig. 29 that they serve to retract
the thread-cells from the surface. Claus suggests that the muscles
are developed from the cnidoblasts.

(b) The plain subectodermal muscle-fibers are of interest. In

E. W. BERGER ON THE CUBOMEDUS.E. 75

Charybdea they lie wholly enclosed within canals of the supporting
lamella (Fig. 32, upper part). They run longitudinally, and near the
base of each tentacle pass out of their canals and become strictly
subectodermal (Figs. 31, 32). This is for Charybdea. In Tripedalia
they rarely come to lie in closed canals as in Charybdea. These
facts show beyond doubt that these muscles are developed from the
ectoderm. Glaus has suggested their ectodermal origin, but did not
demonstrate it. He also suggested that they become inclosed in
canals by the supporting lamella pushing up around them and finally
fusing above them. This, I believe, is demonstrated by the conditions
in Tripedalia (Fig. 29). Here the canals usually remain open, but
occasionally, as in the left-hand canal, one may become completely
inclosed. This condition of things suggests the intra-lamellar muscles
found in actiniarians. The nuclei found in the canals with the muscle-
fibers probably belong to the cells from which the muscles become
differentiated. Glaus figures these muscle-fibers and nuclei, and it may
be added that the supporting lamella he figures, for C. marsupialis,
is much thicker than I have figured it for C. Xaymacana and
Tripedalia cystophora. The number of muscle-canals also is greater
and occupies a much greater depth of the thickness of the lamella.
Since Glaus gives a figure of a transverse section showing the muscles
in their enclosed canals, I have not deemed it necessary to duplicate
his figure. In the transition from a tentacle to a pedalium, the
muscles are most strongly developed toward and at the edges of the
pedalium. This is true for the pedalia in general, and accounts for
the readiness with which they can be bent inwards, as noted in the
physiological part of this paper.

(c) I have found a single ganglion-cell among the cells of the
ectoderm of the tentacles. This showed so plainly that I have figured
it (Fig. 28). Other ganglion-cells no doubt exist, but could probably
not be distinguished from other cells. In its position in Fig. 28 it
appears to be associated with the nettle-cell shown just above it. Its
position is very much the same as that figured by Lendenfeld (25a).

The Endoderm.—The cells of the endoderm of a tentacle are
long and quite slender (Fig. 31). At their bases they are vacuolated
quite like the cells of the ampulla and the canal of the sensory
clubs. They contain a well-formed nucleus with a nucleolus. In
their distal half small light bodies with a dark center are very
evident. These bodies are evidently a secretion.

76 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

Another peculiar phenomenon presents itself in these cells. The
distal part of each cell becomes separated off from its body by what
appears to be the formation of a transverse cell-wall (Fig. 31, c-d). I
have found the ends of these cells quite separated off in some series.
The formation of the walls seems to begin as a thickening at the sides
of the cells, and a section through this region, transverse to the cells,
would appear like Fig. 30. The dots in the centers of the polygonal
areas of this figure are the centrad continuations of the cilia to be
described below. As already remarked in describing the endoderm
of the ampulla, I believe we here have another place of origin of
the "floating cells." The secretion just described moves into the
distal parts of the cells prior to their separation (Fig. 31). In some
series I could see these secretion bodies much more numerous within
the distal ends of the cells than in Fig. 31.

As will be seen in Fig. 31, each of the endoderm cells of the
tentacles has a flagellum that extends into the lumen of the tentacle.
Each flagellum has a thickening just within its cell, which may be
regarded as a basal body. From this basal body, again, a small fiber
extends centrad into each cell. It does not appear that the flagella
are thrown off with the distal parts of the cells ; at all events, I never
found them connected with any of the floating cells except in a few
doubtful instances.

What I have said for the endoderm of the tentacle of Charybdea
applies equally to Tripedalia.

Glaus, in his figure of a transverse section of a tentacle of C.
marsupialis shows the endoderm as cubical. I cannot explain why
there should be such a difference between the endoderm of the
tentacles of C. marsupialis and that of the tentacles of C. Xaymacana
and Tripedalia cystopliora. Glaus does not describe the endoderm in
detail.

The endoderm cells of the pedalia of both Charybdea and
Tripedalia are cubical and possess flagella, basal bodies, and centrad
continuations, quite like those I have described for the endoderm cells
of the ampulla. The double nature of the basal bodies and the centrad
continuations is, however, not so evident. A secretion I did not find.
Histologically, therefore, the endothelium of the pedalia corresponds
rather with that of the ampulla, and that of the tentacles with that
of the peduncle of the clubs.

E. W. BERGER ON THE CUBOMEDUS^E. 77

SUMMARY.

The most important results in the histological part of this paper
relate to the structure of the retinas of the eyes of the sensory clubs.

The retina of the distal complex eye is composed of three kinds
of cells : two kinds of sensory cells (the prism and pyramid cells),
and the long pigment cells (Figs. 1-9). The prism and pyramid cells
have each an axial nerve fiber in their prisms and pyramids respect-
ively. These fibers I could, however, trace only to the neighborhood
of the ^nuclei. But since I could trace similar fibers in the retinal
cells of the simple eyes (Fig. 16) past the nucleus into the subretinal
nerve tissue, I believe that the axial fibers in question also extend
centrad as nerve fibers into the subretinal nerve tissue. Other observers
also figure such fibers as extending centrad as nerve fibers. The axial
fibers of the prism cells have each a dumbbell-shaped basal body at
their entrance into the pigmented part of a cell. The evidence for a
body of such shape in the pyramid cells was not conclusive, though
a basal body for the axial fiber exists. The long pigment cells project
or retract their pigment in light or darkness respectively and thus
seem to serve to check the diffusion of light in the retina. I have
also supposed that these cells may serve for conducting impulses to
the lens, and that the latter is adjustable.

The proximal complex eye (Fig. 13) has only the prism cells
present in its retina, and not two kinds of cells as Schewiakoff has
described (see text, pp. 53, 60, 63) for all the eyes.

The simple eyes (Fig. 12), two on each side of a club, four in
all, also have only one kind of cells in their retinas, and each cell
has a flagellum extending into the vitreous secretion of the lumen.
These flagella could be traced centrad as a nerve fiber (Figs. 12, 16).
Similarly, a nerve fiber could be traced centrad from the flagella of
the epithelial cells of the clubs. Dumbbell-shaped basal bodies for
the flagella of the simple eyes could also be demonstrated, but the
evidence for this in the epithelial cells of the clubs was not so
satisfactory.

Other points of interest are: A secretory epithelium lining the
ampulla of the clubs, and a somewhat similar epithelium lining
the canals of the tentacles (Figs. 7, 27, 31) ; the partial origin of the
"floating bodies" in the canals of the clubs and tentacles and the
stomach pockets from these epithelia (Figs. 18, 19) ; two flagella to

78 JOHNS HOPKINS UNIVERSITY MOEPHOLOGICAL MONOGRAPHS.

each cell of the endothelium of the ampulla and of the pedalia (Figs.
7, 17) ; the peculiar nuclei in the endothelial cells of the ampulla
(Fig. 20) ; the longitudinal muscles of the tentacles being completely
inclosed within canals of the supporting lamella, but near the base
of a tentacle becoming subectodermal. This demonstrates their
ectodermal origin. In Tripedalia it is seldom that any of these
muscles become enclosed as in Chary bdea (Fig. 29).

If to the reader my results seem to embody a somewhat hetero-
geneous detail, he must remember that the work consists partly
in corroborating and partly in supplementing the work of previous
observers, and that, in general, histological detail does not usually
make the most readable paper.

BIOLOGICAL LABORATORY,

JOHNS HOPKINS UNIV., May, 1899.

LITERATURE.

LITERATURE REFERRED TO IN THE SECTION ON PHYSIOLOGY.

I. ROMANES, G. J. a. '75, '77. The Locomotor System of Medusae. Philosoph-
ical Transactions. London. Vol. CLXVI, pt. 1. Vol. CLXVII, pt. 2.

b. '85. Jelly-fish, Star-fish and Sea-urchins. London.

II. MURBACH, Louis. '95. Preliminary Notes on the Life-history of Gonione-
mus. Journal of Morphology. Vol. XI.

III. HESSE, R. '95. Uber das Nervensystem und die Sinnesorgane v. Rhiz-
ostoma Cuvieri. Zeit. Wis. Zool., B. LX.

IV. EIMER, TH. Zoologische Untersuchungen. '74. Wiirzburg Verhandlungen.
VI. Bd.

V. HAECKEL, E. '79. Monographic der Medusen. Jena.

VI. BERGER, E. W. '98. Abstract of Dr. F. S. Conant's Notes on the Physi-
ology of the Medusa?. Johns Hopkins University Circulars. Vol. XVIII, No. 137.

VII. (See also 8, below.)

LITERATURE REFERRED TO IN THE SECTION ON HISTOLOGY.

1. CARRIERE, J. '85. Die Schorgane der Thiere. Munchen u. Leipzig.

2. HAAKE, W. '87. Scyphomedusen des St. Vincent Golfes. Jen. Zeit. f.
Naturwis., Bd. XX., pp. 596-597, 602-604.

3. GLAUS, C. '78. Uber Charybdea marsupialis. Arb. aus dem Zool., Inst.
Univers. Wien., Bd. I.

4. SCHEWIAKOFF, W. '89. Beitrage zur Kenntniss des Acalephenauges. Morph.
Jahrb., Bd. XV, H. 1.

5. PATTEN, WILLIAM, a. '89. Studies on the eyes of Arthropods. II. Eyes of
Acilius. Journal of Morphology. Vol.11.

b. '98. A Basis for a Theory of Color Vision. American Naturalist. Vol.
XXXII, No. 383.

E. W. BERGER ON THE CUBOMEDUS.E. 79

6. APATHY, ST. ''J7. Das Leitende Element cles Nervensystems u. seine topo-
graphischen Beziehungen zu den Zellen. Mitt. Zool. Stat. Neapel., Bd. XII, H. 4.

7. PARKER, G. H. '97. Photomechanical Changes in the Retinal Pigment Cells
of Pala?monites, and their Relation to the Central Nervous System. Bull. Mus.
Comp. Zool. Harvard Coll. Vol. XXX, No. 6.

8. CONANT, F. S. a. '97. Notes on the Cubomedusa?. Johns Hopkins Univer-
sity Circulars. Vol. XVII, No. 132.

b. '98. The Cubomedusa?. Memoirs Biological Laboratory Johns Hopkins
Univ. Vol. IV, No. 1.

9. A REVIEW OF 5b. '99. A Theory of Color Vision. Natural Science. Vol.
XIV, No. 85.

10. HERRICK, F. H. '91. The Embryology and Metamorphosis of the Macroura
(Brooks and Herrick). Natl. Acad. Sciences. Vol. V, p. 454.

11. HERTWIG, O. & R. '78. Das Nervensystem und die Sinnesorgane der
Medusen. Leipzig.

12. SCHREINER, K. E. a. '96. Die Augen bei Pecten und Lima. Bergens
Museums Aarbog.

b. '97. Histologische Studien iiber die Augen der freilebenden marinen
Borstenwiirmer. Bergens Museums Aarbog.

13. HESSE, R. '99. Untersuchungen iiber die Organe der Lichtempfindung bei
niederen Thieren. V. Die Augen der Polychaten Anneliden. Zeit. Wis. Zool.,
B. LXV, H. 3.

14. ANDREWS, E. A. '92. On the Eyes of Polychsetous Annelids. Journal of
Morphology. Vol. VII.

15. WILSON, H. V. '78. Unpublished Notes.

16. GRENACHER, H. '84. Abhandlungen zur vergleichenden Anatomie des
Auges. I. Die Retine der Cephalopoden. Abhandl. der Naturf. Gesellsch. zu Halle.
Bd. XVI.

17. BEER, THEODORE. '98. Die Accomodation des Auges in der Thierreihe.
Wiener klinische Wochenschrift. Nr. 42.

18. WILSON, E. B. '96. The Cell.

19. BERGER, E. W. '98. The Histological Structure of the Eyes of Cubomedusa?.
The Journal of Comp. Neurology. Vol. VIII, No. 3.

20. LENDENFELD, R. Die Nesselzellen der Chidarier. (Review and bibliography.)
Biol. Centralbl. Bd. XVII, Nr. 13.

21. SCHNEIDER, K. '90. Histologie von Hydra fusca mit besonderer Beriick-
sichtigung des Nervensystems der Hydropolypen. Arch. Mik. Anat. Vol. XXXV.

22. MAAS, O. '98. Die Medusen. (Charybdea arborifera, Systematic.) Mem.
Mus. Comp. Zool., Harvard Coll. Vol. XXIII, No. 1.

LITERATURE RFERRING TO THE CENTRAD CONTINUATIONS OF CILIA AND FLAGELLA.

A. HAECKEL, E. '72. Die Kalkschwamme. Vol. I, p. 141; Vol. Ill, PI. 25,
Figs. 3-5.

B. SCHULTZE, F. E. '75. Rhizopodien Studien. V. Arch. Mik. Anat. Bd. II, p. 583.

;

80 JOHNS HOPKINS UNIVEESITY MORPHOLOGICAL MONOGRAPHS.

C. EIMER, TH. '77. Weitere Nachrichten iiber d. Bau des Zellkerns, nebst
Bemerkungen liber Wimperepithelien. Arch. f. Mik. Anat. Bd. XIV, Taf. VII, p. 114.

D. BUTSCHLI, O. '78. Beitrage zur Kenntniss der Flagellaten, u. s. w. Zeit.
f. Wis. Zool. Bd. XXX, p. 269.

E. ENGELMANN, TH. W. '80. Zur Anatomie u. Physiologie d. Flimmerzellen
Pfliiger's Arch. Bd. XXIII.

F. HATSCHEK, B. '85. Entwickelung der Trochophora von Eupomatus unci-
natus. Arb. Zool. Inst. Wien., Bd. VI, p. 139.

G. HEIDER, K. '86. Zur Metamorphose der Oscarella lobularis. Arb. Zool.
Inst. Wien., Bd.VI, pp. 189-194.

H. SCHNEIDER, K. C. '92. Einige histologisehe Befunde an Coelenterata. Jen.
Zeit. f. Nat. 27, N. F. 20.

I. HECHT, EMILE. '95. Contribution a 1'Etude des Nudibranchs Memoirs de
la Societe Zool. de France. T. 8, PI. IV, Fig. 45.

J. MINCHIN, E. A. '96. Notes on the Larva and Postlarval Development of
Leucolosolemia variabalis, etc. Proc. R. Soc., London. Vol. LX.

K. HENNEGUY, L. F. '98. Sur le rapports des ciles vibrales avec les centro-
somes. Arch, d'anat. micros., T. 1.

L. LENHOSSEK, H. '98. tiber Flimmerzellen. Anat. Anz. (Supplement.) Bd.
XIV.

M. PETRE, CARL. '99. Das Centrum fur die Flimmer u. Geisel-bewegung.
Anat. Anz. Bd. XV, Nos. 14 and 15.

N. See also 6.

REFERENCE LETTERS.

a=flagellum in Fig. 27, that is sup- fpyr=axial nerve fiber of a pyramid cell,

posed to extend centrad beyond frc=axial nerve fiber of the retinal

the nucleus. cells of the simple eyes.

b=twin flagella in Fig. 27, of which gc=ganglion cells.

the centrad continuation is seen ind=impression of the lens probably

applied against the distal sur- due to the pressure of weight

face of the cells and to be con- against the surrounding tissue,

tinued centrad. l=lens.

c=:capsule of lens. lp=long pigment cells.

cf=axial fibers of cells extending m=muscle fibers.

centrad. namp=nuclei of ampulla cells.

co=cornea. nc=network cells (Figs. 13 and 16),

concr— concretion cavity. and nettle cells (Figs. 28, 29).

ec— ectoderm. nf =nerve fibers and tissue.

en=endoderm. nlp=nucleus of long pigment cell.

f =flagella. nm=nucleus of muscle cells.

flp = distal fiber of a long pigment cell. nprc=nucleus of prism cell.

fpr=axial nerve fiber of a prism cell. npyrc=nucleus of pyramid cell

E. W. BERGER ON THE CUBOMEDUS^E. 81

nz=nuclear zone. sec=vitreous secretion in the lumen of

pr= prism of prism cell. the simple eyes.

prc=prism cell. sla=supporting lamella.

vb = vitreous body or zone.

pyr=pyramid of pyramid cell. J . . . ,.,

x=(l) the approximate level at which
pyre = pyramid cell. Fig 4 should be cut transversely

pz=pigmented zone. to give Figs. 1 and 3.

rr=retina. (2) the thickening of the support-

s=secretion in endo. of tent, and ing lamella in Fig. 13 to support

ampulla. the lens-

, . , *;=Point of approximation of cells of

sh^shrmkage space. . £,.

lenses in Figs. 7 and 13.

DESCRIPTION OF FIGURES.

ALL. FIGURES, UNLESS OTHERWISE STATED, ARE FROM CHARYBDEA.

Fig. 1. This figure represents a transverse section through a portion of the
vitreous body of the distal complex eye at about the level x of Fig. 4. Three kinds
of areas are seen, namely, the prisms and pyramids with their axial fibers and the
distal continuations of the long pigment cells. Towards the lower left of the figure
the section is a little more distal than at the right and the transverse areas of the long
pigment cells are no more so large as at the right of the figure. The dark granules in
the areas of the long pigment cells represent pigment. Camera lucida sketch. X920.
pp. 45, 46, 48, 49, 50, 51, 52, 54.

Fig. 2. This figure is a camera lucida sketch from a section taken transverse
through the most distal part of the pigmented zone of a slightly pigmented retina of
a distal complex eye. The presence of three kinds of elements is again evident. The
dots without the polygonal areas represent the centrad continuations of the axial fibers
of the prism cells. The lettering explains the other areas, X920. pp. 46, 48, 50.

Fig. 3. This is from a section similar to that of Fig. 1, but a little more distal.
At the right, the section is more distal than at the left of the figure, in consequence
of which the long pigment cells are there taken through their distal fibers. Note the
small shrinkage spaces about the axial fibers of the prisms. The white lines bounding
the prism areas appear as in nature. The pyramid cells are not shown in this figure.
X950. Camera sketch, pp. 50, 51, 52, 54.

Fig. 4. This figure is from a section taken parallel to the long axis of the cells
of the retina of a distal complex eye. It is from a camera sketch, and nothing has
been put into the figure except what could be clearly seen. The lateral boundary
lines of the prisms are not shown. Note the evidence for the existence of three kinds
of cells. X920. pp. 44-52, 54.

Fig. 5. This figure represents a sagittal section through the nuclear and
pigmented zones and the subretinal nerve tissue of a slightly pigmented retina of a
distal complex eye, that had been killed in the dark. Camera sketch. The pyramid
cells are not shown. X900. pp. 47, 51, 52, 53.

Fig. 6. These cells are from a preparation by Conant of a sensory club, macer-
7

82 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

ated in acetic acid. Cell a is evidently an iris cell. The others are probably prism cells
from the proximal complex eye. X900. pp. 44, 48.

Fig. 7. In this figure I represent a sagittal section through the distal complex
eye. In the middle half of the section, the nuclei, the prism and pyramid cells with
their axial fibers, and the long pigment cells with their large distal fibers are all
strictly camera lucida sketched. A portion of the pigmented zone has been left
unpigmented to better show its structure. At the right and above the concretion
cavity is shown a portion of the endoderm of the ampulla. The section is not strictly
in a dorsoventral plane of the club, in consequence of which the cells of the ampulla
are cut diagonally and through their tips. Note the dumbbell-shaped nuclei of the
ampulla cells, as also the masses of secretion. A part of the retina of the proxi-
mal complex eye is shown in the upper part of the figure. X 920. pp. 41-54, 63, 64,
68-71.

Fig. 8. These cells are from a macerated preparation. Cells a, b, c, d may be
either prism or pyramid cells from the distal complex eye or prism cells from the
proximal complex eye. Cells e and f are probably from the right fourth (Fig. 13) of
the retina of the proximal complex eye or from the simple eyes. The unlettered
cells are probably from the simple eyes. Some of these show a distal process. X900.
pp. 48, 62, 65.

Fig. 9. The cells here figured are long pigment cells from the same preparation
as Fig. 6. X900. pp. 50, 51.

Fig. 10. This drawing shows an end view of a group of prisms from the same
preparation as Fig. 6. X900. pp. 46.

Fig. 11. This group of prisms are from the same preparation as Fig. 6. Two of
them are broken off. The fibers seen at the lower end are probably some of the axial
fibers. The fiber at the upper end 1 believe is interprismatic and the distal fiber of a
long pigment cell. X900. pp.46.

Fig. 12. This figure is a summary of my results on the simple eyes. It is from a
camera sketch of one of the distal eyes, but somewhat diagrammatic. The left side of
the figure is proximal, the right side distal. x920. pp. 61, 62, 64, 65.

Fig. 13. Sagittal dorsoventral section of a proximal complex eye. Conant drew
and published this as his Fig. 69. Conant's evidence regarding the axial fibers of the
prism cells was incomplete ; so that, in this respect, he left his figure unfinished. I
have drawn in these fibers and republish the figure. At the right of the retina and
next the lens (the white space) the vitreous body is incomplete and the fibers from the
retinal cells project freely into the space. This part of the retina also remains
unpigmented. Like my Fig. 7, this figure evidently represents a section somewhat to
one side of a sagittal dorsoventral plane of the club, so that the endoderm cells of
the ampulla are cut diagonally or transversely, pp. 41-44, 60, 64-68.

Fig. 14. This is drawn to show how regularly small shrinkage spaces may occur
in transverse sections of the vitreous bodies. This figure is from a transverse section
of the vitreous body of a proximal complex eye. I believe that these spaces are deter-
mined by the axial fibers of the prisms. Prism outlines are not shown. X950.
pp. 54.

E. W. BERGER ON THE CUBOMEDUS>E. 83

Fig. 15. This figure is a drawing of a portion of a transverse section of one of
the simple eyes. Note the flagella from the retinal cells, pp. 62.

Fig. 16. The section of the lower left hand corner of this figure is through a por-
tion of one of the proximal complex eyes, and shows the centrad continuation of the
axial nerve fibers of the retinal cells. The section is such, that, besides the simple
eye, the nuclei of the proximal complex eye (upper part of figure) and two network
cells are cut. X920. pp. 47, 62, 63.

Fig. 17. A transverse section through the tips of the ampulla cells is here shown.
To the left is towards the upper end of the ampulla. The basal bodies with the cen-
trad fibers are in the plane of the section, while the flagella are supposed to extend
below the plane of the section. XI 350. pp. 71.

Fig. 18. These bodies, from within the ampulla cells, contain some of the secre-
tion of the ampulla cells, and resemble the " floating bodies." X1350. pp. 72.

Fig. 19. The " floating bodies " here represented are from the ampulla. Globules
of a secretion similar to that found in the ampulla cells are seen both within and
without the bodies. Note also the two black bodies without the cells and two or
three similar ones within the cells. These latter bodies are of doubtful nature.
X1320. pp.72.

Fig. 20. This figure represents sections of the various nuclei found within the
ampulla cells. X1350. pp.69, 70.

Fig. 21. These cells are from the same preparation as Fig. G. They are
evidently retinal cells from the simple eyes. The tendency of their pigmented
ends to become globular, I believe, is due to their having become isolated before
they hardened during maceration. X920. pp. 62.

Fig. 22. This diagram illustrates the retraction of the long pigment cells.
The dotted lines in the vitreous body mark the outlines of the prisms, while
the continuous lines represent the axial fibers of the prism and pyramid cells-
pp. 45, 46, 48, 49, 53.

Fig. 23. These cells are from the epithelium of a sensory club. They are
from the same preparation as Fig. 6. Flagella are not shown. X900. pp. 64.

Fig. 24. This group of epithelial cells of a club are from the same prepara-
tion as Fig. 6. X850. p. 64.

Fig. 25. This sketch is a transverse section through the tips of the epithe-
lial cells of a club. The polygonal areas are the cells, while the central dots
are the centrad continuations (nerve fibers) of the flagella of the cells. X920.
pp. 63, 65, 66.

Fig. 26. The flagella of the epithelium of a club are in this figure seen to extend
centrad, some beyond the nuclei. Cell outlines are not shown. X920. pp. 64, 65, 66.
Fig. 27. The cells of the lower half of this figure belong to the ampulla, those of
the upper half to the canal of the peduncle. The right side of the figure is towards the
eyes (the ventral side) of the club. Globules of secretion are seen within the
ampulla cells, as also a globule without. The ring above the latter globule is
probably an empty shell of a floating cell. X1320. pp. 68, 69, 71, 73.

Fig. 28. This figure is from a transverse section of a tentacle of Charybdea.

84 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS.

The mass with darkly stained granules is the remains of a thread cell. The
ectoderm and a small part of the supporting lamella only are figured. Note the
large ganglion cell. X920. pp. 74, 75.

Fig. 29. Part of a transverse section of a tentacle of Tripedalia. The endoderm
is not figured. The supporting lamella is seen to be considerably thinner than in
Charybdea. Note the subectodermal muscles, as also the muscle fibers to the thread
cells. X920. pp. 69, 74, 75.

Fig. 30. This is a transverse section through the endothelium of a tentacle of
Charybdea in the line c d of Fig. 32. The dark lines bounding the polygonal areas
are the thickenings of the sides of the walls of the cells in the line indicated. The
central dots are the centrad continuations of the flagella. X920. p. 76.

Fig. 31. This figure is a transverse section through a tentacle of Charybdea at
about the middle of Fig. 32, i. e. so near to where the tentacle joins the pedalium,
that the muscles within the lamella have all come to lie under the ectoderm. The
ectoderm is not shown. X920. pp. 75, 70.

Fig. 32 A longitudinal section through the supporting lamella only, of a
tentacle of Charybdea, is here shown. In the upper part of the figure the muscle
fibers are seen wholly enclosed by the supporting lamella. In the middle of the figure
they are seen to pass out of their canal. In the lower part of the figure, the support-
ing lamella is seen to bend to the right where it becomes continuous with the lamella
of the pedalium. X920. p. 75.

BERGER.-CUBOMEDUSAE.

ec.

MffiSG

— f AasS* We* "•/'

BERGER.-CUBOMEDUSAE

PLATE

/re

fr.c.

HgJi

RgJS.

® S"®

fig.20.

Fin.'.".'.

BERGER.-CUBOMEDUSAE.

PLATE

KffZS.

fr.

n.i-

III'.

s In.

VITA.

The writer of the foregoing dissertation, Edward William Berger,
was born in Middleburg Township, near Berea, Ohio, on November
29, 1869. He is the youngest son born to Karl Gottlob Berger and
Christiane Pauline, daughter of Karl Gottlieb Gellrich, who, leaving
their home in Silesia, of southeastern Prussia, in 1854, came to make
America their adopted country.

The writer's preliminary education was received in the public
schools of Lorain, Ohio. In 1886 he entered German Wallace College,
of Berea, Ohio, from which institution he graduated in 1891. In
1892 he entered Baldwin University, also at Berea, and graduated
in 1894. In the fall of the same year he entered Johns Hopkins
University, taking up special studies in Zoology, Physiology, Botany
and Chemistry.

Two months of the summer of 1897 were spent in Jamaica with
the Johns Hopkins Marine Laboratory, and two months, in 1898, at
the United States Fish Commission Laboratory, Wood's Holl, Mass.
The writer has twice held a University Scholarship.

BIOLOGICAL LABORATORY,

JOHNS HOPKINS UNIVERSITY, May, 1899.

Memoirs from % Biological Caboratoij)

OP THE

JOHNS HOPKINS UNIVERSITY

IV, 5

WILLIAM K. BROOKS, EDITOR

OPHIURA BREVISPINA

A DISSERTATION PRESENTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

THE JOHNS HOPKINS UNIVERSITY

BY

CASWELL GRAVE

(Reprinted from the Memoirs of the National Academy of Sciences)

BALTIMORE
THE JOHNS HOPKINS PRESS

1900

As my name appears upon the title-page of this memoir, it is
proper for me to state that my share in the work has been that
of the instructor under whose direction the work has been done.
The discovery that this Ophiuran is of peculiar interest and that
it is unusually favorable for the study of the problems of the
morphology of Echinoderms, was made by Dr. Grave ; and the
results which are here detailed are his work.

W. K. BROOKS.

CONTENTS.

Page.

1. Introduction

83

2. Historical sketch

84

3. Distribution and habits

84

4. Physiological notes

85

5. Early stages

87

6. Stage "A." Origin of tho anterior enterocreles

87

7. Stage " B." Origin of the livdrocu-Ie

88

8. Stage "C." Closing of the blastopore and formation of the mouth

89

9. Stage "D." Kotation of the hydroca'le completed

9G

10. Stage "E." Second pair tentacles formed

92

11. Stage "F." Invagiuatiou of the nervous system

94

12. Stage "G." Degeneration in the larval organ begun

95

13- Stage "H." Formation of subnenral sinns

96

14. Kelation of tho larva to adult

96

15. Larva of Antedon rosacea and Ophiura brevispiua compared

98

l(i. Literature cited

98

17. Description of figures

99

81

OPHIURA BREVISPINA.

By W. K. BROOKS aiid CASWBLL GRAVE.

INTRODUCTION.

During the summer of 1808 it was my privilege to occupy the table of the Johns Hopkins
University in the United States Fish Commission laboratory at Woods Hole, and while here I
rediscovered the peculiar Ophiuran larva, which was first found and figured by KROHN (7).

Finding the lame he described in the open sea KROHN did not know to what species they
belonged; but the larva1, the development of which is the subject of the greater part of this paper,
came from eggs laid in aquaria by Ophinra brerispiua. It is not likely that the same species of
Ophiuran occurs both at Funchal, where KROHN did his work, and also at North Falmouth, where
my material was obtained, but it is very probable that species belonging to the genus Ophiura
have similar larval forms.

Among Ecliinoderms, where a direct development from the larva to adult occurs, that is,
without the usual highly specialized intermediate pelagic larva, we usually have to do witli a
species which in some manner takes care of its brood; but in O. breoispina the larva; are free
swimming, they being provided with a well developed loeomotor apparatus, yet the usual Ophiurid
pluteus larva is as completely omitted as it is from the life history of the viviparous Amphlura
squamata.

From the fact that the usual pluteus skeleton is begun in the larvse of 0. brevispina one is led
to suspect, however, that at some period in its history the species possessed a larva more nearly
like a pluteus than at the present time. On the other hand, on account of the resemblances which
exist between the larva} of 0. brevispina and Antedon rosacea (treated of in another place) we may
suppose a close phylogeuetic relationship exists between them. If, as many zoologists believe,
the criuoids have retained more nearly thau any other group the characters of the primitive
Echinoderm stock, then in the larva of 0. brevispina we may have one which has retained
unmodified its primitive characteristics.

In this paper, however, the facts only of development are taken up, and the question of the
bearing which this larva may have on any theoretical discussion concerning the interrelationships
of the Ecliinoderms is suggested here in order that the reader may keep the subject before him
while studying the paper. The points of resemblance between the Ophiuran and Antedon larvre
are enumerated in a chapter further on.

The method used in the preparation of the material for microscopical study, and which gave
good results, is as follows : The larvie were taken up into a pipet with as little water as possible,
and squirted into a small bottle containing a solution of sublimate acetic (98 parts of a sat. sol.
HgCl2 being used to 2 parts of glacial acetic acid). After from two to live minutes the sublimate
solution was drawn off gently, leaving the larv;v at the bottom where they had settled. Then 50
per cent alcohol was added, which in a few minutes (5) was drawn off and replaced by 70 per cent
alcohol, in which a little iodine had been dissolved. In a few hours (3-12) this was changed for
clear 75 per cent alcohol, in which the larva- remained until needed for laboratory study. After
staining lightly in acid carmine, so as to facilitate their orientation, the larva; were dehydrated in
the usual way and cleared in oil of cloves. From the clove oil they were oriented by a modifica-
tion of the PATTON method. After an impregnation with 55° paraffin, series of sections three

83

84 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

microns in thickness were made in three planes, transverse, longitudinal sagittal, and longi-
tudinal horizontal. The sections were stained on the slide with KLEINENBERG'S htematoxylon.
Other methods were tried, but none proved so satisfactory as the one just described. The
shrinkage in echinoderm tissue, which usually accompanies the unmodified paraffin method, was
not to be seen in the tissue of these larvae, due, no doubt, to its unusual thickness.

It has been thought best to make the following list of terms which are used synonymously in
the text of this paper in the description of the larvie. Those in the same line can, in most cases,
be interchangeably used.

Dorsal-aboral-above-over.
Ventral-oral-below-under.
Auterior-forward-before.
Posterior-back ward-behind.

In the drawings of the larva-, when the ventral side is up and the anterior end is nearest the
top of the page, then the reader's left is also left in the figure.

For convenience in description, the various stages taken to illustrate the life history of the
species have been designated by letters of the alphabet, this method seeming preferable to one in
which age is used as a distinguishing character, since the progress of development at any age
depends so intimately on the varying conditions of environment.

I take this opportunity to acknowledge my indebtedness to Dr. C. P. SIGEKFOOS, at whose
suggestion I began the study of Ophinran development.

I was aided very materially while at the Fish Commission laboratory by Prof. H. C. BUMPUS,
who placed at my disposal every facility for work at his command, and to him, also, I am greatly
indebted for many suggestions in methods of rearing larv;i- at the seashore.

To Professor BROOKS, under whose direction my work has been done, are due my warmest
thanks for the interest with which he has followed me in my studies and for the many valuable
suggestions he has offered from time to time during the year.

HISTORICAL SKETCH.

The species of Ophiuran, Ophiura brevispina, the life history of which is the subject of this
dissertation, was first discovered and described by Thomas Say in 1825 (12).

Since this time the species has been rediscovered and renamed as many as three times. It is
probably best known at present by one of its synonyms, Opltiura olivaceu, which was given to it
in 1865 by THEODORE LYMAN (8). In his earlier works LVMAN distinguished between 0. ollvacea
and O.brevispina, but in his Challenger report on the Opliiuriihc and Astrophytidw (9) he places
the two species together as one under its earlier name, -which, although less descriptive of the
species than that given by Lyman, it is probably best to retain.

In 1852 ATERS described the species under the name Opliioderma olipacenm in Vol. IV of the
Proc. Bost. Soc. Nat. Hist.

LUTKEN also described it as OpModerma scrpens in 1856.

DISTRIBUTION AND HABITS.

Ophiura brcvispina is a very widely distributed species, it having been reported from points
along the Atlantic coast from Brazil to New England.
It has been taken from the following localities:

1. Bahia, Brazil. 6. Beaufort, North Carolina.

2. Port Antonio, Jamaica. 7. Old Point Comfort, Virginia.

3. St. Thomas, Bahamas. 8. Sag Harbor, New York.

4. Cape Florida, Florida. 9. Dartmouth, Massachusetts.

5. Tortugas. 10. New Bedford, Massachusetts.

11. North Falmouth, Massachusetts.

That part of North Falmouth Harbor which is inhabited by the species is very shallow, its
depth at low tide not exceeding 1 fathom.

The bottom is covered with a mat of living and dead grasses and alg;e, and in this tangle
the ophiurans live, together with a great variety of crustaceans, molusks, and worms.

MEMOIKS OF THE NATIONAL ACADEMY' OK SCI KNOKS. 85

Tlio usual color of the species is an olive green, with darker bands on the anus and sometimes
with :i clouded disk.

Through the blending of their colors with the seaweed the ophiurans are greatly protected
from their enemies, and it is difficult, even when looking for them, to see them among the seaweed
so long as they do not move.

It is quite common to lind a small Amphipod crustacean clinging to the arms of dredged
specimens, and from the structure of the crustacean it is probable that the two species live
together commensally. What benefit either animal can derive from the association it is difficult
to see.

One pair of the thoracic legs of the crustacean is so modified as to form a structure beauti-
fully adapted for clinging to the round ophiuran arms. The last segment but one of each of this
pair of legs is Y shaped. At the cud of one arm of the Y is attached a movable segment, the end
segment of the leg, which when shut down upon the end of the other arm of the Y incloses a
triangular space in which the ophiurau arm is held.

The body of the crustacean is colored and banded in such a manner as to simulate closely the
color and banding of the ophiuran arms.

When placed in aquaria with their host, the crustaceans cling to the ophiuran arms until the
water becomes depleted of oxygen, when they leave the arms and swim about the edge of the dish
apparently much alarmed.

In examining the stomachs of the ophiurans one finds bits of other animals, such as crustacean
appendages and the skeletons of young horseshoe crabs. From this it is probable that the crea-
tures are scavengers, since an active crustacean would hardly be captured by so slow and poorly
armed an animal as an ophiuran. None were ever observed to eat anything when kept in the
laboratory, and it is quite out of the question to observe them in their natural habitat, since they
are nocturnal animals remaining hidden during the day.

The ophiurans were first examined for sexual elements early in June, and at that time the
eggs were very large but adhered closely together in the gouads. The sperm appeared to be fully
formed but were iionmotile.

From this time on until the middle of August the species was regularly watched and exam-
ined, and on July 16 the first ripe eggs and sperm were obtained. A great number of specimens
had that day been dredged and placed in aquaria dishes of fresh, filtered sea water. One week
later a great number of adults were again brought in and placed under the same conditions as
those which had spawned in the laboratory the week before, but this time very few eggs were
obtained, and all subsequent attempts to get the ophiurans to spawn were unsuccessful.

From this it would seem that the breeding season is extremely short.1

The time of day at which spawning occurred corresponds well with the time at which 1 have
noted it to take place in Ophiopliilus aculeata and Ophiocoma echinatn, that is, between 8 and 10
o'clock p. in.

PHYSIOLOGICAL NOTES.

The locomotor movements of an ophiurid, upon a casual observation, seem to consist of an
uncoordinated writhing and twisting not calculated to bring the creature to food or a place of
safety except by chance; but a more careful study shows them to be the result of an orderly and
nicely coordinated mechanism.

The rapid strides which characterize the movements of a brittle star are in strong contrast

1 During the summer of 18S19, after this paper bad gone to press, my experience with the species was very dif-
ferent from the above. Specimens brought into the laboratory early in June threw eggs and sperm, but the eggs,
after passing through the early segmentation stages, ceased to develop. The eggs were probably immature, and
were spawned only because of the bad condition of the water in the aquaria, but spawning always occurred early
in the evening at the tinn1 when it would have occurred under normal conditions. Why unripe eggs should develop
at all, or why eggs mature enough to begin their development should not be mature enough to complete it, ia ail
interesting question.

This phenomenon was repeated every few days until July l'f>, when about one-fourth the number of eggs
spawned developed into normal larva-. This is ten days Inli-r than tlie date when eggs became mature at Woods
Holl. From the fact that the water is much warmer at Beaufort than at Woods Holl om; would expect to tiud the
spawning season earlier at the latter place.
10396 li

86

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

with the slow creeping movements of a starfish or sea-urchin, the difference being due to the
employment of different locomotor mechanisms in the two cases; the starfish and sea-urchin
depending entirely upon their tube feet and spines while in the ophiurids, the arms themselves
are the efficient locomotor organs, they being used much as we use our arms in swimming.

The arm of an ophiurid consists of a large number of segments, each of which contains a
central calcarious ossicle. The calcarious ossicles of adjacent segments articulate with each other
like the vertebne of the spinal column, and are joined together by two pairs of muscles in such
a manner that motion is possible in all directions. This mechanism is aided in producing the
locomotion of the creature not only by the arm. spines, where they are present, but by the foot
tentacles. These latter organs, which are the homologues of the tube feet of other echinoderms,
have been previously regarded as having given up their locomotor function entirely, but 1 shall
show further on that this is not true in the genus Ophiura.

The experiments I carried on last summer on the movements of ophiurans resulted in little
that is new, but on account of the confirmation my notes and photographs give to PREYKRS' work
(11) on the same subject, it has been thought advisable to publish them.

In the usual method of progression one arm precedes, it taking no other part, apparently, than
to point out the way; the two arms adjacent to and behind the anterior arm make the stroke; the
remaining arms are dragged behind, acting as a rudder.

FIG. 1.

FIG. -j.

FIG. 4.

FIG. 5.

No preference as to which arm should precede could be found in an adult ophiurau, each arm
being equally capable of going before, making the stroke, or following behind.

If greater speed is needed, for example, to get away from a strong stimulus, the arm which
precedes may also take part in the stroke, its contractions being made simultaneously with those
of the side arms. This added force, if produced repeatedly on one side, would soon change the
course of progression, but this-difficnlty is overcome by an alternation of the stroke of the preced-
ing arm, first on one side, then on the other (text tig. 2).

In a third method of normal locomotion the arms are arranged as is seen in text fig. 5, in
which only one arm follows, acting as the rudder. This leaves two pairs of arms for the stroke,
but the anterior pair is usually most vigorous in its contractions.

Since no physiological differentiation into anterior, posterior, or lateral parts is to be found in
ophiurids, the creatures are under no necessity of turning the body when a change in the direction
of progression is to be made. The arm which finds itself pointing in the new direction to be
traveled takes the lead, although it may have been either lateral or posterior in position in the
previous movements.

As has been mentioned before, the foot tentacles aid in making the strokes of the arms efficient
in propelling the body. After a stroke has been made, while the arms are being drawn forward
and extended for a new stroke, the tentacles can be seen moving actively about, but as the arms
come to rest for the backward movement the tentacles are thrust down against the substratum
and cease to move. The tentacles thus fit themselves into the inequalities of the surface and
afford fixed points for the arms to pull against. The tentacles of the posterior arms act in the
same way, and are efficient in preventing the force of the stroke being lost in side motion.

In ophiurans with long arm spines these latter structures may perform the function just
described for the foot tentacles, but in the genus Ophiura the arm spines are very minute and
closely applied to the sides of the arms.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 87

It is interesting to note the wonderful coordination of locomotor movements immediately
following tlie amputation of three of the arms. In tin's case, il the nerve ring has been uninjured
one of the remaining arms takes up the part of guiding and balancing, while the other strokes
first on one side then on the other (text h'gs. 3 and 4).

When the central nerve ring is cut at any point the coordination in movement is impaired, and
when cut in five places, between the arms, it is lost entirely.

When placed on its aboral surface an ophiuran quickly turns over. The method used is quite
definite; two adjacent arms straighten out so that together they form a straight line. On these
arms as an axis the body revolves, being pushed over by the .three remaining arms, but mostly by
the median one of the three.

EARLY STAGES.

The mature eggs are opaque and vary in color from an olive green to an orange yellow.
Those of the same individual, however, are constant in their coloration. Until quite well
developed the larva1 retain the color which was on the eggs at the time they were laid.

For echinoderms the eggs are very large, being 0.3 millimeter in diameter.

Soon after they are fertilized the eggs throw off two membranes, the first of which is much
thicker than the second.

When first laid and during their early development the eggs float, but when their cilia are
formed the larva1 are able to swim below the surface.

As I did not know that any special interest would be found in the life history of the species,
I did not carefully observe the early stages while living, nor preserve material for future study,
and as I have stated elsewhere, all later attempts to get other material were unsuccessful.

This makes it necessary to begin this paper with the description of a late gastrula in which
the first pair of enterocosles have already begun to form as lateral pouches from the anterior free
end of the archeuteron (figs. 1-3).

Larvie in this stage of development will be designated as "A."

STAOE "A," 36 HOURS OLD.
(Figures 1, 2, and 3.)

At the age of 30 hours the larva* swim actively, they being uniformly covered with cilia

(fig. 1).

The shape of the larva1, is an oval, the length being to the shorter diameter as 2 is to 1.

The animal or anterior polo is slightly more pointed than the posterior vegetative one. The
ventral surface is distinguished by the presence of the blastopore, which latter has been pushed
from its posterior position to a ventral one by the rapid growth of the ectoderm of the dorsal
surface of the larva.

An apical plate of taller cells is present at the anterior end, but I could not see that the cilia
at this point were any longer than those which cover the other parts of the larva (fig. 3, ap).

From the blastopore, through which its cavity opens to the exterior, a large archenteron
projects forward into the blastociele.

The remainder of the blastocwle, not taken up by the archenterou or its pouches, is filled with
a close network of mesenchyme cells. This meseuchyme tissue is shown in fig. 3, mes, which is a
longitudinal sagittal section of "A."

From the anterior free end of the archenteron a large pouch is in process of being cut otf.
This pouch projects to the right and left as horn-like processes, which latter are to be considered
the rudiments of the right and left anterior euterocceles (fig. 2, aer and ael).

As to the method of gastrulation I can not at present speak from observation on larvae in
which it is just taking place, but from a study of the stage now under consideration some idea can
be gotten as to how it has proceeded. In figs. 1 and 3 we see a cellular plug (cp) protruding from
the blastopore and also extending far into the archenteric cavity. In some cases it extends even
into the enterocu-le pouch. The contour of this cellular mass is ragged, which is also true of both
the outer and inner surfaces of the wall of the archeuteron and the inner surface of the ectoderm.

These facts seem to indicate that gastrulation does not take place by invaginatiou, as is usual
in echinoderms, but that the larva before gastrulation is a solid, plauula-like affair, aud later the

88 MEMOIRS OF THE NATIONAL ACADEM? OP SCIENCES.

archenteron is formed by a splitting away of the central core. In the same way the plug of cells
is probably formed by the hollowing out of the solid archenteron.

Beside their ragged outline the walls of the larva have another peculiarity in their structure,
for, judging by the number and position of the nuclei, they are from two to three cells iu thickness
(tig. 3).

Cell walls are not distinguishable iu any stage of development.

STAGE " B," 42 HOURS OLD.
(Figures 4 and 5.)

According to BURY (2) the hydrowele does not have the same origin in all the groups of
echinoderms. lie found that it originates in the crinoids, sea-urchins, and starfishes from
the left anterior enteroccele, but iu the ophiurids it grows out from the anterior end of the left
posterior enteroccele.

This observation, which BURY records with apparent hesitation, I can completely confirm, as
will be seen in the description and figures of l'B."

Externally the appearance is the same as in "A," but the internal structures have undergone
a great change.

The anterior pouches, the cavities of which in "A" were connected both with each other and
with the cavity of the archenteron, are now separate and distinct. The connection between these
structures still continues, however, in their fused walls. The left pouch is a little larger than the
right and lies behind and to the left of the latter (fig. 4, ael).

Just below the anterior pouches there is to be found a third pouch, which is growing out from
the left side and anterior end of the archenteron (fig. 4, hy). It protrudes anteriorly and partially
covers the two anterior euteroca-les. The cavity of this pouch, which is the rudiment of the
hydrocu'le, is in wide communication with the archenteron.

From the wall forming the convex sides of the hydrocwle there are, even at this early stage
in its formation, five outgrowths which are the beginnings of the radial canals of the adult
ophinran (fig. 4. 1, 2, 3, 4, and •">).

The whole hydrocu-le is curving round to the right to encircle the oesophagus, which latter is
making its first appearance iu this stage as a shallow but definite pit iu the central part of the
ventral ectodermal wall (fig. 4, oe).

To avoid confusion the hydroca'le was spoken of above as arising from the archeuteron, but,
as will be seen in the transverse section (fig. 5), taken in a plane posterior to the origin of the
hydroru'le, a differentiation is taking place in the archenteron which enables us to distinguish in
it the rudiments of two structures, the posterior enterocceles and the stomach. By a longitudinal
circular furrow the archenteron is being cut horizontally into a large ventral pouch, the posterior
enterocoales (pe) and a smaller dorsal one, the stomach (s). This stomach rudiment bends around
the posterior end of the posterior euterociele and opens to the exterior through the blastopore
(fig. 4).

It is from the left side and anterior end of the ventral pouch that the hydrocu'le grows out,
hence the confirmation of BURY'S statement that it arises from the left posterior enterocce.le iu
ophiurids.

In most echinoderms the posterior enterocreles originate as paired structures, and if the
statements of BURY and MCBRIDE are correct, that the left posterior enterocirle of the larva
forms the hypogastric body cavity of the adult, and the right posterior enterocu-le goes to form
the epigastric cii'lom, then, according to this, the large ventral pouch, which I regard as the
fused right and left posterior enterocu-les, really represents the left only, because it takes no part
in the formation of the epigastric body cavity of the adult ophiurid, but, with the left, does pass
directly into the hypogastric.

The origin of the epigastric enterocu-le is discussed in the description of Stage " C," in which
its rudiment is first found over the stomach.

My reason for regarding the ventral pouch of u B " as the fused right and left posterior entero-
cceles, is that at the time of its origin it is symmetrically disposed on either side of the plane of
larval bilateral symmetry.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 8i>

(ioTO (5), too, lias shown that the hypogastric, enterocrele in the starfishes is not formed from
the left alone, hut in it are to be found the left and the greater part of the right posterior entero-
cu'les.

The cellular plug of cells, which in "A'1 fills the archenteric cavity, becomes divided by the
furrow which separates the archenteron into enteroeu'le and stomach and a part of it becomes
inclosed in the cavities of each of these structures (fig. 5, cp).

STACK "('," 4X Hoi its OLD.
(Figures 6, 7, and X.)

The external form of the larva, which in this series of embryos is six hours older than "B,"
has been changed by the appearance of two lateral thickenings of the ectoderm a little posterior
to the median transverse plane (fig. G).

The blastopore, which in "15" was open to the exterior, has closed, leaving no trace of its
former position.

The mouth and (esophagus, which existed in "B" only as a shallow ectodermal pit, now have
the form of a deep, hollow tube (figs. G and 7, in and oe), which projects vertically inward until it
passes through the hydrowele and beyond the posterior enteroccele, when it curves back under
the latter to fuse with the anterior wall of the stomach.

The stomach and posterior enteroccele are still in open communication, as in "B," but the
furrow in "G" has deepened, and the process by which the two structures are being separated is
almost complete (fig. 7).

Although the walls of the (esophagus and stomach are fused, their cavities are still separate.

This condition renders it easy to see just what part is played by the ectoderm in the formation
of the alimentary canal, the entire oesophageal cavity being surrounded by ectoderm.

In "B" the hydroccele communicates with the posterior enteroccele by a wide opening, and
at the same point in " C " the two structures are still in communication, but the connection has
been narrowed down to a small tube (fig. 7, he).

Beside this connection with the posterior enterocosle, a second tube has been formed, joining
the left anterior enteroccele with the hydroccele (fig. 7, st). This new tube, which is the rudiment
of the stone canal, enters the hydroca-le at the same point with the tube connecting the latter
with the posterior enterociele.

The left anterior euteroco^le lies to the left of the (esophagus, and dorsal to the left half of
the hydrocu'le (tigs. (! and 7, ael).

It is to be noted that, although we now have a larva possessing both hydroccele and stone
canal, there has been as yet no pore canal formed. This is a marked reversal in the sequence of
the formation of these structures from what might be expected from the order of their appearance
in other known echinoderms, the pore canal arising usually before the formation of the hydroccele,
while the stone canal appears much later than either.

Returning to the hydrociele, we find it a horseshoe-shaped structure astride the (esophagus
(figs. G and 7, hy). The bulging areas which are to form the radial canals of the adult are much
longer and more regular in size than in "B." The radial pouch, which lies to the right of the
cesophagus and at the end of the right horn of the horseshoe, will hereafter be spoken of as radial
canal 1, since it arises from -that part of the hydroctele which was first to bud out from the
posterior enteroccele. The other radial canals, passing to the left over the (esophagus, will be
designated as 2, 3, 4, and 5. Radial canal 5 lies in this stage over the opening of the stone canal.

The rotation of the hydrocoele around the (esophagus from its original left position, which
was begun in " B," has continued to such an extent in UC" that half of it lies to the right of the
median sagittal plane of the larva and half to the left. Radial canal 3 lies in this plane and
points directly toward the anterior end of the larva (fig. 6).

With its rotation the hydroc(ele also moves bodily toward the posterior end of the larva,
carrying with it the (esophagus. The (esophagus, coming in contact with the anterior wall of the
united posterior enteroca-les, causes the latter to be pushed in at the point of contact. As the
process continues, those parts of the posterior euterocceles lying on either side of this in-pushing
area are forced to flow forward around the oesophagus and under the hydroccele; thus we have

90 MEMOIRS OF THE NATIONAL ACADEMY OP SCIENCES.

the posterior enteroc(ples becoming- horseshoe-shaped, the two horns of which lie under the horns
of the hydroccele (figs. (5 and 7, he and by).

Lying dorsal to the stomach we find a small enteroccele which was not present in "B," or if
present, not in this position. It is the rudiment of the body cavity, which in the adult lies aboral
to the stomach and which has been recently appropriately termed the epigastric enteroccele (figs.
C and 7, ee).

As to the origin of this structure I have no direct observations to give, but certain facts have
led me to believe that it is formed from the right anterior euterociele. These facts may be summed
up as follows: In "B" no epigastric enterocoele exists, but the two anterior euterocceles (fig. 4,
aer and ael) lie side by side anterior to the stomach and the posterior enterocudes. In " C " (figs.
G and 7, ee) an epigastric pouch, equal in size to the right anterior enteroccele of "B" is to be
found, but by the side of the (esophagus only the left anterior enterocrele remains (figs. 6 and 7,
ael).

During the six hours which intervene between " B" and "C" it seems hardly possible that a
complete formation of the epigastric enteroccele should have taken place or that there should
have been time for the complete degeneration and disappearance of the right anterior pouch;
sufficient time may have elapsed, however, for the migration of the right anterior enteroccele to a
position behind the stomach.

Against such an interpretation as the above there is the fact that in no other case has the
epigastric enterocn-le been observed to take its origin from the right anterior pouch. It has been
described as arising from the right posterior enterocu'le, however, as has been referred to before, in
all the groups by BURY, and his observations have been corroborated by both McBniDE and
GOTO in the starfishes.

STAGE "D," 60 HOURS OLD.

(Figures 9-14.)

The changes which have taken place in "C" to produce "D" are very marked.

The cilia have disappeared, except in four transverse rings or bands, three of which extend
entirely around the body of the larva. The third ring, counting from the anterior end, is inter-
rupted by the aboral disk on the ventral surface.

This third ciliated ring first appears on the lateral bulges, which were described in "C," and
the fourth ring appears on a second pair of lateral bulges which originate behind the first pair
near the posterior end of the larva.

The shape of the larva is no longer oval, but the posterior end has widened laterally and
become somewhat dorso ventrally compressed (fig. 9). The anterior end has not changed in shape
and may be thought of as forming the handle of the now club-shaped larva.

The enlarged posterior end of the larva contains all its organs and is the part which will enter
directly into the formation of the adult ophiurid.

From its homology with the preoral lobe and larval organ of Aster in a gibhosa I have called the
anterior end of the larva the larval organ. It disappears with the metamorphosis into the adult
form.

The larval organ is also homologous with the stalk of the Antedon larva, although in the
ophiurid larva it never functions as an attachment organ. When swimming, the larval organ
precedes. It is filled with a network of mesenchyme cells (fig. 11, mes).

Internally the changes have been even greater than the external ones we have just considered,
for it is during this period of development that the rotation and readjustment of organs takes
place, which is present in all echinoderms at some stage of their development.

The hydrocu'le, which has begun its rotation about the (esophagus as an axis in "C," has
completed it in UD" and reached its definite position.

That part of the hydrocu'le which in "C" was situated on the left of the plane dividing the
larva into bilaterally symmetrical halves, now lies on the right side of the same plane and vice
versa. (Compare figs. G and 9.)

A revolution of 180° has taken place in the hydroc<ele since "C," to which if the 180° of
rotation be added, which took place up to the time of UC," we have a total rotation of 360° in the

MEMOIRS OF THE NATIONAL ACADEMY OP SCIENCES. 91

hydroccele. Radial pouch 1, finally, after having passed around tlie (esophagus, comes to rest
at the point where it originated, lladial poiicli f>, it. will be noted, is carried only halt' as far
as radial pouch 1, or from its point of origin on the left to a point opposite on the right of the
irsophagus. (Compare figs. (! and fl, (1) and (5).)

This great amount of rotation seemed so peculiar that I hesitated for some time to believe it,
and was led to suppose instead that while the hydroco'le moved to the right the other organs
lying above it rotated an equal amount to the left.

The early closure of the blastopore and the central position of the mouth in the early stages
make such a view as the latter seem possible, and as it may suggest itself to those who study figs.
<> and 7, 1 will give below the points which seem to me, directly or indirectly, to prove that the
hydrocu-le revolves under the enterocceles and stomach, rather than that the latter twist over
the hydroccele:

(a) The ectodermal bulges, nearer the posterior end in "C" (tig. 0), are the same as those
nearer the posterior end of " D" (fig. 9), on which the third ciliated band is situated.

(b) If the latter view is the correct one then radial canal 3 points toward the same end of
the larva in both "C" and "D" (figs. 0 and 9), but in "C" the end toward which it points is
anteriorly directed in swimming and in "!)'' it points away from the end which precedes. It is
hardly thinkable that in any stage in its development the anterior end of a larva should change
its physiological function and become the posterior end.

(r) By any other view than the one I have adopted the blastopore, or the point where it
existed before closing, would be anterior and the larval organ posterior in position. In all known
echinoderm larva?, however, the blastopore marks the posterior end, and in all cases where it
occurs the larval organ originates from the anterior end of the larva?.

(d) It may be recalled, also, that in the readjustment of parts which takes place during the
metamorphosis of other echinoderm larvie the rotation is almost entirely confined to the hydroco>le.

As the hydroc(ele passes around the (esophagus the tube connecting it with the left horn of
the hypogastric enteroccele becomes broken and the left anterior enterocu-le, together with the
tube connecting it with the left horn of the hydroccele, are carried anteriorly around the (esophagus
(fig. 9, ael and st). In "D,'' then, we find the stone canal on the right side of a line dividing the
larva into symmetrical halves, instead of to the left of the same line as it is in "C." (Compare
figs. C and 7 with 9.) The anterior enteroccele comes to rest immediately in front of the stomach
and (esophagus.

From the point where the stone canal enters the anterior enteroccele the pore canal grows
out, passes dorsally to the ectoderm, with which latter its walls fuse, and an opening the water
pore (figs. 9 and 11, pc.) breaks through. Thus in this stage the eieloin and hydrocwle are first
connected with the exterior.

In ':C'' the circular water-canal had not closed, but existed in the form of a horseshoe, the
concave side of which opened posteriorly, but as the rotation of the hydroccele takes place its
horns grow toward each other until they meet. A fusion of their walls then takes place at the
point of contact and a complete ring is thus formed. The part of the ring canal, the formation of
which has just been described, lies between radial canals 1 and 5 in fig. 9. The opening of the
stone canal into the water ring is situated in "C" at the base of radial canal 5, but by means of
the rotation of the hydroc(ele about the (esophagus, together with the growth of the ends of the
horseshoe, this opening is carried away from its position at the base of radial canal 5 toward
radial canal 1. It always remains, however, nearer the former than the latter; in other words, it
comes to lie definitively in the right <i<lnnlitix between radial canals 5 and 1. (Compare figs. 6
and 9.)

The radial canals, which existed in "C" as simple pouches from the convex side of the hydro-
cojle, have in " D" each become three-lobed. Near the tip and from the sides of each canal a pair
of pouches has budded out, each of which is about equal in size to the end of the canal which lies
between and beyond them (fig. 9, et and tl). In these three structures we have the rudiments of
the end tentacle and the first pair of foot tentacles of the ophiurid arm.

When we were last considering the hypogastric enterociele it was in the form of a crescent,
the horns of which were very short and its central part very wide. Into its concavity, which was

92 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

anteriorly directed, the oesophagus fitted. The horns of this enterocosle, beginning in "C"to
grow over the hydrocci'le, continue the process during the rotation of the latter, the horns of the
crescent growing at the expense of the thickness of its central part, ar.d in "D" we have this
enterociele lying directly over the hydroca'le in the form of a perfect horseshoe (tig. 9, lie).

Between the ends of the horns of the hypogastric cu-loni lies the anterior enteroceele. The
walls of these structures come together and fuse in such a way that they together form a hollow
circular ccelom surrounding the stomach and lying over the somewhat smaller water vascular ring
(fig. 9).

In the four interradii, marked by their positions between radial canals 1 and 2, 2 and 3, 3 and
4, and 4 and 5, four pouches of the hypogastric enteroccele grow downward, outside the water,
vascular ring, forcing themselves between the radial canals; a fifth pouch, similar to those just
described, is formed from the left anterior enterocu-le in the remaining interradius between radial
canals 5 and 1 (figs. 9, 12, 13, and 14, hip 1-2, 2-3, 3-4, 4-5, and ipax 5-1). These five pouches
are the rudiments of the outer perihrcnial ring, which will be more fully considered in the
succeeding stages.

The stomach, after being entirely cut off from the hypogastric enterocoale, was drawn forward
during the rotation of the hydrocoele, and the (esophagus was carried in the opposite direction,
so that in "D" the stomach lies almost directly over the u-sophagus (figs. S and 13, oe and s). The
partition, which in "C" separated the cavities of these two structures, has disappeared in "D,"
and the cesophageal cavity opens into that of the stomach. There is present, then, in "D" the
definitive alimentary canal of the adult ophiurid.

The "cellular mass," which in "15" and "C" was being divided into two parts by the
constriction separating the archenteron into enteroccrle and stomach, is to be found, in sections
of "D," in the cavities of both the above structures (figs. 11, 12, 13, and 14 cp).

Lying immediately above, or aboral to, the stomach is to be found the epigastric euterocu'le.
It has enlarged considerably during the interval between "C" and "D," but is not yet of sullicient
size for its walls to touch those of the hypogastric cu^lom, and hence in this stage no circular
aboral mesentery is to be found.

STACK •')•:," liG HOURS OLD.
(Figures 15-21.)

The thickening of the ventral ectoderm which was begun in "D" has continued during the
six hours which intervene between "D" and "E" and has spread to the sides of tlie larva (ligs.
15 and 19-21).

Near the edge of this thickened oral disk are to be found five groups of rounded elevations of
the ectoderm (fig. 15, I, II, III, IV, and V). The three elevations, of which each group consists,
form the angles of an isosceles triangle the apex of whicli points away from the mouth of the
larva (fig. 15). These elevations or evagiuated papilla1 lie immediately below and inclose the tips
of those branches of the radial water canals which form tlie rudiments of the end tentacle and
first pair of foot tentacles of each arm (figs. 19 and 21). In this way each tentacle grows into its
ectoderm, the latter closing around it as it pushes out.

The function of these tentacles in the adult being mainly a sensory one, it is interesting to
note that they receive their ectoderm from part of the same thickened oral area which gives rise
later to the adult nervous system.

The ciliated bands in "E" do not differ in appearance and position from those in UD," but
since they were not figured in the earlier stage it may be well to refer to them again in connection
with figs. 15 and 10, cb 1, 2, 3, and -1. The first or most anterior band surrounds the larval organ
quite near its tip.

Near the first band, and parallel to it, runs the second one .also around the larval organ. The
third ciliated band is separated from the second by a much wider space than that which separates
the first and second bands. Were it not interrupted on the oral disk the third ciliated band would
lie in the line separating the biviuin and trivium — that is, between arms I and II on the one side
and IV and V on the other. The fourth band, passing just posterior to the group of ectodermal
elevations lying under the branches of the third radial water tube, surrounds the posterior end of
the larva.

MEMOIHS OF THE NATIONAL ACADEMY OF SCIENCES. 93

In "E" the cavities of the oesophagus and stomach have become obliterated, and the two
structures appear in section as one solid mass of cells (fig. 20, st and oe). No degeneration in
their size, however, is to be observed, and their outer walls remain well defined, the (esophagus
retaining its connection with the ectoderm. As will bo seen later, their lutnen reappear and they
become the definitive alimentary organs of the adult ophiurid.

Returning to the consideration of the water system, we find in "E" instead of one pair of
tentacles on each radial canal, as in " 1),'' there are two pairs present (fig. 17, tl and t2), the
second pair having grown out of the radial canal between the first pair and the water ring. The
second pair is much smaller than those which were first to be formed, and, contrary to what
one would expect, this discrepancy in size does not disappear as time goes on. This is also true
in the sea-urchins, in which the primary tube feet in the larva are enormously larger than those
which are subsequently formed. The primary tube feet in this case gradually diminish in size
after the adult form is reached.

As a rule, among echiuoderms the tube feet or tentacles are formed centrifugally from the
radial canals; that is, between the end tentacle and the last pair of tube feet or tentacles already
formed. This process keeps the undifierentiated growing point of each arm at its tip, but in this
ophiuran, and the same is true of Autedou, the formation of the tentacles begins in a centripetal
manner; that is, the second pair of tentacles appears, not between the end tentacle and the first
pair, but between the first pair and the ring canal.

This second pair of tentacles is the rudiment of the buccal tentacles, and although differing in
both function and position in the adult from that of the foot tentacles, is nevertheless entirely
homologous with the latter. This homology is shown by their origin and the fact that for a time
after forming they are directed away from the mouth toward the end of the arm just as is the case
with the foot tentacles. After a time, however, as will be seen later, they turn back and point
toward the mouth, thus showing that in this second pair of outgrowths from the radial canals
we have to do with the first pair of buccal tentacles of the adult. After budding, as we see, from
the radial canals, they migrate to a position on the ring canal, with which we find them connected
in the adult.

In " E" the buccal tentacles have no ectoderm nor rudiment of such, the ectoderm under their
tips being as yet undiflereutiated from, the oral disk.

The hypogastric euteroco?le has assumed a more pentagonal shape than in "D," it having
grown out over the radial water canals (fig. 17, he). These projections of the hypogastric eutero-
coale will continue to grow with the growth of the arms and become the brachial extensions of
the body ctelom.

The interradial pouches of the hypogastric euterocosle, which were beginning to form in "D,"
have pushed down further and further between the radial canals until, coming in. contact with the
ventral ectoderm, they bend over, inserting themselves between the ring canal and the oral disk
(figs. 17-21, hip 1-2, 2-3, 3-4, and 4-5). In the same way the pouch from the anterior enteroctele in
the stone canal iuterradius has grown under the water ring. In these five iuterradial euteroccelic
outgrowths, as has been mentioned before, we have the rudiments of the outer peritutmal sinus
of the adult. The process by which this perihrernal sinus is formed in Ophiura brevispina agrees
in every detail with its method of origin in Asterina yibbosa as described by McBniDE (10).

The epigastric enterocuele is iu much the same condition as that in which we left it in ;'D," it
being as yet too small to meet and form a mesentery with the dorsal edges of the hypogastric
cffiloru (fig. 19, ee).

The stone and pore canals, too, have changed very little during the interval between " D "
and " E." From the ring canal at a point to the left of the origin of radial canal 1 the stone canal
passes upward and opens into the right postero-dorsal part of the anterior enteroco'le. The pore
canal begins at the same point where the stone canal ends, the two canals thus having a common
opening into the anterior enterocoale or ampulla?. The pore canal extends from the enterocu?le to
the dorsal surface of the larva, where it empties through the dorsal pore at a point a little to the
right of the median sagittal plane. These two canals, although extending in the same direction,
do not lie in the same straight line, the pore canal being set a little anterior to and to the right of
the stone canal (figs. 17, 18, and 21, st and pc; also fig. 11).
10396 3

94 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

STAGE " F," 5 DAYS OLD.
(Figures 22-30.)

Although " F " is separated from the stage last described by a considerable space of time, the
changes in the larva which have been brought about are easy to follow.

The larva is considerably larger than iu '' E," and has reached its full development. From
this time on the larval organ gradually degenerates and is finally completely absorbed by the
developing star (Compare figs. 22 and 31 lo.)

The external form of the larva has been changed by the appearance of a number of elevations
and depressions iu its outer surface, the ciliated bands beiug elevated upon circular ridges (fig.
22, cb 1, 2, 3, and 4), while at points on the sides of the disk beyond the end tentacles projections
in the ectoderm have made their appearance, these being the rudiments of the ophiuran arms (fig.
22,1,11, III, IV, and V).

The larval organ is cylindrical, but the disk has continued its dorso ventral flattening.
(Compare figs. 22 and 20.)

The first and second ciliated bands are situated in the same places as in " E." The third,
while retaining its old position, has grown in upon the ventral disk toward the mouth (fig. 22, cb 3).
On the ventral side of the larva the fourth band has shifted from its old position behind the third
radial canal to one on the interradii between arms II and III and III and IV. It has also become
interrupted on the oral disk in a manner similar to the third ciliated baud (fig. 22, cb 4).

The depressions before referred to are caused by the invagination of the nervous system, which
structure has been forming since "D" in the thickened oral disk of ectoderm. Immediately below
the water ring and radial water canals the thickening has increased more rapidly than at other
points, thus producing a ring-shaped internal ridge, from which extend five radial thickened
ridges. These rudiments of the nerve ring and radial nerves bulge inwardly, no evidence of their
presence being apparent on the outside. When the thickening process has been completed the
whole nervous system gradually sinks in, leaving a circular groove from which five radial grooves
pass out. This is the stage in the formation of the nervous system which has been reached in
"F" (fig. 22, eg and rg). The invagiuation process begins at the ends of the radial nerves, just
inside of the curved tips of the end tentacles, the nerve ring being iuvagiuated last of all.
(Compare figs. 23-27.) As development goes on the edges of the grooves gradually close over the
nerves, the closure taking place in the same order as the invagination proceeded — that is, first
over the ends of the radial nerves, then finally, after gradually traveling up the radial nerves,
closing over the nerve ring.

By the meeting and subsequent fusion of the edges of the grooves, part of their cavity
becomes cut oil' from the exterior and is left below the nervous system as the subneural space.
But this will be referred to again iu an older larva, in which the process of its formation is more
nearly completed, it having begun in a few only of the most advanced larva' of Stage " F."

The nervous system shows a differentiation into two distinct layers, a fibrous one nearest the
water system and a cellular layer lying below the fibers (figs. 23-29). The nuclei of the cellular
layer are oval, with their long diameter perpendicular to the fibrous layer.

Above the nervous system, separating it from the water system, is to be found the outer
periha-mal space (figs. 23-30, opr). Eecalliug the condition of the perihsrinal system in Stage "E,''
we see that the ends of the iuterradial projections from the hypogastric and anterior euterocosles
have grown out over the nervous system, spreading in both directions until the outgrowths of each
interradial pouch meet those of its adjacent fellows in the radii over the origins of the radial
nerves; here the ends of the pouches fuse, and together they grow out over the radial nerves as
the radial periluvmal sinuses.

In the, starfishes, where the formation of this perihicmal system has been observed, it is said
that no fusion takes place between the diverticula of the interradial pouches of the hypogastric
and anterior enterocceles when they meet in the radii, but that a mesentery is formed at the points
where the diverticula come in contact. This mesentery is described as continuing to the end of
the arms, separating the radial spaces into two parallel cavities.

Nowhere could I find such a mesentery iu sections of the larvaj of 0. brevispina, nor could I
feel sure that it exists in the adult ophiuran.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 95

McBRiDE (10) and GOTO (5) both agree that iu starfishes the inner perihnemal ring sinus
arises from the anterior euteroccele, although they differ as to the method of its formation. In
none of the larva; I have is the structure iu question fully formed, but in Stage "F"-a cavity is
arising, as an outgrowth from the anterior enterocade in the stone canal interradius, which I
take to be the rudiment of the inner perihiemal ring space. It lies to the left of the stone canal
near the median sagittal plane of the larva. From the posterior side of the ventral end of the
anterior enterocu'le the outgrowth takes its origin, then extending posteriorly until past the water
ring it bends over and grows down until its end reaches the nerve ring at a point inside the outer
perihiviual sinus. Here the end of the pouch in question begins to spread under the nerve ring
in both directions, parallel to the outer periLuemal ring (figs. 24 and 26, ips). This coincides
exactly with its method of origin in Asterina gibbosa as described by McBRiDE.

Although the outer periluemal ring is entirely cut off from the body cavities at this stage,
there still remain traces of the interradial pouches which gave rise to it (fig. 29, hip 1-2 and 4-5).

The hypogastric enteroccele itself has changed very little since Stage "E," but the epigastric
has enlarged to such an extent that its edges now meet the edges of the hypogastric and a circular
aboral mesentery is formed (figs. 23-29. cm).

In the water system considerable growth is to be noted in the tentacles, the end and first
pair of foot tentacles being capable of protrusion considerably beyond the disk. By means of
these tentacles the larva- are able to cling tenaciously to the surfaces of foreign bodies, it requiring
a strong jet of water from a pipette to detach them. Special notice was taken to be sure that it
was the tentacles and not the larval organ which was used as a means of attachment.

The second pair of tentacles (buccal tentacles) have acquired their ectoderm in this stage and
they protrude, like the other tentacles, over the radial nerves into the radial grooves (figs. 22
and 20, t2).

The axial sinus or ampulla is present in "F," it being that part of the anterior enteroccele
which remains after the pouches have been cut oft', which will form the inner perihtemal and part
of the outer pcriha-uial systems (figs. 24, 25, and 20, ax sin).

It will be noted that beside the, ampulla, which is situated anteriorly to the stone canal, there
are two other cavities near the stone canal to be accounted for (fig. 20, sin a and sin b). I can
not be sure of their origin, but I believe that they also come from the anterior enterocoele. I have
distinguished them by the letters a and b, as they are probably the same cavities as those so let-
tered by McBuiDE (10) in his figures of Amphiura.

The cavity McBRiDE has lettered b', and which he thinks represents the degenerated right
hydroecele, I have been unable to find in any of my sections.

The stomach and oesophagus are in a condition the same as we found them in "E;" that is,
without luiueii.

STAGE "G," 5} DAYS OLD.
(Figures 31 and 32.)

Larvre a few hours older than "F " show a decided degeneration in the larval organ (fig. 31, lo),
but otherwise the external appearance of the two stages is about the same.

The grooves caused by the invaginatiou of the nervous system have begun to disappear by
the growing together of their edges, and instead of the deep furrows we find a slightly depressed
line where the edges of the grooves have met (fig. 28, rg).

In the nervous system a pair of tentacle nerves has been formed from each radial nerve (fig.
31, nl). They grow out laterally from the radial nerves at points proximal to the first pair of foot
tentacles, around which latter they grow and to which they belong. No nerves as yet have
appeared to supply the buccal tentacles.

In "F" the buccal tentacles had only begun to curve away from the end of the arm; but in
"G" this proximal bending has continued until they curve over the nerve ring and point toward
the mouth.

Beside this change in the water vascular system we find in "G" the first appearance of the
rudiments of the polian vesicles. They are four in number and are in the form of small interra-
dial pouches growing distally from the convex wall of the water ring (figs. 31 and 32, pv). There
is no poliaii vesicle present in this species in the stone canal interradius.

96 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

As no perceptible change has taken place since " F" in the organs not referred to above, the
description of them given in the previous chapter will serve equally as well for "G" as for "F,"
and the figures of these organs in "G" may be examined in connection with their description
in " F."

STAGE "H," 8 DAYS OLD.
(Figures 33 and 34.)

In the oldest larv* I have, the metamorphosis has been almost completed. The larval organ
has nearly disappeared, that part of it which yet remains being found sticking to the edge of the
aboral disk of the young pentagonal star.

When living the little ophiurids clung to the bottom and sides of the aquaria dishes. Although
the ciliated bauds were still evident on the disk their free swimming habits had been wholly
given up.

The pore canal still opens on the aboral surface, but with the growth of the latter it is
traveling toward the edge of the disk, and by a continuation of this process the oral surface will
ultimately be reached.

As the closure of the grooves over the nervous system took place, circular areas below the
tips of the tentacles were left open, the tentacle pores, and through these the tentacles, were able
to protrude and withdraw themselves.

The subneural sinuses which had begun to be formed in "F" have been completed in the
eight-day larva (figs. 31 and 34 ss). In "H," then, the nervous system is cushioned below by the
subneural sinus and above by the outer perihrenial ring.

The stomac-h, which for so long a period has been at a standstill in its development, has begun
to grow, its sides pushing out between the epigastric and hypogastric body cavities. The lumen
of both stomach and oesophagus have reappeared (fig. 34 s). The glandular structure which makes
the walls of the stomach so complicated in the adult has not begun to form in "H," the walls being
simple and one cell in thickness.

No figure of " H" as a whole object has been made for the reason that the skeletal plates should
be included, and material adequate to a complete study of them is at present not in my possession.

RELATION OF LARVA TO ADULT.

The hydroccele is the first organ to show radial symmetry in the developing larva of Opldura
Ireritiirina, and from the time when this organ has completed its rotation about the ci'sophagus it
shows a definite relation to the plane of bilateral symmetry of the larva.

The hydroecele is not only radially symmetrical, but bilaterally symmetrical, since it is divided
into symmetrical halves by the plane which passes through radial canal 3 and through the inter-
ratlius of the stone canal. This plane coincides with the plane of bilateral symmetry of the larva.
The other parts of the star are built about the water vascular system; hence it, as a whole, bears
a similar relation to the larva as was initiated by the hydroccele.

No secondary twisting of the various parts of the star occurs, and its relation to the larva
remains constant as it began, and throughout the life history of the species the following state-
ments hold true: Ventral and dorsal in the larva are equivalent to oral and aboral iu the adult.
Although no physiological differentiation exists, if we regard that part of the adult as anterior
which was anteriorly directed in the free swimming larva, the triviuin is anterior, the bivium is
posterior.

In the foregoing I have confirmed, in an ophiurid, the conclusions drawn by GOTO from his
studies on a starfish. In his work on the development of Axierinn i>aUi(la GOTO (5) thought he
was able to prove the coincidence of bilateral symmetry, which obtains in the adult starfish, with
the plane of bilateral symmetry of the bipinnarian and brachiolarian larva1.

The study of the relation of larva to adult in the starfishes is made most difficult, however,
by the independent origin and subsequent twisting of the parts of the star. At the time of their
origin no two parts of the star bear the same relation to the larva. The relation of each part to
the larva also changes as metamorphosis proceeds.

The facts just enumerated admit of other conclusions than those deduced by GOTO, and no

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 97

two investigators have readied the same conclusion. The point of view from which the subject
Las been approached is not the same in all cases, but the results obtained by those who have
studied the question admit of being reduced to the same basis; that is, the relation of the planes
of bilateral symmetry in larva and adult.

CUENOT (4) in his latest work denies the existence of any known relation between them.

SEMON (13), working on a holothnrian, found the two planes in question to coincide, but his
conclusion is based on the supposition that the dorsal mesentery of the adult is the same as that
of the auricularia larva, which supposition BURY has since shown to be incorrect.

BURY (3), after working on members of all the groups of echinoderms. concluded that the
plane of bilateral symmetry of the larval form coincides not with the plane dividing the adult
form into two symmetrical halves, but with the plane of radial symmetry.

McBRiDE's (10) observation on a starfish, Asterina gibbosa, led him to adopt about the same
view as that of BURY. He found that the plane of radial symmetry of the star makes an angle
of 70° plus with the frontal plane of the larva, but may, without error, be considered as 90°. This
is equivalent to saying that the plane of radial symmetry of the star is parallel with the sagittal
plane or the plane of bilateral symmetry of the larva, and is also reducible to the statement that
the planes of bilateral symmetry of the larva and adult are at right angles to one another. Thus
right and left in the larva become aboral and oral in the adult.

The difference in results arrived at by GOTO and McP.RiDE are due almost wholly to the
stages in the metamorphosis selected in each case for the study of the question, GOTO selecting a
very late stage, when the larval body had all but disappeared, while the stage chosen by McBEiDE
is an early one, in which the rudiments of the star are just appearing.

If the five groups of echiuoderms have sprung from a common stem after radial symmetry
had been established, then in the metamorphosis which is found in all the groups there should be
discoverable a unity of relation between larva and adult. It is hard to conceive of the radial
symmetry of echinoderms as having been independently acquired by each group, although it is
easy to see how secondary changes may have arisen in the metamorphosis since the groups separated.

The five groups of echiuoderms stand isolated from one another almost as completely as does
the echinoderm phylum from the other phyla of the animal kingdom, and it is not my intention at
this time to enter into a discussion of the interrelationships of echinoderms. I wish, however, to
point out an interesting series of facts presented by members of the Asterid, Crinoid, and Ophiurid
groups which may have a bearing upon the subject, and in the same connection I wish to call
attention to how well McBRiDE's hypothetical ancestor of the Asterids and Crinoids (10, fig. — ),
when details are not too closely compared, fits into the facts of the larva of Ophiura brerispina,.

In one of the Asterids GOTO has shown that toward the end of metamorphosis the almost
complete star sits as a cap at the posterior end of the larva, with its aboral end posterior, its oral
surface, anterior, the bivium dorsal, and the trivium ventral.

In Autedon, like the starfish, the rotation brings the developing crinoid head to the posterior
end of the larva, but differing diametrically from the starfish in that the oral instead of the
aboral surface of the crinoid is posterior; but this difference does not in any way affect the
homologies between the two errouus as has been sunnosed.

metamorphosis is approx-
1 in the larva is ventral

3 disk to rotate in such a

lobe, it fairly represents

link of it as rotating in

obe, then it more nearly

e aboral surface, while in
9. brevispina the place of
ny other echinoderm, the
job near the edge of the
aboral surface between arms I and V.

98 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

COMPARISON OF THE LARVA OF ANTEDON ROSACEA WITH THAT OF OPHIURA

BREVISPINA.

While I was studying the larva of Opliiura brevispina, characters were constantly being
found which reminded me of the larva of Autedon as described by BURY (1). Some of these
points of resemblance are no doubt only superficial, but others are such as to make it worth the
while to devote a short chapter to the similarities of the two larva?.

The entire ciliation of the very young larva} gives place in both to a series of transverse
ciliated bauds, five in Autedon, four in Opliiura. The band nearest the anterior end of the Ante-
don larva, however, is small and incomplete. Two bauds only in each case surround that part of
the larva from which the disk is formed.

The blastopore in both larvre, after shifting from a posterior position to one on the ventral
surface, closes and the archeuteron loses its connection with the ectoderm and lies free in the body
cavity.

Iii the seven-day embryo of Autedon and Stage "C'' of Opliiura the hydrocoele is a horseshoe-
shaped structure lying in the posterior ventral part of the larva? with the open end directed
'anteriorly, and in each case the plane of radial symmetry of the hydrocu-le is at right angles to
the plane of bilateral symmetry of the larvae.

In the formation of the paired tentacles from the radial water canals the process is begun
ceutripetally in both larva?, the second pair of tentacles appearing between the first pair and the
water ring instead of between the first pair and the eud tentacle, as is the case in the other groups
of echinoderms.

In the five-day Antedon larva and those stages represented by "D" to "F" in Opliiura the
stalk and larval organ are strikingly similar, both in shape and position, the two structures being
anteriorly directed in swimming.

The stem of the Antedon larva, as a result of metamorphosis, comes to be an aboral structure,
and just betore the disappearance of the larval organ from the ophiuran larva it is to be found as
a small knob, not in the center of the abornl disk, it is true, but on its edge. In the starfishes it
may be recalled that the preoral lobe disappears on the oral surface of the metamorphosing star.

To the above larval characteristics may be added the similarity which exists in the disposition
of the alimentary and cirlomic systems in the adult forms.

In both Crinoids and Ophiurans the digestive apparatus is confined to the disk.

The body cavity is continued into and to the ends of the arms. When a transverse section of
a pinnule of Autedon is compared with a transverse section of an ophiurid arm, the following
striking correspondence is found in the parts: Aborally, segmentally arranged calcarious ossicles
and muscles are present; a continuation of the body cceloin runs between and oral to the muscles;
connected with and on each side of this central brachial body cavity are two other cavities, the
subtentacular canals of Antedou. These latter in the Ophiuraus are connected with periheemal
space in each vertebral segment.

The radial water tube lies between the subtentacular canals, and in each segment sends out a
pair of tentacles. The tentacles iu both the Crinoids and Ophiurids are devoid of the terminal
suckers, which are so characteristic of the other echinoderms.

Separating the radial water tube from the nerve cord is to be fouud the radial periluumal
sinus.

In ophiurans a subneural space is present which is not represented in the crinoid arm. This
is due to the fact that in Antedon the nervous system is superficial, while in Opliiura it has been
iuvagiuated, and with its iuvagination a space has also been carried in below it.

LITERATURE CITED.

1. BURY, H. "Early Stages in the Development of Antedon rosacea." Philosophical Transactions, vol. 170. 1888.

2. BURY, H. " Studies in the Embryology of the Echmoderms." Q. J. Mio. So., vol. 29.

3. BUHY, H. "The Metamorphosis of Echmoderms." Q. J. Mic. Sc., vol.38.

4. CUENOT, L. " Etudes Morphologiques sur les Echinodenues." Archiv. Biol. t. XI.

5. GOTO, S. "The Metamorphosis of Asterias pallida, with special reference to the fate of the Body Cavities. '
Journal of the College of Science Imp. Univ Tokyo, Japan. Vol. X, Ft. III. 18M8.

6. GRAVE, C. "Notes oil the Development of Ophi ma olivacea." Zool. Auzeig. Bd. XXII. No. 580. 18!)9.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 99

7. KROIIN, A. " Uobcr eineu neuen Entwiekeluugsinodus der Ophiuren." Mueller's Archiv. f. Anat. u. Phys., 1857,
p. 369.

8. LYMAN, T. " Ophinra broviepina Say." Pt. I. 111. Cat. Mus. Comp. Zool., Harvard Coll. 1865, p. 16.

9. LYMAX, T. " Report on the Ophiuridse." Vol. V. Challenger Report.

10. MeBiiiDE, E. W. " The Development of ABterinagibbosa." Q. J.Mic. Sc., vol. 38.

11. FRKYER, \V. " Ueber die Bewegungen der Seesterne.'' MittU. a. d. Zool. Stat. z. Neapel. den. 1886.

12. SAY, T. "Ophiura brevispinn," Journal Phil. Acad., Vol. V. 1825. Page 149.

13. SEMON, R. '• Die Entwickelnng derSyuaptadigitataunddieStamniesgeschicliteder Echiuodermeu." Jonaisch.
Zeitschr. f. Naturwiss. XXII. 1888.

EXPLANATION OF PLATES.

The figures illustrating this paper were drawn to the same scale of magnification, 330 diameters, and were
reduced one-half in reproduction.

Figs. 1, 2, 4, 6, 7, 9, 17, 31, and 32 were reconstructed from series of transverse and sagittal sections.

In all cases the part of the figure which is nearest the top of the plate is either anterior or else ventral in the
larva.

ABBREVIATIONS USED.

ael Left anterior enterocojle.

aer Eight anterior entorociule.

ap Apical plate.

ax sin Axial Sinus.

b Blastopore.

cb 1, 2, 3, and 4 Ciliated bands.

eg Circular groove.

cm Circular mesentery.

cp : Cellular plug.

d Dorsal pore.

ee Epigastric enteroctcle.

et , End tentacle.

he Canal connecting posterior en-
teroctcle and hydroca'le.

he Hypogastric euteroecele.

hip 1-2, 2-3, 3-i, 4-5 Interradial pouches of the hypo-
gastric enteroccele.

hy Hydroccele.

ips Inner perihsemal sinus.

ipax 51 Interradial pouch of axial sinus.

lo Larval organ.

m Mouth.

mes Mesenchyme.

ill First pair tentacle nerves.

nr Nerve ring.

oe (Esophagus.

od Oral disk.

opr Outer perihtemal ring.

pc Pore canal.

pe Posterior enterocoeles.

pv Poliau vesicle.

rad 11 Radial nerves.

rg Radial grooves.

rps Radial perihsemal space.

sin a Sinus "a".

sinb Sinus "b".

8t Stone canal.

88 „ Subneural sinus.

8 Stomach.

tl First pair foot tentacles.

t2 Buccal tentacles.

wvr Water vascular ring.

I, II, III, IV, V Arm rudiments.

1,2,3,4,5 Radial water canals.

100 MEMOIKS OF THE NATIONAL ACADEMY OF SCIENCES.

PLATE I.

FIG. 1. Larva in Stage "A," seen from the right side, the right half of the ectoderm removed and the mesenchynio

omitted.

FIG. 2. The same larva, seen from the ventral side as a transparent object.
FIG. 3. Median longitudinal section of a larva in Stage "A."

FIG. 4. A larva in Stage " B," viewed from the ventral side as a transparent object.
FIG. 5. Transverse section of a larva in Stage " B '' in a plane halfway between the blastopore and the point where

the hydroeele is connected with the archenteron.
FIG. 6. Larva in Stage " C," seeu from the ventral side, the ventral half of the ectoderm, the mesenchymo, and part

of the nigophagus removed.
FIG. 7. The left half of the same larva.

FIG. 8. Transverse section through Stage "C" in a plane indicated on fig. 6 by the letters a-i.
FIG. 9. The reconstructed internal anatomy of a larva in Stage •' D," the ventral ectoderm removed and with it part

of the oesophagus.
FIG. 10. An outline drawing of fig. 9, on which are indicated by lines the planes of the sections which follow in

figures 11, 12, 13, and 14.

FIG. 11. Longitudinal section taken through a larva in Stage <;D" in the plane indicated on tig. 10 by the Hue m-n.
FIGS. 12, 13, and fig. 14 of Plate II. Transverse sections taken through Stage "D" in planes indicated on fig. 10 by

the lines a-b, c-d, and «-/.

PLATE II.

FIG. 15. Ventral view of a larva in Stage "E," to show ciliated bands and first appearance of the arm rudiments.

FIG. 16. Dorsal view of Stage "E," showing the ciliated bands.

FIG. 17. A recoustruction of the anatomy of a larva in stage "E," the ventral ectoderm removed.

FIG. 18. An outline drawing of fig. 17, on which are indicated by liues the planes of the sections shown in figs. 19,

20, and 21.
FIGS. 19, 20, and 21. Transverse sections taken through larva in Stage "E" in planes indicated on fig. 18 by the

lines r-s, t-n, and v-w.
FIG. 31. Reconstruction of the anatomy of a larva in Stage "G." In this case as in all the other reconstructions

the ventral surface is np and the ventral ectoderm removed.

FIG. 32. An outline drawing of the water vascular system of a larva in Stage "G," seen from the ventral surface.
FIG. 33. Transverse section of Stage "H," taken through the region of the stoue canal.
FIG. 34. Transverse section of a larva in Stage "H," taken through the stomach.

PLATE III.

FIG. 22. Ventral view of the fully developed larva before metamorphosis has begun. Stage "F."

FIG. 23. Outline drawing of fig. 22. The lines indicate the planes of the sectious, which have been drawn to show

the anatomy of a larva in Stage "F."
FIGS. 24, 25, 26, and 27. Longitudinal sections of a larva in Stage " F," the planes of which are indicated on fig. 23

by the lines a-b, c-d, m-n, and x-y.
FiGS. 28, 29, and 30. Transverse sections of a larva in Stage "F," the planes of which are indicated on fig. 23 by

the lines c-o, d-o and e-o.

MEMOIRS NATIONAL ACADEMY OF SCIENCES. VOLUME VIM.

PLATE I.

ap
•"•••.-:.

.••t*« iV

Fig. 9.

Pig. 13.

CASWELL GRAVE DEL.

MEMOIRS NATIONAL ACADEMY OF SCIENCES. VOLUME VIII.

PLATE II.

cbl

cba

cbs

ho :

et

nr

cm

Fig. 31.

1ASWELL GRAVE DEL.

MEMOIRS NATIONAL ACADEMY OF SCIENCES. VOLUME VIII.

PLATE III

e
d

IV
c

opr

ax sin

ax sin wvr

opr

he

;pr
radn

he

Fig. ai.

Ptg. 25.

Pig. 22.

Pig. 27.

Fig. 26.

opr

opr

..•eg

rad n

t T 12

••-^P^ ;,-"*K""*%\

b :W*Q&F;:>1,,

rps

he

cm

Fig. 30.

wvr ,-e

ASWELL GRAVE DEL.

6. THE ANATOMY AND DEVELOPMENT OF CASSIOPEA XAMACHANA.1
BY ROBERT PAYNE BIGELOW.

INTRODUCTION.

During the summer of 1891 the Marine Laboratory of the Johns Hopkins University
was stationed in the Island of Jamaica. It was at Port Henderson, — a little hamlet
situated at the west side of the mouth of Kingston Harbor, at the point where the Salt
Pond Hill, eight hundred feet high, descends to the salinas and mangrove swamp*
surrounding the mouth of the Rio Cobre. On the other side of the hill, and to the south
of it, there is a considerable body of salt water, known as the Great Salt Pond. It is
completely separated from the sea, but only by a beach of sand, which at its narrowest
part is not more than a few rods in width ; and it is said by those who live near that in
times of storm or freshet this barrier may be broken through.

One morning in June Dr. G. W. Field was hunting birds along the seaward shore of
this pond, and came upon a little bay that forms a deep indentation in the barrier and is
connected with the pond by a narrow inlet. The bay is overhung by low cashaw and
mangrove trees. At one side is a sunny sandy spot where a crocodile had made its bed,
and a fresh zigzag mark showed where it had recently slid into the water. A school of
fish was circling about in the clear water, and barnacles and sea-anemones spread their
tentacles from the submerged roots of the mangroves, while the bottom at the inner end
of the bay was completely carpeted by a colony of beautiful rhizostomatous medusae.

A few very small specimens might be seen swimming about, but most of the medusae,
especially the larger ones, would not ordinarily leave the bottom. They lay there upon
their backs, with their voluminous, branching mouth parts spread out over their discs,
which were motionless, except for occasional flaps of their margins. If any of these
animals were disturbed, they would, however, swim about like ordinary medusae ; but
before long they would settle down again and assume their usual attitude upon the bottom.

1 An earlier draft of this paper was accepted in May, 1802, by the Board of University Studies in the Johns Hopkins
University as a thesis for the Degree of Doctor of Philosophy. During the years 1891-92 and 1892-93 I held the Adam T.
Bruce Fellowship in the Johns Hopkins University, and WHS thus enabled to make a second journey to Jamaica. Publication
has been delayed in order that the results of this journey might be incorporated in the paper, and it is hoped that the greater
accuracy and completeness thus obtained have added materially to its value.

192 ROBERT PAYNE BIGELOW ON

Within this limited area there were countless numbers of them, and in many places they
were so thickly spread that their margins touched upon all sides, or even overlapped.

The spectacle presented by this collection of medusae was truly marvelous ; and in
order to show something of it to the rest of us, Dr. Field gathered a pailful of specimens
and brought them to the laboratory. Upon examination they were all found to belong to
a single new species of Cassiopea, — a genus of which only one species was known to
occur outside of the Eed Sea, Indian Ocean, and southwest Pacific ; l and this pailful,
taken up at random, contained both adults and young in various stages of growth.

Professor Brooks made drawings of some of these, and then I made a visit to the
Salt Pond to obtain more of the young medusae, and at the same time I collected sub-
merged bits of wood and stems of plants. My hopes were more than realized when, upon
examining these objects in the laboratory, I found them thickly studded in places with
scyphistomas in various stages of development. I was particularly delighted when I
noticed in one of the largest larvae certain glistening spots in the bases of the tentacles
and found, on putting them under the microscope, that they were unmistakably masses of
calcareous bodies that would form part of the marginal sense organs of the adult. They
excited my interest, especially as I had been studying the development of these structures
in Discomedusae ('90) , and had been unable hitherto to obtain the early stages.

After this discovery I began, with the advice of Professor Brooks, to make a careful
study of this species, with the intention of carrying the investigation of its anatomy and
development as far as the limited amount of time at my command would allow. Prelim-
inary accounts of my results were published in 1892 (Bigelow, '92, ((, l> and c) . In the
spring of 1893 I had another opportunity to visit Port Henderson with a party from the
Johns Hopkins University, and was able to make important additions to my earlier
observations.

During the first visit to Jamaica I was unable to find Cassiopea outside of the one
locality that I have described, and, although both the adults and the young in nearly all
stages were present at this place in such great numbers, searches for males and for females
with ripe eggs were equally fruitless. The great abundance of young and the range in
their apparent ages was, therefore, surprising, until I found that the scyphistomas were
multiplying freely by budding, in a manner to be described later on. During my second
visit I found this species as abundant as ever in this locality, and I also found a number of
adult specimens in several of the small shallow lagoons among the mangroves in the rear
of Port Royal. But these were all females, and it was still impossible to obtain males
or eggs that would develop.

1 Alilmii.tjli I'Vwkcs ('82) iiK-nt.ilios Ms Cussiviini frmiiloiia, Lamarck, of Key West and the Torluiias, with Pulydonia
frondosa Ag., it is nevertheless a true Cassiopea, not a Polyolonia.

CASSIOI'KA XAMACIIAXA. 193

Full-grown medusae could be kept in good condition in ;it|ii!iri;i for a number of days,
and could be kept alive for weeks; while the young medusae and seyphistomas would
thrive there an indefinite time, if there were a little pond ooxe at the bottom of the
aquarium and the water were changed twice a day. Indeed the growth and multiplication
of the scyphistomas would proceed actively under these conditions. By keeping the larvae
in shallow dishes 1 was able to watch the whole course of non-sexual development ; but the
development from eggs remains unknown to me because of the impossibility of finding any
that would develop. It was not until a few days before we left Jamaica in 1891 that I
discovered the habit that the very young free-swimming larvae have of hiding beneath the
bits of bark and the like to which the scyphistomas in the aquarium were attached, and
therefore the greater part of my work on the early stages of development was done during
the second expedition.

After a few words concerning technique I shall begin with a systematic description of
the species, followed by an account of the anatomy of the adult, and the remaining part
of the paper will contain what I have learned of the development from the observation of
the living animals while in Jamaica, and by the study of sections of preserved material,
carried on chiefly at the Biological Laboratory of the Johns Hopkins University, but in
part also at the Marine Biological Laboratory of Woods Hole and at the Biological
Laboratory of the U. S. Fish Commission in the same place.

I wish to express my thanks to Professor W. K. Brooks for the advice and encourage-
ment that he gave me while I was doing this work as one of his students, and I am also
indebted to Professor C. 0. Whitman, to the Hon. George M. Bowei-s, and to Professor
H. C. Bumpus for the many courtesies received while at Woods Hole.

TECHNIQUE.

For the preservation of the very young larvae a one quarter saturated solution of
picric acid, with 2% of sodium chloride added, gave good results. Erlicki's fluid with the
same addition, and J% osmic acid followed by Erlicki's fluid, did fairly well for scyphis-
tomas, but the best specimens obtained were those killed in the following mixture :

10% solution copper sulphate 100 c. c.

saturated solution corrosive sublimate 10 c. c.

As soon as they were killed, the specimens were placed in 5%, bichromate of potassium
and left there until hardened, after which they were washed in 35% alcohol containing a
trace of hydrochloric acid and preserved in 70% alcohol. Excellent preparations of the
medusae were obtained by this same method, and Flemming's fluid also gave good results.

194 ROBERT PAYNE BIGELOW ON

SYSTEMATIC PART.

Genus Cassiopea PERON and LESUEUR (1809). — This genus, as limited by
Haeckel ('80) , may be defined as follows : Discomedusae without tentacles and without a
central mouth opening ; provided, instead of the latter, with numerous oral funnels at-
tached to the ventral, or axial, side of the eight oral arms, which are pinnately or tricho-
tomously branched, have a subcylindrical, or subconical, gelatinous support continuous
to the tips of the principal branches, are provided with numerous club-shaped vesicles
among the oral funnels, and are without appendages on the dorsal, or abaxial, side ;
also with four interradial gonads in the aboral wall of the four separate subgenital cav-
ities ; sixteen marginal sense organs (rhopalia) ; and thirty-two radial canals connected
by a network of anastomosing branches.

For the sake of clearness this definition is made to include the characters of the
family Toreumidae Haeckel, to which this genus belongs, and the purely generic charac-
ters are italicized.

Cassiopea xamachana

Cassiof>ea xamachana BIGELOW, Zool. Anzeiger, no. 393, 1892, pp. 212-214.

(?) C.frondosa FEWKES, Bull. mus. comp. zool., vol. 9, no. 7, 1882, pp. 254-259.

Diagnosis. — A Cassiopea with a disc-like umbrella, concave on the aboral side;
when regular, with eighty short and obtuse marginal lobes separated by deep grooves on
the surface of the exumbrella (in each of the sixteen parameres three velar lobes between
two ocular ones) ; white markings on the exumbrella, consisting of a circular band with
a diameter somewhat greater than that of the concavity, within this sixteen oval or elip-
tical spots lying in the radii of the rhopalia, and on the outer side eighty marginal spots,
one for each marginal lobe ; oral arms rounded and slender, never angular, exceeding the
radius of the umbrella by at least one half of its length, and bearing nine to fifteen
primary branches which are, in turn, copiously branched, giving the whole appendage a
spatulate outline ; very numerous small oval vesicles attached at the axils of the small
branches and thickly massed upon the oral disc of adult females, and many small and a
few large, flattened, linear vesicles attached one at the axil of each of the larger branches

'This name, suggested by Professor Brooks, must stand as printed in the preliminary description of the species,
according to the cunvnt, rules of nomenclature, followed by the Boston Society of Natural History. But it should have
been written xinnnn^itui. from A'liniaycn (the a; is pronounced like ch in the German rich), the Indian name for the island of
Jamaica, as written by the early Spanish historians (see Hen-era, Novi orbis pars duodeeima, sive descriptio Indiaa occiden-
talis, 1024 ; also Encycl. Brit., Pth ed., article, Jamaica). The form Xaymaca given by Bridges, Annals of Jamaica, 1827,
and followed by several subsequent authors, is probably a misprint.

CASSIOPEA XAMACHANA. l!i:,

and to the canals on the oral disc, the thirteen largest vesicles being one at the axil of
the largest branch on each arm and one at each junction of the canals on the oral disc ;
oral funnels entirely wanting on the oral disc in adult females, but present in immature
specimens.

Special description. — A detailed account of the anatomy of this species will be given
in the anatomical portion of this paper. It is intended here to call attention merely to
the features that distinguish our species from its nearest allies. Cassiopea xamachana
resembles very closely two medusae that inhabit the Red Sea and Indian Ocean, Cassiopea
iiiiili-iiiiicda Eschscholtz and C. polypoides Keller ('83) , but it seems, nevertheless, to be
distinct from either.

Upon comparison with the descriptions of C'axxiopea andromeda given by Tilesius
('29), Haeckel ('79), and Vunhoffen (88), and with the figures of Tilesius ('29) and
Forskal (1776), C. .raiiiachanti appears to differ from this species in .the following
particulars : The exumbrella is not merely flat, but is concave ; besides the ninety-six white
spots on the exumbrella, there is a broad circular band of white more or less connected
with all of the marginal spots (Fig. 35), the oral arms are more thickly branched and are
longer, exceeding by one half to two fifths the radius of the umbrella, instead of being
only one third longer ; moreover the arms have none of the flattened appearance figured
by Tilesius and mentioned by Haeckel.

C. xamachana differs from C. polypoides in having more slender oral arms, with five
to seven pairs of branches instead of three, and with fewer very large vesicles, and these
apparently not so large and always flattened. The color pattern in the two species is
nearly the same, except that in C. xamachana the three white spots on the three velar
lobes of each paramere are seldom widely separated from the circular band of white.
The colors in the pattern, however, differ considerably in the two species. The ground
color in C. xamachana is never light brown, but is always much darker, a greenish brown,
usually with a distinct shade of blue on the subumbrella. The large oral vesicles are
never sky-blue, rose-colored nor white, but are yellowish green, often with a bluish green
stripe ; and, while the margins of the oral funnels are deep brown, they are always fringed
with the white digitella,

C. xamachana is easily distinguished from C. ornata Haeckel by the presence of
large oral vesicles and by the more extensive branching of the arms ; and it differs from
C. mertenxii Brandt ('38), C. depressa, Haeckel ('80), and C. picta Vanhoffen ('88), in the
number of marginal lobes on the umbrella. C. ndrosia Agassiz and Mayer ('99) differs
also in number of marginal lobes and in coloring. Fewkes ('82) has described a medusa
from Key West and the Tortugas \inder the name " Cassiopea frondosa Lamarck," which
he regards as identical with Poly clou in frondosa Agassiz. From the description given by

ROBERT PAYNE BIGELOW ON

Fewkes it is impossible to identify his species positively. But a comparison of his
figures with living specimens of both sexes of P.frondosa shows at once that the two
species are distinct ; while a comparison with C. xamachana shows so close a resemblance
that I am inclined to think that Fewkes has discovered one of the varieties of our species,
described in the next section. Not only is P. frondosa perfectly distinct from C. xama-
chana, but I think we are justified in retaining the former, for the present at least, in a
separate genus ; and there can be little doubt that Lamarck's Cassiopea frondosa dwell-
ing in the " Ocean of the Antilles " with its " maryine decem-lobata " is none other than
Agassiz's Polyclonia frondosa. Therefore, even if it should be proved that the form
described by Fewkes is the same as the subject of the present memoir, the name that I
have given to it will hold, nevertheless, as the designation of the species.

Variations. — If we compare the average dimensions of various organs, expressed in
thousandths of the diameter, with the maxima and minima, as may be done by examining
the third, fourth, and fifth columns in Table 1, p. 201, it becomes evident that there is a
very considerable amount of variation in the relative size of parts of C. xamachana.

In the oral arms, not only does the relative size vary, but the number and the
arrangement of the branches are both variable. Moreover this variability exists between
the different individuals. In nineteen specimens examined the maximum number of
branches found on one arm was sixteen, the minimum nine, and the greatest difference
on any one individual was four.

The most striking variations in C. xamachana, however, are to be found in the
structures at the margin of the umbrella. These are highly variable in this species, and
have been found to be variable, although to a less extent, in other medusae. It is
unfortunate, therefore, that in his beautiful systematic work on the medusae Haeckel
should have found himself forced to distinguish the genera chiefly by differences in the
marginal structures. He himself notes the variability in the number of parameres of
Pofi/r/onif( froiidoxd. Agassiz and Mayer ('99) found in one specimen of C. ndrosia
eighteen rhopalia, and in another twenty-two.

The number of rhopalia was counted in twenty-seven specimens of C. xamachana. Of
these ten were found to have sixteen, the typical number for the genus, and twelve had
more than sixteen, three having seventeen, and three more, eighteen. The largest
number on one individual was twenty-three. There were five specimens with fewer than
sixteen rhopalia, but only two had less than fifteen, and both of these showed correlated
abnormalities in the mouth parts and subgenital cavities. One had fourteen rhopalia, four
oral arms, and two subgenital cavities and gonads. The other had ten rhopalia, only five
oral arms, with three oesophageal canals leading from the stomach to the canal system of
the arms, and three normal subgenital cavities and one very small vestigial one (Fig. A.) .

CASSIOPEA XAMACIIANA. |;,7

Redundancy of mouth parts is not nearly as common as of the marginal structures.
Only two cases were observed. One specimen with an additional pair of oral arms in one
interradius had seventeen rhopalia. The other had eleven arms, with five snbgeiiital
spaces and gonads, and this one had twenty-two rhopalia. On the other hand, live speci-
mens were found with twenty or more rhopalia, and perfectly normal mouth parts.

It will be seen, then, that the number of rhopalia, which has been taken as the
principal generic character in the group, is a highly variable one. The number of

Fig. A. Section through the stomach of a specimen with only 10 rhopalia and 5 oral arms, to show the abnormal
arrangement of gonadia and oesophageal canals. In the region marked a the margin of the umbrella presents a wide
space in which there are no rhopalia. go = gouad. For explanation of the other lettering see Explanation of Plates.

marginal lobes in each paramere has been taken as one of the principal specific characters,
and this is, likewise, highly variable. The variation consists principally in the interpolation
of a small secondary lobe between two typical ones. Even in a regular and typical
specimen, such as is shown in Fig. 35, the position that would be taken by these secondary
lobes is indicated by small ridges on the dorsal surface. A specimen with a large number
of rhopalia is as likely to have the marginal lobes in each paramere arranged t^ypically as
one having a smaller number. Conversely, a specimen with fifteen to seventeen parameres
is as likely as not to vary from the typical form. The variation may consist in the addi-
tion of two secondary lobes in the paramere, the addition of four lobes, or in a quite
irregular arrangement ; and this modification may affect all of the parameres alike or only
a portion of them.

Throughout all of these modifications of the margins there is manifested a constant
regard, as it were, for the symmetry of the parts. It is very seldom that an additional
rhopalium appears as if attached fortuitously in some irregular way. Almost always either
it is in the midst of an entirely new paramere or else there is a distinct line of symmetry
running between two adjacent rhopalia that evidently correspond to an originally single
one. In other words, a paramere has been incompletely doubled, and the two parts are

198 ROBERT PAYNE BIGELOW ON

bilaterally symmetrical to one another. We find all degrees of this doubling in the adult
from a double-headed rhopalium * to two complete parameres, and the same process may
be seen in the forked tentacles frequently found in the larvae (Figs. 14 and 21, x) . With
the exception of the forked tentacle and the double-headed rhopalium, these stages of
duplication are well represented at u, w, x, y and z in Fig. 30.

The radially arranged stripes and spots on the exumbrella, which, with a circular
band, form the color pattern described in the next section, vary in number with the
rhopalia and marginal lobes. But when two rhopalia are close together there may be
only one corresponding rhopalial stripe, and it will then occupy a position intermediate
between the two. For example, in the specimen mentioned before as having only ten
rhopalia, two of the rhopalia were very close together and there was but one rhopalial
stripe corresponding to them. The other rhopalia were evenly spread, except that they
were absent from one rather wide section of the circle. The rhopalial stripes, nine in all,
were placed in a corresponding manner, and were absent from the corresponding area.

There is also a wide degree of variation in the extent of fusion between the circular
band and the marginal spots. The spots on the velar lobes are usually not fused to the
circular band in young specimens, and they frequently remain distinct in adults. It
was found, however, that this is more usually true of specimens from the Salt Pond than
of those from Port Royal. I thought that I could see, also, correlated differences in the
sizes of certain of the mouth parts, and I was thus led to inquire if there were a division
here of the species into two races. For this purpose Table 2, p. 201, was constructed.
From this it will be seen that the specimens from the Salt Pond (var. ^1) have on the
average longer oral arms and shorter vesicles than the ordinary specimens from Port
Royal (var. B) , while the stomach is of the same size in the two groups. Whether these
slight cliff ei-ences are in any way connected with the probable difference in density of the
water in the two localities, experiment alone can determine.

In the third column of this table dimensions are given of some specimens from Port
Roj'al (var. C) that are so different from the rest that they might be regarded as of a
distinct species.2 Suspicion of their being merely sports is aroused, however, by the fact
that only two specimens (female) were found living among a large number of the
usual form.

The most striking peculiarity of these two specimens was the great number (forty
to fifty) of uniformly large oral vesicles, two to four centimeters in length, scattered over

1 Fewkes, ('82) has observed similar double-headed rhopalia, and it is on account of the variability of the marginal
structures that he regards Polyclonia as merely an abnormal Cassiopea.

8 It is possible that this variety may be the same as Cassiopea frondosa Lamarck of Fewkes, although his figures do not
show any large vesicles on the proximal parts of the oral arms.

CASSIOPEA XAMACIIANA. 199

the whole extent of the mouth parts. Among them the central, radial, and primary
vesicles were hardly distinguishable, although so easily recognized by their greater si/e
in the typical form of the species.

Another peculiarity was a projection of the mesogloea on the oral side of each
subgenital osteum, so that the interradial diameter of the oral disc was considerably longer
than the radial diameter, as shown in the table. These specimens presented also some
peculiarities of coloring, which will be noted in the next section.

Color. — The coloring of this semi-transparent animal consists of certain white mark-
ings, together with shadings of subdued tints of brown, green and blue, that are often
very beautiful.

If we turn the aboral side (Fig. 35) of the medusa toward us Ave find often a brown-
ish band encircling the disc at the periphery of the concavity and shading off gradually
on both sides. Deeper within the mesogloea there is a much wider white circular band
extending under the brownish one ; and white bands, or spots, extend in a radial direction
outward from this along the marginal ridges. There is one spot to each ridge, and it
reaches nearly to the tip of the marginal lobe. The bands on the rhopalial lobes are inter-
rupted, however, by a roughly circular, transparent area over each rhopalium ; and in
many specimens, especially young ones, the interrhopalial (velar) spots are not fused with
the circle. On the inner side of the circular band of white there is a circle of large white
spots, ''rhopalial spots," or stripes, that lie deep in the substance of the exumbrella and
are visible through the mesogloea, one in the radius of each rhopalium. The spot is
elliptical in outline, and extends from the white band to a point about two fifths of the
distance between the periphery of the concavity and the edge of the stomach. These
spots, while usually continuous with the band, like the marginal spots, are not always so.

At the centre of the umbrella the stomach and subgenital cavities may be seen
through the mesogloea as a reddish brown circular area with a diameter of about one
fourth of the total diameter of the disc ; while surrounding the stomach there is a deep
blue halo with points that extend outward between the last-mentioned bands of white.

Now if the animal be allowed to return to its usual position, the subumbrellar surface
will be found to be pretty evenly stippled by the greenish brown cells in the mesogloea.
Apparently beneath this stippling there is a blue pigment forming a circle around the
margin of the stomach and extending outward in broad bands, one along each interrho-
palial radius, nearly or quite to a large, more or less distinct patch of blue, that lies close
to the margin between every two rhopalia. The radial canals, and the fine, connecting
network of tubes, appear as rather indistinct, opaque, white lines.

The mesogloea of the oral arms is transparent and colorless, except for an opaque
white stripe beneath the dorsal surface of each arm, of the same character as the white

200 ROBERT PAYNE BIGELOW ON

markings of the umbrella. There is a similar stripe on the dorsal side of each of the
larger branches which may, or may not, be continuous with the stripe on the main stem.
The bases of the oral funnels are of a delicate blue color, which often extends to the
brachial canal. The margin of each funnel is a deep brown, that shades off over the
blue ; while the small tentacles, or digitella, that spring from this margin are pure white.
The larger tongue-shaped vesicles on the arms and oral disc -have a greenish yellow color
with a bluish green longitudinal stripe. The smaller vesicles on the arms are colored in
a similar way and are inconspicuous, but the cluster of very small vesicles that occupy the
greater part of the oral disc has a very different appearance, being lightly tinted by fine
reddish brown pigment spots.

The two specimens that I have called variety (7, are somewhat differently colored.
The markings are yellowish white. The circular white band is indistinct. The rhopalial
bands are interrupted at the margin of the concavity of the exumbrella, and stop short
about half a centimeter from the rhopalial hood. At the margin of the umbrella there is
a white spot on each rhopalial lobe and a strap-shaped spot on each velar lobe. The
centre of the umbrella is whitish and opaque, so that the stomach does not show through.
The oral arms are translucent, milky white tinged with brown, and without distinct wlu'te
markings, except on the dorsal side of the principal branches. The large oral vesicles
are yellow and greenish yellow, with a brown centre.

Size. — The diameter of the largest specimen found is 24 cm., while the smallest
specimen that contained eggs measured G.5 cm. The average diameter of twenty-three
adult specimens was 13.7 cm. The relative sizes of the parts are exhibited in the follow-
ing tables. Table 1 shows what may be regarded as the normal proportion for the species,
though, to be sure, some of the measurements were made only on Port Royal specimens.
Some of the dimensions were measured on a smaller number of specimens, and for these
the average diameter of the umbrella of the specimens on which these measurements
were made is given separately to show the correct proportions. Table 2 furnishes a
means of comparing the proportions in the three varieties.

CASSIOPEA XAMACIIANA.

TABLE 1. DIMENSIONS OF SPECIMENS FROM BOTH SALT POND AND TORT ROYAL

(VARIETIES A AND 7i.)

Number of

Average di-

Average dimensions

Maximum Minimum

spcvnm'iis
nu-asured

mensions in
centimeters

in thousandths of
average diameter

(In thousandths of diameter
unless otherwise noted)

Remarks

Diameter of umbrella

21

14.33

1000

24 cm.

9 cm.

Length of arms meas-

ured from centre of

Salt Pond and

oral disc

•21

11.21

782

888

642

Port Royal

Length of central

specimens tak-

vesicle

21

2.71

189

287

103

en together.

Length of primary

vesicle

21

1.82

126

190

71

Diameter of umbrella
Length of radial
vesicle

10

10

15.79
3.01

1000
190

24 cm.

250

12 cm.
117

Port Royal

only-

Diameter of umbrella

11

13.62

1000

16.3 cm.

6.3 cm.

Salt Pond and

Diameter of stomach

11

4.22

311

333

292

Port Royal.

Diameter of umbrella

3

18.66

1000

24 cm.

15 cm.

Thickness of umbrella

3

1.33

71

73

70

Port Royal.

Total depth

3

3.59

191

212

173

Diameter of umbrella
Diameter of oral disc

o

2

20.50
8.00

1000
390

24 cm.
441

17 em.
354

Port Royal.

TABLE 2. AVERAGE DIMENSIONS OF THE THREE VARIETIES, GIVEN IN THOUSANDTHS
OF THE AVERAGE DIAMETER OF THE UMBRELLA.

Variety A— Salt Pond

Variety R— Port Royal

Variety C— Port Royal

Length of arms

850

715

818

Length of central vesicle

157

217

Radial vesicle

190

163

Primary vesicle

112

138

167

Diameter of stomach

298

311

322

Thickness of umbrella

71

80

Total depth

191

213

Radial diameter of oral disc

390

394

Interradial diameter of oral disc

390

470

Locality. — Great Salt Pond, and mangrove swamps (" The Lakes") in the rear of
Port Royal, Jamaica. — BIGELOW.

(?) Moat outside Fort Jefferson on Garden Key, Tortugas Islands, and Mangrove
Keys, near Key West, Florida. — FEWKES.

202 ROBERT PAYNE BIGELOW ON

ANATOMY.

Form of the Body. — To one who is familiar with the cyaneas, aurelias, and the
like, of our northern coast, the shape of this medusa appears very strange. The aboral,
or exumbrellar surface (Fig. 35), instead of being convex in Gassiopea xamachana, as
it is in the great majority of medusae, is concave when the animal is at rest, except for a
slight convexity over the stomach, and except in the region of the thinner marginal part
of the umbrella, where also it is convex. The surface of the subumbrella, on the other
hand, is convex, except in this same thinner marginal area, where it is in turn concave.
The umbrella thickens very gradually from its margin to the centre, and the elevations
and depressions of its surface have very gentle slopes, so that its general shape is much
nearer that of a flat disc than the dome-like form of most medusae (Fig. 35).

A circular column arises from the centre of the oral surface of the umbrella. It is
broad, but very short; and a few millimeters from the umbrella it loses its circular
outline, owing to eight stout arms that spring from it at regular intervals (Fig. 34).
These are smooth and rounded, except along a line on the oral side, where they bear
the fringe of oral appendages, and they are long and much branched. This column is
the oral disc, and its arms the oral arms.

TJie Structure of the Mesogloea. — By far the greater part of the mass of the oral
arms and disc, as well as the umbrella, is composed of a firm, elastic, gelatinous
substance, the mesogloea, and it is to this that the shape of the body is due.

The description given by Keller ('83), of the structure of the mesogloea in G.poly-
poides would apply almost equally well to our species. The mesogloea consists of a
hyaline matrix, in which are imbedded certain fibres and three kinds of cellular elements.
Most of the fibres appear to be analogous to connective tissue fibres, and take a general
course through the mesogloea at right angles to the surface. Others seem to be proto-
plasmic. At any rate, they may be observed to proceed from the star-shaped cells that
are scattered throughout the jelly.

The cellular elements are: the star-shaped cells, just mentioned; vesicular bodies,
found in certain restricted localities ; and the green cells, which, as it will be shown later,
are symbiotic plants.

The star-shaped cells remind one of osteoblasts, and are probably analogous to them,
in that they are concerned in the formation of the jelly. Hamann ('8l) has called
them colloblasts. They are small, often somewhat elongated, and have a well-marked
nucleus.

CASSIOPEA XAMACIIANA. 203

The vesicular bodies give rise to the white markings that were mentioned in the
description of the species. These vesicles are much larger than the colloblasts. Each
one seems to be made up of a wall of exceedingly minute refractile granules, surrounding
a clear space. That this body is a cell, however, is shown by the presence of a nucleus
pressed closely against one side.

The green cells, or zoanthelae, are widely distributed throughout the mesogloea, but
are most abundant in the umbrella. They are not infrequently found imbedded in the
eudodermal epithelium. The living cells have a greenish brown color, which they
impart to the animal as a whole. They are of considerable size, are globular,
without projections of any kind, and are generally to be found in clusters of two or more
(za, Figs. 52, 56, and 63). Each one contains a nucleus and numerous granular bodies,
and apparently is surrounded by a cell wall ; but the latter is hard to distinguish from
the edge of the adjacent matrix.

Keller thought that similar bodies in C. pal ypoides could not be algae, because he
failed to find any evidence of a cellulose cell wall. He regarded them, therefore, as
essential elements of the " mesoderm."

In those of my specimens, however, which have been killed in Erlicki's fluid and
stained with borax carmine, the nucleus of these cells is found to be red, while the gran-
ular contents of the cell are bright green, and there are often one or two green bodies
present that are as large or larger than the nucleus. They have all the appearance of
chlorophyl bodies, and it is well known that the chlorophyl of some algae is not readily
removed by alcohol. Moreover, in teased preparations treated with iodine solution these
cells are found to be filled with granules that quickly turn deep blue, — evidently starch.
The test for cellulose with iodine followed by sulphuric acid, gave, however, unsatisfac-
tory results. The outline of the cell would become distinctly darker, but not recognizably
blue. In the same way with chloriodide of zinc, a very marked reaction for starch was
obtained, the granules becoming almost black ; but so long as the object was viewed
by direct transmitted light, no reaction for cellulose could be detected with certainty.
On the other hand, when the light was thrown upon the object obliquely by means of a
condenser with a central diaphragm, 'the effe'ct was quite different. The starch granules
became a deep ultramarine, and the parts of the cell not occupied by the starch appeared
distinctly violet, showing without doubt the presence of cellulose.

Both starch and cellulose, as well as some form of chlorophyl, having been demon-
strated in them, there can be no further question that the green cells in Cassiopea are
symbiotic algae.

The Oral Arms and their Branches. --The eight oral arms (Fig. 34) arise from
the central oral disc at about equal intervals; and when an arm is extended, the distance

204 ROBERT PAYNE BIGELOW ON

from the centre of the oral disc to the tip of the arm about equals three quarters of the
diameter of the umbrella. But the arms .are very contractile, and may be shortened to
half this length. The arms are slender and graceful in shape, the mesogloea tapering
very gradually to the tips of the finest branches. The branches are arranged alternately. "
The largest one, which is the first formed, is at a point about two thirds the length of
the arm from its base. From this point the branches decrease in size gradually toward
the base of the arm, and more rapidly toward the apex. The general outline of the
arm, therefore, including its branches, is roughly spatulate.

The Oral Funnels and Brachial Appendages. — Just below the surface of the
oral side of each arm there is a longitudinal tube, the brachial canal, that ramifies to
each branch, and finally opens to the exterior by funnel-shaped oscula (os., Fig. 34) at
the tips of the numerous ultimate branches, and at many places along the course of the
tube. The margins of these oscula, or oral funnels, are provided with short tentacle-like
projections, the digltella. These are covered by an epithelium containing nettle cells,
and each has a gelatinous axis in which there are transverse plates of greater density
than the rest of the jelly, and these give the structure the cellular appearance first
described by Hamann ('81) . The epithelium lining the tubes and funnels is
ciliated.

There open also into the brachial canals the lumina of the oral vesicles (v., Fig. 34) .
These structures, as already stated in the diagnosis, have their points of attachment in
the axils of the branches. All except the smallest are flattened laterally. The smaller
ones are oval in outline, the larger ones linear. At one side near the apex there is a
cluster of short processes that Hamann has homologized with digitella.

The Oral Disc. - - Although the eight oral arms seem to be placed at equal
distances and to be alike, they are morphologically in pairs, each pair being homologous
to one of the. four lips of a semostomous medusa, — an aurelia, for example. The line
that separates two members of a pair is therefore, according to Haeckel's nomenclature
a perradius. The brachial canals from each pair of arms, on entering the oral disc,
converge and unite into a single radial tube that is continued to the centre of the disc,
where it unites with the other three. In this way the course of the tubes on the oral
disc forms a pattern that resembles a Maltese cross. The larger central vesicle is
attached at the centre of the cross. In a living specimen 11 cm. in diameter this
measured 3 cm. in length. There are four other vesicles that most nearly approach the
central one in size, and these arise from the radial canals near the junction of the
brachial canals, and I have called them, therefore, the radial vesicles. In full-grown
individuals there are eight more vesicles upon the oral disc, a little smaller than the last,
one on each brachial canal distal to the junction. It is only near the periphery of the

CASSIOPEA XAMACIIAKA.

disc that the canals are provided with oral funnels. For most of their course on the disc
the canals give rise to the very small vesicles, finely speckled with a reddish brown
pigment, that already have been mentioned. These have nettle batteries at their tips,
and are so numerous as to cover completely the greater part of the disc and to hide the
course of the canals. This mass of small vesicles, however, is not acquired until late.
Specimens as much as 6 cm. in diameter will be found to be without them. In such
specimens we have the five largest vesicles, and a number of oral funnels are scattered
along the canals, just as they are upon the arms. This replacement of the oscula on the
oral disc by small vesicles has been observed to occur also on adult females of Polyclonia
frondosa, but not in the males (Bigelow, '93). It is not improbable therefore that a
similar difference between the sexes may be discovered in our species of Cassiopea.

The Subgenital Cavities and the Digestive Tract. — At each of the four points
of junction of the brachial canals there is a slit-like passage, oesophageal canal, dipping
vertically into the mesogloea of the disc, and opening into the stomach. The latter is
a lens-shaped cavity, with a gently arched roof. Its floor consists chiefly of four lozenge-
shaped areas, where the body wall is very thin and plaited in radial folds (Fig. 34) .
These thin parts of the body wall form the roofs of the subgenital cavities, which open
to the exterior, each by an elliptical orifice, osteum, (x, Fig. 34) in the side of the oral
disc near the subumbrella and in the angle between two pairs of arms (inter radial) .
The gonad appears as a band which crosses this membrane tangentially at its greatest
width. Just central to each gonad there is a multiple series of very many small gas-
tric filaments forming a narrow band parallel to the ovary. These are ciliated, and
provided with nettle and gland cells. The portion of the floor of the stomach not made
up of these lozenge-shaped membranes is bounded by the firm mesogloea of the oral disc.
This area has the shape of a Maltese cross, and it is in the arms of this cross, between the
subgenital cavities, that the passages from the oral canals open into the stomach.

Near its periphery the floor of the stomach is marked by radial grooves. These are
continued, each into one of the radial canals that extend outward from the edge of the
circular stomach to the marginal region of the umbrella. There are regularly thirty-two
of these, sixteen in the radii of the rhopalia, and sixteen interrhopalial. When the
number of rhopalia is increased, the number of radial canals may or may not increase in
proportion. There are often thirty-four or thirty-six of them. The canals in the
radii of the rhopalia are larger and more nearly straight than the interrhopalial ones, and
all are connected by a fine network of anastomosing branches, among which no distinct
circular canal can be recognized. The meshes in the network of canals are connected by a
plate of endodermal cells, the endodermal lamella. This lamella is also in contact with the
subumbrellar ectoderm along a line encircling the umbrella a short distance from its mar-

200 ROBERT PAYNE BIGELOW ON

gin, so that there is a complete sheet of endoderm separating the subumbrellar from the
exumbrellar mesogloea.

Musculature. — The exumbrella is devoid of muscles, but on the opposite side
there is a continuous sheet of muscle fibres, which is spread over the subumbrella,
except a narrow zone at its margin, and is continued over the oral arms to their finest
ramifications, and also into the subgenital cavities.

Most of the fibres on the subumbrella, do not take an evenly circular course, but are
undulating. They form in this way a series of double " arcades " like those found by
Haeckel in other species of Cassiopea. There is one of these double arcades for each
interrhopalial space. The surface of the mesogloea in this region is grooved. The
sheet of muscle fibres lies directly upon it, and is therefore corrugated, the grooves
being parallel with the fibres.

The muscular layer upon the oral arms is smooth, and its fibres take a longitudinal
course, extending to the digitella and oral vesicles.

In the subgenital cavities the arrangement of the muscle fibres could not be made
out; but their presence was revealed by the squirming movements of the thin membrane
that separates the subgenital cavity and bears the gonacls and gastric filaments.

Structure of the Marginal Sense Organs. — Each rhopalium has a pigment spot
on the aboral side near the extremity, and each one lies in a deep sensory niche. The
dorsal sensory groove, common in the Pelagidae, Aurelia, etc., is entirely lacking;
although Keller found in C. polypoides a slightly depressed thickening of the ectoderm
that corresponds to it. The sensory niche and rhopalium are, with the exception of the
pigment spot, similar in all essential particulars to those found in Pelagia. The
rhopalium is the only organ in the sensory niche (Fig. 56). It is a hollow, finger-like
projection attached by its base to a low ridge that runs along the roof to the central wall
of the niche. This ridge is penetrated longitudinally by the continuation of a radial
canal from the stomach, and the lumen of the rhopalium opens into the distal end of this
canal. In the distal half of the rhopalium the lumen is nearly obliterated by the increase
in thickness of its endodermal lining. Here the endoderm, instead of being a columnar
epithelium as elsewhere, is a mass of parenchyma-like cells, each of which contains a
large calcareous concretion, a so-called otolith. A thin, supporting membrane separates
the endoderm from the ectoderm. At the distal extremity of the rhopalium the ecto-
derm is a thin, cuboidal epithelium, while over the rest of the surface it is a thick, sensory
epithelium resting on a thick network of fine nerve fibres. This, in turn, rests on the
supporting membrane. I have observed no ganglion cells in this layer of nerve fibres,
which is continued under the epithelium of the rhopalial ridge to the central wall of the
niche, where it becomes imperceptible. There are no thickened bands of these fibres

CASSIOPEA XAMACHANA. 207

running to ciliated pockets, such as are found in Dactylometra (Bigelow, 590), and the
fibres probably spread out finally into a thin network underlying the general epithelium
of the subunibrelhi.

The one feature in which this rhopalium differs from what is found in Pelagia is the
presence of the pigment spot, already mentioned, lying on the aboral side of the
rhopalium immediately above the centre of the mass of concretions. This area is
probably sensitive to light, but it only differs from the rest of the sensory epithelium in
that here the superficial cells are deeply colored by a yellowish brown pigment. A
more careful examination would undoubtedly show the histology of this structure to be
similar to what Schewakoff ('89) has found in Aurelia.

HABITS.

The species of Cassiopea and the closely related genus Polyclonia find their habitat
usually in quiet lagoons among the mangroves along the shores of the tropical seas. The
mode of life of several species has been described by Brandt ('38) on the authority of
Mertens, L. Agassiz ('62), Gray ('69), A. Agassiz ('si), Archer ('8l),Fewkes ('82), Guppy
('82), and Agassiz and Mayer ('99). Cassiopea xamachana is no exception to the rule
either in its habitat or its sedentary mode of life.

When the young medusa is set free from the strobila it is an active swimmer. It
gradually becomes less active as the mouth parts acquire their adult structure, and by
the time the animal has reached a diameter of two centimeters it has definitely taken
up its abode upon the bottom. It lies there, as described in the Introduction, with the
oral appendages upward, and seldom changes its position unless disturbed. The con-
cavity of the exumbrella is an important aid in maintaining this posture against the
action of waves and currents. The gelatinous tissue is firm and elasti-c, and causes the
umbrella to assume its normal shape when the subumbrellar muscles are relaxed. The
slight suction thus produced when the medusa comes to rest on a flat surface gives it
such a hold that a certain amount of force is required to remove it.

Usually, however, the water in the lagoons is very quiet, and there is more danger
from its stagnation than from its motion. A Cassiopea is enabled to avert this danger by
the slight swimming movements of the thinner marginal part of the umbrella. By
ineans of these rhythmic contractions the water is drawn in on all sides, and then is
driven upward and away. A healthy specimen lying undisturbed on the bottom of an
aquarium was observed during seven minutes to make on the average 19.7 contractions
of the umbrella per minute. In this way the animal is enabled to draw to itself a fresh
supply of oxygen and of its minute food material,

208 ROBERT PAYNE BIGELOW ON

The oral arms and their branches are usually spread out so as to cover the sub-
umbrella completely, but they are almost always in motion, bending to one side or the
other, and they may be flexed aborally until the tips come within the umbrellar
margin, or extended until they reach far beyond. Besides these general movements,
the various appendages of the oral arms have movements of their own. Muscular con-
tractions may be observed also in the thin membrane that separates the stomach from
the subgenital cavities, and they probably serve to renew the water that bathes the
gonads.

When the oral disc has been severed from the umbrella, both parts may remain
alive for several days, and both retain their powers of spontaneous movement. The first
effect of the operation is often to throw all of the parts into a strong tetanus, but shortly
afterwards the pulsations of the umbrella may be renewed at a rate considerably more
rapid than normal, — 32 to 34 per minute in one case. When the medusae have been
for some time under unfavorable conditions, it frequently happens that the part of the
body-wall surrounding the periphery of the stomach is ruptured, and thus the mouth parts
as a whole become separated automatically from the umbrella.1

While at Bimini, Bahamas, in 1892, I observed that the food of Polycloniafrondosa
consists chiefly of copepods and other small Crustacea, and that these are caught by the
combined action of the oral vesicles and the oscula ('93, p. 106). If a copepod strikes
a vesicle, the vesicle bends quickly so as to cover the mouth of the adjoining oscu-
lum, and the copepod is thus enclosed in a trap. Artificial stimulation would cause the
same reaction.

Experiments on Cassiopea xamachana made to determine whether or not this
species obtains its food in the same way gave negative results. Stimulation of an oral
vesicle causes only a slight bending on the side stimulated. Although these vesicles are
provided with batteries of nettle cells near the tip, I was unable to see that they played
any part in the taking of food. It may be that they are protective, but the sting is
very feeble.

Examinations of the contents of the stomach were almost equally unsatisfactory.
The contents of twenty-two stomachs were examined, and of these fourteen contained
only a clear, very viscous fluid and, in some cases, a few apparently ripe eggs, more or
less distorted, some small, colorless cells, probably sloughed off from the endodermal
epithelium, and green cells identical with the " zoanthelae " found in the mesogloea. One»
01- more copepods and other Crustacea were found in five specimens. In one of these one
small amphipod was found, and in another the cornea of an unknown crustacean.

1 While this paper is passing through the press a paper has appeared by E. W. Berger (1900), in which he gives the
results of experiments made by F. S. Conant upon Polyclonia and Cassiopea to test the effect o£ removing the rhopalia.

CASSIOPEA XAMACHANA. 209

Diatoms and other algae, exclusive of the above-mentioned zoanthelae, were found
in four cases out of the twenty-two. Almost always the first impression on opening a
stomach was that it was empty; and it was only by careful examination that the contents
of the stomach could be discovered. In two cases, however, a considerable amount
of material was found in the stomach. In one of these the stomach contained, besides
the usual eggs, zoanthelae, debris, etc., the remains of many copepods, some nematods,
a zoea, and some diatoms. In the other one there was found a compact greenish mass,
about one centimeter in diameter, composed of granular debris and diatoms of various
species, together with some desmids, Oscillaria, foraminifera, infusoria, Vorticella, and
some fine filaments with spirally arranged contents.

This species exhibits the power of regeneration of lost parts to a marked extent.
Specimens were frequently met with in which branches of the oral arms, or even portions
of the margin of the umbrella, had evidently been formed recently to replace parts that
had been destroyed. Moreover, branches of the oral arms that had been cut off were
observed to regenerate oscula and vesicles at the central end.

ONTOGENY.

Historical Review. — Numerous studies upon the reproduction of various animals
by budding have shown that the formation of organs in the bud may take an entirely
different course from the development of homologous organs in the sexually produced
embryo. As the observations to be described in the sequel were made entirely upon
larvae that were observed, or supposed, to be asexually produced, they cannot settle
any of the disputed points in regard to the development of sexually produced larvae.
Nevertheless it will be of interest to compare the sexual with the asexual mode of
ontogeny, especially as the development of a scyphistoma from a bud has never before
been fully described.

The development of scyphomedusae from the egg has been studied in comparatively
few forms. Several species of Aurelia have been studied by Sars, Haeckel, Schneider,
Glaus, Goette, Frank Smith, and Hyde. Two species of Cyanea have been studied by
McMurrich and Hyde. A species of Chrysaora has been studied by Glaus; and
Kowalewsky, Metschnikoff, Krohn, and Goette have traced the very interesting abbrevi-
ated development of Pelagia noctiluca. The nearest ally of our species that has been
studied with any degree of completeness is the Mediterranean rhyzostome Cotylorhiza
tuberculata, which has been the subject of investigation by Glaus and Goette.

Prom the description of the process of budding given in the next section, it will be

210

ROBERT PAYNE BIGELOW ON

d -

t— d

noted that the larvae produced by budding are set free in a form resembling planulae,

and it is the planula stage in sexual reproduction which is the earliest that can be

compared with any stage to be described in this paper.

All agree that the planula of the Scyphomedusae is a more or less oval, ciliated

larva, somewhat flattened on one side. It consists of two layers of cells surrounding a

cavity which is completely shut off from
the exterior. The first step toward the
development of the scyphistoma is an in-
vagination of the ectoderm at the posterior
pole of the planula. According to Goette
and Miss Hyde, the endoderm is invagina-
ted at the same time in such a way that
two endodermal pouches remain in the
plane of the long diameter, — " Haupte-
bene," Goette, — (Fig. B, and 6-6 Fig.
D), one on each side of the invagination,
while the endoderm is pushed entirely
away from the oral pole in the plane of
the short diameter (Fig. C, and c-c Fig.
D) . The ectodermal invagination is the
oesophagus (Schlund) , its external open-
ing is the mouth, and the endodermal
evaginations are the first pair of gastric
pouches. Soon an opening is formed
(Schlundpforte) between the base of the
oesophagus and the central stomach, and
at the edge of the opening the ectoderm
fuses with the endoderm. Then the sec-
ond pair of gastric pouches is formed.
According to Goette these are produced in
Cotylorliiza and Pelagia entirely from the
ectoderm of the lower part of the oesopha-
gus (Fig. F) . According to his view the
lower edges of these evaginations coincide
with the margin of the gastro-oesophageal
opening (Schlundpforte) , and are therefore

at the level of the upper edges of the first pair of pouches. As these pouches extend out-

rigs. B-G. Two stages in the development of the mouth
and gastric pouches in a sexually produced scyphistoma,
according to Goette. Figs. B to D are sections in the three
dimensions of space of a larva in which the oesophagus is
invaginated and the first pair of gastric pouches are formed.
Figs. E to G are similar sections of a larva in which the open-
ing from the oesophagus into the stomach (Schlundpforte) 1ms
been established and the second pair of gastric pouches are in
the process of formation, b, c, d, etc., indicate the plane of
Figs. B, C, D, etc.

CASSIOPEA XAMACIIANA. 211

ward the ectoderm is pushed out until the principal radii) e-e and f-f, Fig. G) gradually
become equal. This results in the formation of the flattened peristomal disc (Figs. E
and F) . At the same time the wall separating the oesophagus from the first pair of gastric
pouches (Taschenvorhang) is split upward until the openings into the two pairs of pouches
are upon the same level, and the " Taschenvorhang " is reduced to a low ridge, while the
oesophagus is very much shortened. Portions of the original lining of the lower part of
the oesophagus persist as the covering of the inner edges of the septa which separate the
four gastric pouches. The larva is now in what Goette calls the scyphula stage. He
regards this stage as of great phylogenetic importance, showing clearly, he thinks, a
close genetic connection between the Scyphomedusae and the Anthozoa ; so that these
groups should be placed in a single class, Scyphozoa, to distinguish them, on the one
hand, from the Hydrozoa, including the hydroids, hydromedusae, and, on the other, from
the Siphonophorae.

This view, which is confirmed by Miss Hyde, is antagonized by Glaus, and has given
rise to a prolonged controversy. The parties to this dispute are practically in accord
in regard to the facts of observation, as represented by the figures in their latest
contributions. It is in the interpretation of these facts that they differ mainly. Clans
f 90) admits that there is an ectodermal invagination previous to the formation of the
mouth, and that the lining of the proboscis is ectodermal. He maintains, however, that
the condition represented in Figs. B, C, and D, is due to a severe contraction of the
animal, and is without morphological significance. He denies that there is any
oesophagus, " Taschenvorhang," or " Schlundpforte," in the sense that Goette uses these
terms ; and with these he rejects the idea of a close affinity between the Anthozoa and
scyphomedusae, with the correspondingly sharp distinction between the latter and the
hydromedusae. He admits, however, that a distinction of importance between scyphis-
tonia and hydropolyp is to be found in the possession by the former of an ectodermal
lining of the proboscis, and in the presence of gastric pouches and septa.

Goette's position has been strengthened considerably by his latest contribution on
this subject ('93), and his conclusions are confirmed in nearly every particular by Miss
Hyde. Goette's figures appear to be camera drawings of serial sections, and are a
great improvement over the rather diagrammatic illustrations in his earlier papers.
They are not entirely convincing, however, for the ectoderm and the endoderm grade
into each other so that it is impossible to determine the exact boundary ; and the
material is so subject to distortion during the preservation that it is often difficult to
determine whether a given fold of the epithelium is of morphological value or is merely
an artifact.

Miss Hyde fails to confirm the view that the whole of the second pair of gastric

212 ROBERT PAYNE BIGELOW ON

pouches (and hence, according to Goette, five eighths of the peripheral digestive tract of
the medusa) is of ectodermal origin. On page 550 of her paper she speaks of only the
roof, i. e., the lining of the oral side, of these pouches as formed from oesophageal ecto-
derm, and her sections bear out this view.

The foregoing brief summary of the present state of knowledge regarding the early
stages in the development of the sexually produced scyphistoma will serve as an
introduction to what is to follow. Mention of the work of others upon the later stages
will be made when we come to the corresponding periods in the development of
Cassiopea xamachana.

The Formation of the Bud. — In 1841 Sars described the budding of scyphistomas
that were supposed to belong to a species of either Aurelia or Cyanea. The buds, accord-
ing to this account, may grow out directly from the main part of the body of the larva,
or they may be produced on stolons extending outward from the foot. In either case,
several buds may be formed, apparently in various positions on the scyphistoma at one
time. The figures show the buds attached to the parent and provided with a well-
developed crown of tentacles at the distal end. Agassiz ('60) found a similar process of
budding to occur occasionally in Aurelia. Goette ('87) has confirmed these observations,
and has found that the larvae of Cotylorhiza tuberculata also produce buds.

In Cotylorhiza the bud is formed as an outgrowth from the body of the scyphistoma ;
and as it grows it gradually approaches the shape of its parent, but its relative position is
just the reverse of what Sars found ; for the distal end forms the stem, and the proximal
end begins to flatten, out into a circumoral disc. In this condition the bud is set free, and
swims about, rotating on its long axis, with its distal end forward. The moiith is formed
at the point where the constriction finally separates the bud from its parent, and the larva
fixes itself by the opposite end.

Glaus ('92) has found that the scyphistomas of Cotylorhiza not only produce buds,
but that they produce them in large numbers. Scyphistomas reared from eggs that had
been laid in September, 1890, were kept alive in the aquaria at Trieste and Vienna
throughout the following winter and spring. No change was observed after the larvae
had reached the sixteen-tentacle stage, until the following July, when budding occurred.
The process was not restricted to a few well-nourished individuals, but seemed to be a gen-
eral and repeated phenomenon, and it resulted in a large increase in the number of larvae.
Clans' s brief description of the formation and fate of the bud, accompanied by three
figures, corresponds perfectly with what I have to describe in the following pages. The
strobilization, he says, took place in August, and was monodiscous.

Another case of rapid multiplication of scyphistomas by budding is described by
Lacaze-Duthiers ('93) . A colony of these larvae, of unknown origin, was discovered in

CASSIOPEA XAMACIIANA. 213

an aquarium at Banyuls in 189'2. No strobila or ephyra was observed up to October,
1893, but in the meantime the number of. scyphistomas had increased enormously. The
buds appeared as elevations on the side of large individuals. The base of this elevation
became elongated into a filamentous stolon carrying the bud at its tip. The bud fixed
itself to the glass, gradually developed tentacles, and finally became independent by the
disappearance of the stolon.

In Cassi&pea xamachana the process of budding is an important, if not the chief, factor
in the perpetuation of the species. On looking over collections of scyphistomas taken
from the Great Salt Pond during May, June, and July, a considerable number was found
with buds attached in various stages of development (Figs. 1, 2, and 26), and budding
continued in the aquaria.

There is no stolon. The first visible rudiment of the bud is a slight swelling on one
side of the calyx just above where it tapers into the stem. It involves all three layers of
the body wall (Fig. 36) . At an early stage in the growth of the bud the four septal
muscles may be found as four slender cords of cells embedded in the mesogloea and appar-
ently growing out from a thickened area of the ectoderm at the apex of the bud (sm,
Figs. 38 and 39) . This appearance seems to indicate that the septal muscles of the bud
are formed, as in sexually produced scyphistomas, by an ingrowth of the ectoderm.

Careful study of serial sections through young buds shows, however, that the
septal muscles of the bud are connected with septal muscles of the parent. In
Fig. 37 the course of the septal muscles is reconstructed from a series of longi-
tudinal sections. The muscle sn^ passes around the base of the bud on the side
away from the observer and gives rise to two branches, sm\ and .sj;/;/a. These extend
toward the apex of the bud, and each one divides dichotymously. Muscle sw4 gives
off a branch which fuses with sm^ at the base of the bud. Sm2 produces three branches
which unite into a single branch. This branch extends into the base of the bud, but it
was impossible to trace it further. It may be connected, perhaps, with the muscle sm'.,,
which was traced for a short distance from the apex of the bud.

On account of the presence of the muscles, it is possible to study the relation of the
plan of symmetry of even young buds with the symmetry of the parent. A series of
sections made at right angles to the long axis of the bud shows that the vertical perradii
of the bud lie in the plane of one of the perradii of the parent. But in transverse
sections it is impossible to trace the muscles of the bud, except for short distances,
owing to their extreme fineness.1

However, the study of longitudinal sections makes it reasonably certain that the sep-

1 The position of the bud is always perradial, although in some preparations it appears to be interradial, owing to the
obliquity of the sections, as in Fig. 30,

214 ROBERT PAYNE BIGELOW ON

tal muscles of the bud are derived from outgrowths of one or both of the septal muscles of
the parent which lie in the interradii adjoining the perradial area of bud formation. If this
be true, then every part of the young bud is formed from the corresponding part of the
parent, viz., — ectoderm from ectoderm, mesogloea from mesogloea, endoderm from endo-
derm, muscles from muscles, and digestive tract from digestive tract. There is no indica-
tion of any method of budding of the kind described by Lang ('92) .

In the young bud the mesogloea is very thin, so that the ectoderm and endoderm are
very nearly in contact. The evagination gradually increases in size, becoming first hemi-
spherical and then more elongated. At the same time a constriction appears close to the
body of the scyphistoma, which deepens until the bud becomes a spindle-shaped body
attached to its parent by a short and narrow stem consisting of a film of mesogloea
covered with ectoderm, the digestive cavity of the bud being entirely closed (Fig. 1) .

Scyphistomas are never found with more than two buds attached. When two buds
are present they are always of different ages, and the elder is always attached to the
apex of the younger (Fig. 2) .

The Planula-like Larva. — When finally constricted off, the bud is a simple, spindle-
shaped, hollow body, without trace of mouth or tentacles. It is like a planula in form and
habits. The whole surface is covered with cilia, and it swims about, rotating from right
to left upon its long axis. In swimming, the distal end is directed forward. While swim-
ming, the larva is constantly changing its shape, assuming in a few minutes various forms
from an elongated spindle to a short heart-shape (Fig. 3, A, B and C) . It swims near
the bottom, hiding under any object that it may find there. When it strikes an
obstacle, it may rest there quietly, or it may rotate slowly upon its long axis. In all
its movements it reminds one very strongly of Agassiz's description of the planulae
of Aurelia.

The larva is white, speckled with a few greenish brown spots. It is rather opaque,
but much of the structure may be seen in a living specimen. A longitudinal section shows
that the ectoderm consists of a deep layer of very narrow and closely-packed columnar
cells (Fig. 39) . The mesogloea contains a few green cells (producing the greenish brown
spots), and some widely scattered colloblasts. The layer is thickest at the equator,
diminishing gradually to a very thin layer at both ends. The four septal muscles (sm)
are seen clearly at the distal end of the larva, embedded in the mesogloea and united with
the ectoderm near the apex (Fig. 38) . In one specimen, not yet detached from the
parent, it was possible to trace one of the muscles all the way from the proximal to the
distal end. The muscle fibres are already differentiated and line the tube of mesogloea,
while the nuclei occupy a central position.

The endoderm is a columnar epithelium, rather thin, with the cells closely packed

CASSIOPEA XAMACHANA. 215

together and coarsely granular at the proximal or posterior end of the larva. Toward
the equator the cells become higher ; and at the distal, or anterior, end they are large and
clear. The character of the cndoderm makes it possible to identify the anterior end of
the swimming larva with the distal end of the bud.

The Formation of the Mouth. — The first change to be seen in the swimming larva
is the formation of. the mouth, which occurs two or three days after the larva has been set
free. When writing my preliminary paper ('92 a) , I was in doubt as to the relation
between the poles of the bud and those of the larva. The better and more abundant
material obtained during the second visit to Jamaica proved conclusively that the proxi-
mal end of the bud forms the oral end of the larva, just as it does in Cotylorhiza, accord-
ing to Goette. The identification is made easy by the polar differentiation of the
endoderm described in the previous section.

When first seen the mouth looks like a minute pin-hole in the posterior end of the
larva (m, Fig. 4 B] . In longitudinal sections at this stage the first indication of the for-
mation of the mouth is the disappearance of the mesogloea from a small area at the poste-
rior end, so that there is no longer a distinct boundary there between the ectoderm and
endoderm (in, Fig. 40) . At the same time a small dent appears in the outer surface.
This deepens until it forms a minute tube connecting the endodermal cavity with the
exterior (Fig. 41) . The mouth thus formed gradually widens and becomes slightly
funnel-shaped.

Further stages in the development of the mouth are represented in Figs. 5, 6, 7. In
In Fig. 6 there is a distinct circular groove which outlines the base of the proboscis and
separates it from a rudimentary peristome. In Fig. 7 the peristome is well developed,
and the mouth is widely open.

There is no invagination of the ectoderm connected with the formation of the mouth,
and there is no oesophagus, " Schlundpforte " or '• Taschenvorhang." So, if Goette's
account of the formation of the mouth in the sexually produced scyphistomas be accepted,
we have here a case where an agamogenetic differs to a marked degree from the
gamogeuetic course of development.

The Scyphistorna. — With the elongation of the forward end, the formation of ten-
tacles, and the development of four gastric pouches, the free-swimming larva becomes a
typical scyphistoma.

By the end of the third or fourth day after the bud has been set free, the forward
end of the larva has elongated to form a stem equal in length to the rest of the body
(Figs. 7, 8, 9, and 11) . The end of the stem becomes expanded, generally flattened, and
the epithelium covering it becomes deeper than the rest (Fig. 42) . This epithelium pro-
duces a secretion which serves to fasten the larva to some solid object. Fixation usually
takes place during the fourth or fifth day, but the time varies greatly.

215 ROBERT PAYNE BIGELOW ON

Development of the Tentacles. — During the third day the peristome appears as a minute
ridge surrounding the posterior end of the larva, a short distance from the mouth. The
first series of tentacles arises during the following twenty-four hours as four perradial
angles in the margin of the peristome (Fig. 8) . Four interradial tentacles appear almost
simultaneously with these, or sometimes considerably later (Figs. 9, 10, 11, and 12).
The elongation of the tentacles takes place rapidly, so that at about the end of the sixth
day the peristome is surmoimted by a crown of eight tentacles, which equal the proboscis
in length.

With the broadening of the peristome the differentiation of the body of the scyphis-
toma into stem and calyx becomes apparent externally (Figs. 13 and 14). When the
eight perradial and interradial tentacles have become long enough to reach some distance
beyond the mouth, eight adradial tentacles appear in the angles between them. In
Fig. 14 the adradial tentacles are distinctly developed, and two of the tentacles of the first
series are bifurcated near the base. Figs. 15 and 16 represent the typical scyphistoma in
the sixteen-tentacle stage. The tentacles are now long and graceful, and thickly dotted
with batteries of nettle cells.

When fully developed, the scyphistoma is about one millimeter and a half in diameter ;
and it is provided typically with thirty-two tentacles.1 But there is as much variation in
the number of tentacles in the scyphistoma as there is in the number of sense organs and
parameres in the adult. The number of tentacles is seldom less than thirty-two, often
greater. The way in which this variation takes place is indicated in Figs. 21 and 14.
All degrees of anomaly may be observed, from a bifurcated tentacle shown at a: in Fig. 21,
through the condition represented in Fig. 14, to two completely separated tentacles occu-
pying the position of one typical one.

The tentacles of a well-developed scyphistoma, when fully expanded, exceed the
length of the body several times. According to the position in which they are held, the
tentacles may be divided into two series. Those of one series are held nearly erect, while
those of the other series, consisting of the alternate tentacles, are bent backward until
their tips nearly touch the ground upon which the animal rests. The action of the ten-
tacles in capturing prey may be observed in a small aquarium, under the microscope. As
soon as a tentacle comes into contact with a small floating body, such as a copepod, it is
whipped quickly into the mouth, and at the same instant the side of the mouth toward
the tentacle is opened more widely. On one occasion I saw two tentacles make captures
at the same time, and the mouth expanded in both directions at once, showing a close
co-ordination between the movements of the tentacles and of the mouth.

1 Specimens have been found with twenty-four tentacles, but it is uncertain whether this is a regular stage in develop-
ment between the sixteen- and thirty-two-tentacle stage, or a duplication of tentacles of the earlier stage.

CASSIOPEA XAMACHANA. '_> | 7

Development of the (ratstnc Pom-hex. — While the final result is the same, the method
of development of the gastric pouches differs entirely in our larva from the process as
described by Goette. The free-swimming larva is somewhat flattened laterally (Fig. 4),
and the long and short diameters are in the planes of the perradii. But there is no
evidence that the gastric pouches in the long diameter are formed any earlier than those in
the short diameter, and the ectoderm plays no part in their development. The formation
of the gastric pouches is usually described as a process of evagination ; but in this case, at
least in the earliest stages, the delineation of the pouches seems to be due rather to the
ingrowth of the septa.

These appear at the time of the formation of the first tentacles (Fig. 8) as four
minute vertical folds of the endoderm (sep, Fig. 43) , placed equidistantly in the angle
formed by the peristomal fold (compare Fig. 42) . The mesogloeal portion of the septum
is at first very thin, and in the four-tentacle stage (Figs. 8 and 43) is no higher than the
thickness of the endodermal layer of cells. The septal muscles do not penetrate the septa
at this stage. While well-developed below, they can be traced upward only to within
35 or 41 /z. * from the base of the septa.

In the eight-tentacle stage (Fig. 13) the septum is still very small, and the septal
mesogloea near the margin of the peristome is very thin, hardly thicker than a cell wall.
But at the central margin it has increased in thickness, and now the septal muscle may
be traced from the stem upward through this thickened portion of the septal mesogloea to
the peristomal ectoderm (sm, Figs. 44 and 47). It is impossible to determine whether
the new portion of the septal muscles is formed by growth upward of the muscles already
present in the stem, or whether it is the result of a proliferation of the peristomal ecto-
derm which may grow downward and fuse with the older portion of the septal muscles.

The four septa are now complete, and divide the digestive cavity into a large
central stomach, extending into the stem, and four shallow marginal gastric pouches.
The gastric pouches expand with the growth of the peristomal disc, so that the ecto-
derm and endoderm remain in close contact at the margin ; while the central edges
of the septa retain their original relative position. Thus the gastric pouches and the septa
become deeper as the larva increases in size. This is evident in the sixteen-tentacle stage
(Figs. 15 and 45 to 49) . It will be shown later that the relation of the interradial ten-
tacles to the septa is variable. In this stage, however, the distal part of the septal meso-
gloea has begun to disappear, so that immediately under the interradial tentacles the
endoderm of adjacent pouches is fused (Figs. 45, 46, and 47) . In the fully developed
scyphistoma this area of fusion is perforated, so that there is a communication between
adjacent pouches, forming the " Ringsinus" of German authors. In the specimen with

1 35 or 41 (i = about one seventh of the length of the larva.

218 ROBERT PAYNE BIGELOW ON

forty-two tentacles, from which the section represented in Fig. 51 was taken, this per-
foration was very small. Figs. 52 and 53 represent a little later stage, in which the
opening has become much wider.

Each septal muscle is a solid cord of cells, with a single layer of longitudinal muscle
fibres iu its periphery. In the peristome the fibres of the septal muscles spread out in a
fan-shaped arrangement toward the margin.

Four slight interradial depressions in the peristome may be observed as early as the
four-tentacle stage. They are deeper in the eight- and sixteen-tentacle stages, and the
septal- muscles may be seen to join the ectoderm at their bottoms. These depressions may
be homologous with the septal funnels of the Stauromedusae, or they may be merely the
result of the contracted condition of the larva. The question is of no importance, for
there can be no doubt about the homology of the muscles ; and whether the}' are solid or
hollow is merely a matter of detail. In later stages (Figs. 51 and 52) sections seem to
show that the peristomal depressions have deepened centrally so as to leave the inser-
tion of the septal muscles high up on the peripheral side. But here, again, before any
morphological conclusions can be drawn, account must be taken of the growth of the
perradial angles of the proboscis and of the effect of the contraction of the septal and
peristomal muscles.

Relation of Septa to Interradial Tentacles. — According to Goette, the interradial
tentacles are always interseptal in origin. That is, two of these tentacles are produced as
diver ticulae from each of the second pair of gastric pouches (ectodermal) . and their subse-
quent position in the planes of the septa is due to a secondary shifting. He finds
in this important evidence in favor of his theory of the close affinity between the
Scyphomedusae and the Anthozoa, for the tentacles of the latter are also invariably inter-
septal. (Jlaus, on the other hand, finds that the interradial tentacles of Aurelia and Coty-
lorhiza are variable in origin. According to his observations ('91 and '92), an interradial
tentacle may be interseptal, that is, arise as a diverticulum of a single gastric pouch, the
endoderm growing out and pushing the ectoderm before it ; or it may arise in the plane of
a septum by the union of two endodermal diverticulae, one from each of the adjacent
pouches. He holds, therefore, that the origin of the tentacles cannot be used as evidence
to uphold Goette's theory.

My observations on Cassiopea are in perfect accord with those of Glaus. A number
of series of transverse serial sections made from young scyphistomas in the eight- and
sixteen-tentacle stages were studied carefully with the aid of camera sketches drawn on
transparent paper. By this means the relations of the parts could be determined accu-
rately ; and the results are embodied in the series of diagrams, Figs. H to L. In the
eight-tentacle stage, according to Goette, the gastric pouches iu the long diameter, /• r,

CASSIOPEA XAMACHANA.

219

should each give rise to a single tentacle, while the pouches in the short diameter, r' r',
should each produce three. From a glance at Figs. H and J it will be seen that this is
not the case in Cassiopea. To

be sure, most of the tentacles at H I ^

this stage are intcrseptal in
position, but the interradial ten-
tacles arise as often from pouch-
es in the long diameter, r r, as
from those in the short diame-
ter, r' r' ; and in both Figs. H
and I there is one interradial
tentacle that is distinctly septal
in position. The septa are still
complete, so that there can
hardly be any chance of a shift-
ing of position. In the sixteen-
tentacle stage, there is a ten-
tacle in the plane of each septum
(Figs. K and L) ; but here the
perforation of the septa has
commenced, and a shifting of
relative position is possible.

Even at this stage irregularities are common. For example, one of the septa in Fig. K is
placed asymmetrically with relation to the tentacles, and two tentacles are wanting in
Fig. L.

The Strdbila, — Development of the Rhopalia. — When the scyphistoma has reached a
diameter of about two millimeters, there appear the first characters that are distinctive of
the strobila. The first noticeable change in this direction takes place at the bases of the
tentacles of the more erect series. This change may be regarded either as the outgrowth of
a conical lobe from the margin of the circumoral disc bearing the tentacle at its tip, or as
a conical widening of the basal portion of the tentacle. The former view is probably the
better. At about this time there appear in the tentacle, just beyond the apex of the cone
from which it springs, a few glistening white bodies. These are the so-called otoliths, and
mark the beginning of the formation of the rhopalinm (Figs. 17 and 18). The tentacles
containing these will be called the rhopalial tentacles.

These concretions, or otoliths, increase in number until they form a conspic-
uous mass, while the basal cone begins to broaden laterally. This is now distinctly non-

Figs. H-L. Diagrams illustrating the space relations between the septa
;n ul the tentacles, observed in five young scyphistomas of Cassiopea xama-
chana. r — radii of the long diameter, r' — radii of the short diameter.

220 ROBERT PAYNE BIGELOW ON

i
contractile, and may be spoken of as a marginal lobe of the periatome (Figs. 19 and

20). In the specimen shown in Fig. 21 we see the first indication of strobilization. The
upper, expanded part of the calyx is separated from a conical, lower portion by a slight
groove. The marginal lobes have become semicircular in outline, and a slight elevation is
noticeable on the aboral side of each rhopalial tentacle immediately external to the mass
of concretions. The epithelium at this point is pigmented, and forms the first rudiment
of the eye (oc. Fig. 22). Fig. 23 illustrates a more advanced stage, where the proximal
part of the tentacle is beginning to take on its final shape, and is separated by a pro-
nounced bend from the distal portion, which is still functional as a tentacle.

We come finally to a stage in which, while the long distal part of the tentacle retains
its characteristic structure and remains completely functional, the short proximal part
has become completely differentiated into a rhopalium. Fig. 54 is from a longitudinal
section of such a tentacle. The rhopalial part has assumed nearly its final shape. The
differentiation of its ectoderm into sensory epithelium, eye-spot, and layer of nerve fibres,
is complete. It has a lumen that extends outward to the solid chorda-like endoderm of the
distal part of the tentacle, and opens toward the centre into a gastric pocket. The endo-
dermal lining of the lumen is a columnar epithelium, the more distal cells being deeper
and containing the concretions. Compare Fig. 54 with Fig. 53, which, being interradial,
was certainly destined to be a rhopalial tentacle.

The growth of the marginal lobes, which were semicircular at the stage of Fig. 21,
has continued, and each lobe has now produced two secondary ones, one on each side of
the rhopalial tentacle. These are connected by a slight ridge that crosses the base of the
tentacle on its aboral side (h. Fig. 54) . The secondary lobes are the rhopalial lobes of
the margin of the umbrella (Fliigellappen of German authors) , and the connecting ridge
is the hood (Deckplatte) that covers the rhopalium. These marginal structures may be
seen in Fig. 24, and this brings us to another stage in the development of the rhopalium,
the absorption of the distal part of the tentacle.

In the strobila shown in Fig. 24, the rhopalial tentacles have a very different
appearance from what we have seen before. They are shorter than the other tentacles,
and are much swollen at a point just beyond the eye-spot. The distal portion is begin-
ning to degenerate. This process, when once begun, proceeds rapidly. During the few
hours that were spent in making this drawing, the rhopalial tentacles were reduced in
length nearly one half. The eye-spots and concretions were conspicuous, and in each of
the former there was a slight cup-shaped depression. This is the earliest stage in which I
observed slight medusa-like movements of the ephyra disc. The tentacle at this stage is
in a process of degeneration for about fifteen hundredth* of a millimeter outward from the
ocellus. In this area of degeneration ( t. Fig. 55) the endodermal cells are broken down,

CASSIOPP:A XAMACHANA. 221

the supporting membrane has disappeared, and the inner boundary of the ectoderm is
indistinct. The axial mass of this part of the tentacle is made up of loose particles of a
finely granular substance, in which may be seen many small and deeply stained nuclei.
There are also a number of green cells that apparently escaped into the central mass
when the supporting membrane broke down. There is evidently a free communication
between this mass of disintegrating material and the digestive cavity, through the
rhopalial canal.

The method by which the shortening of the tentacle is brought about would seem to
be as follows : The axial cells adjoining the cells that bear the concretions (Fig. 54) first
break down. Why they should do so, and at this particular time, I cannot say. This
disintegration proceeds centrifugally, and it is accompanied by a dissolution of the sup-
porting membrane. The ectodermal cells then either begin here and there to break
down while still in place, and the resulting debris is squeezed into the central cavity; or
else, the cells migrate, or are squeezed inward and then disintegrate. The continuity of
the remaining ectoderm is maintained, however. The products of degeneration prob-
ably pass through the rhopalial canal into the digestive tract. As this process continues,
the inward movement of the ectodermal cells is more rapid than their disintegration, so
that when the distal part of the tentacle is reduced to the size of the rhopalial part (Figs.
25 and 55), it is a solid mass of small cells with small nuclei that stain dark. Some of
these cells contain a large vacuole and have the nucleus pushed to one side. Scattered
among the small cells, there are a number of globular bodies as large as, or larger than,
the green cells, and completely filled with coarse granules that stain deeply witli safranin;
no nucleus is visible in them. The ocellus has now become distinctly cup-shaped (oc.
Fig. 55).

At about this time the interrhopalial tentacles begin to be absorbed in their turn
(Fig. 25) . The umbrellar margin has in the mean time grown out beyond the insertion
of each interrhopalial tentacle, on its aboral side, into two lobes with a hood between
(Figs. 25 and 26). This structure, although smaller, corresponds exactly to the rhopalial
lobes and hood, and is further evidence for the homology between the tentacles and the
rhopalia. The drawing reproduced in Fig. 25 was made between the hours of eleven
in the morning and two in the afternoon. At five o'clock of the same day the tentacles
had been reduced to one third the length shown in the figure, and the absorption of the
rhopalial tentacles was very nearly completed.

In the later stages of the absorption of the interrhopalial tentacles, the broken-down
material is evidently forced in some way into the radial canal. The rhopalium (Fig. 56)
is practically complete at this stage. The point (x) where the last trace of the tentacle
proper disappeared, is still distinguishable in sections by the presence of small cells with
indistinct cell walls, and by the absence of otoliths.

222 ROBERT PAYNE BIGELOW ON

These observations, then, confirm those of Glaus ('83,) who, without going into the
details of development, maintained that the rhopalia are modified basal portions of the
tentacles of the scyphistoma; and they contradict Goette's statement that the rhopalia
are developed independently of the tentacles.

Other Phenomena of Strobilization. — While the marginal structures are undergoing
the metamorphosis that has just been described, important alterations are taking place in
the general shape of the body. The horizontal constriction first noticed in Fig. 21 has
deepened (Fig. 24), while the fold below it has heightened, and the upper portion has broad-
ened and flattened, until the condition shown in Fig. 26 is reached. At this final stage
the upper portion has all the characteristics of a free-swimming medusa (ephyrula), except
that it is attached by a slender aboral stem to the centre of a goblet-shaped basal polyp.
The four interradial depressions in the peristome of the earlier stages have become nearly
flattened out, all that remains of them being the hollows between the projecting radial
angles, or pillars, of the proboscis, which have now become very prominent.

At the stage of Fig. 21, there may be noticed on the proboscis eight patches of
thickened ectoderm containing many nettle cells. These nettle batteries are arranged
symmetrically, one on each side of each pillar of the proboscis. At a little later stage
(Fig. 24) the batteries have become invaginated. forming cup-shaped depressions, thickly
crowded with nettle cells in all stages of development.

During the time when the larva is being differentiated externally into an upper and
a lower portion, internal changes are taking place in the former, which result in the
disappearance of structures characteristic of the scyphistoma and the appearance of
others distinctive of the medusa.

The orifices in the gastric septa have become relatively larger (cs. Fig. 52 and re.
Fig. 57) until the septa are reduced to columnar pillars, columellae, connecting the upper
and lower walls of the bodj^ and pierced longitudinally by the septal muscles (c. Fig.
58) . The columellae are called by German authors "Septalknoten," but they are not
homologous with the so-called " Septalknoten," or areas of adhesion, in the Peromedusae.
The columellae of the Peromedusae are the walls of the large septal funnels where they
pass from the subumbrella to the exumbrella, and, according to Haeckel's figures, are
separated by the gonads from the areas of adhesion.

In the fully developed scyphistoma of Cassiopea, the septal muscles are solid
throughout their length, and there is no cavity corresponding to the septal funnels,
which, according to Goette, are well developed in Aurelia. But in the strobila there
does appear a slight depression extending a very short distance into the end of each
septal muscle. In Figs. 57 and 58, where this is well marked, much of the depression
may be due to the strongly contracted condition of the specimens; but other specimens

CASSIOI'KA \\M\CIIANA. 223

not so contracted show at least the deeper part of the cavity, whicli therefore, ni:i,y be
truly a vestige of the septal funnel.

At an early stage of strobilization there may be noticed a short conical projection
from the central edge of each colnmella. 1 1 extends also around the sides. These pro-
jections arc probably the rudiments of the (irst four gastric filaments, which are dis-
tinctly developed. at the time when the ephvrula is set free ('//', Fig. 58) .

While the septa are shrinking to become the columellae, ridges appear opposite
each other on the upper and lower walls of the peripheral part of the digestive tract
between the bases of the tentacles. The epithelial membranes at the summits of oppo-
site ridges unite, and thus there is formed a series of lines of adhesion extending inward
from the periphery and dividing the space into a series of radial canals, each ending in
a tentacle. The two discs of mesogloea never fuse along these lines of adhesion, but the
endoderm remains between them as the endodermal lamella, or cathammal plate. At the
stage of Fig. 24 the lines of adhesion occupy about half the space from the margin to
the columellae.

The lower disc of the strobila remains simply an annular fold of the body wall until the
metamorphosis of the upper disc is nearly complete. The septa,! muscles in this region
bend outward with the rest of the body wall (Fig. 57) . At, length, however, the endoderm
grows out toward the periphery as four shallow pouches, leaving septa, between them
whicli contain the longitudinal muscles. Very soon alter this the septa, are perforated so
as to allow a fusion of the endoderm at their upper angles (Fig. 58). In the last stage
of strobilization (Figs. 20 and 5'J) the longitudinal muscles may be traced from the
peristome through the columellae and the mesogloea of the exumbrella to the narrow
isthmus where (lie ephyrula disc joins the basal polyp. The latter has now a well-
developed peristome ( Fig. 5!l), and the mesogloea in this region is very thin. Just in
the isthmus the muscles have disappeared, but they may be found again in the peristome
of the basal polyp and traced fora distance close under the epithelium to the edges of
the septa, where they bend abruptly downward, and continue through the septa into the
stem.

Although seldom visible in the living specimen, sections show that the basal polyp
at this stage possesses eight short tenacles (Fig. 5!>). It, lias also an annular fold of the
ectoderm, closely surrounding the isthmus (Fig. 5'J and p. Fig. GO) . This fold is the
rudiment of a new proboscis, which is without doubt entirely ectodermal in origin. But,
as Goette has pointed out, if does not follow from this that the lining of the proboscis is
ectodermal in sc.yphistoma.s developed from the egg. Pulsating contractions of the
umbrella are first noticed at the time when the rhopalial tentacles begin to be absorbed
(Fig. 24). They are then feeble and at long intervals. At the stage of Fig. 26 these

224 ROBERT PAYNE BIGELOW ON

movements are. rapid and violent. The rhythm is interrupted by few pauses, and these
are short. The result of these movements is that the thin wall of the isthmus is
ruptured, and the ephyrula is set free.

After this separation, the basal polyp has the appearance represented in Figs. 27 and
28. It is a scyphistoma with seventeen short tentacles and a rudimentary proboscis
(Figs. 61 and 62). The proboscis and the tentacles grow rapidly, so that in a few days
it is impossible to distinguish a regenerated basal polyp from a young scyphistoma in the
sixteen-tentacle stage, except that the former has a somewhat thicker stem. It may be
inferred from this complete regeneration of the basal polyp that it undergoes repeated
strobilization, as Clans l has found to be the case in Aurelia.

The Ephyrula.' — The ephyrula of Cassiopea is very different in appearance from the
corresponding stage in ordinary scyphomedusae with eight rhopalia. Cotylorhiza has an
ephyrula resembling the same stage in the semostomatous medusae. Good figures of
this are given by du Plessis and Clans, and there is a striking difference between these
figures and Figs. 21) and 30 in this paper, which are camera drawings of well-preserved
ephyrulas of Cassiopea, mounted in balsam. Fig. 2H represents a young Cassiopea that
has not long enjoyed a free existence. The general shape of the umbrella is like that of
the adult, and there is the same concavity in the centre of the exumbrella, while the
margin curves in the opposite direction, as in Fig. 64. The typical ephyrula of Aurelia
or Cotylorhiza has eight marginal arms with two lobes at the end of each, and between
each pair of lobes there is a rhopalium. In Cassiopea, structures corresponding to these
arms are present to the number of sixteen, or often more. But these do not destroy the
general circular outline of the animal, for the}' are connected by thin areas on the
umbrella, alternating with an equal number of ridges, which at an earlier stage bore the
interrhopalial tentacles on their under sides.

AVe have, then, at this stage the marginal zone of the umbrella marked by a number
of short radial ridges separated by an equal number of thin areas. The ridges are in
line with the radial canals. At the peripheral end of each ridge the margin of the
umbrella is produced into two lobes, those adjoining the rhopalia being well marked,
the others small and inconspicuous (//. Fig. 30).

In Fig. 29 there are seventeen, and in Fig. oO, twenty-three, rhopalia. The latter
is an unusually large number, and it will be noticed that the number of marginal lobes
has not increased in proportion, SO that irregularities of the margin occur in many places,
as described in the section on variations.

At this stage the rhophalia have come to lie, as in tin- adult, wholly within the margin
of the umbrella, and project from its subumbrellar surface. The interrhopalial tentacles

1 SIT fiiiit-iiciu-, dims ('92).

CASSIOl'KA XA.MACIIANA. 225

have totally disappeared. Th<' lines of adhesion separating the radial canals arc faintly
visible as radiating lines of greater transparency.

The four lips of the mouth are spread out into a cross-shaped figure, and one
may look directly through the lumen of the oesophagus into the stomach and see the
four gastric filaments (Figs. 29 and 30). Each one of the four lips is nearly square, and
from its two outer angles there are two grooves that extend obliquely inward until they
meet and form a V. The point of the V is in an angle of the oesophagus, along which
there is a groove that is continuous with the other two grooves, and that extends into
the stomach. On the interradia.1 side of each of the eight labial grooves, there may be
seen a small roughly circular area that is less transparent than the rest. These areas are
the nettle batteries, first seen in the strobila. The margins of the lips are provided
with numerous small processes, the digitella, which are arranged in a single continuous
series.

Fig. Go is a section of an ephvrula that has just become free. In this stage there is
still an opening through the aboral wall of the stomach, and one may see the last vestige
of the connection between the columella and the exumbrella, which contains also the
degenerating remnants of the septal muscle.

At a little later stage, when the opening in the roof of the stomach has closed,
both the septal muscles and the septal funnels totally disappear. Sometimes one, some-
times the other, is the first to vanish.

The Later Stages. — The later stages in the development of Cassiopea will be
treated very briefly. While the umbrella remains at first unchanged, the metamor-
phosis of the month parts is inaugurated by the growth of the two outer angles of each
of the more or less quadrate lips, so that they are soon drawn out into extended lobes
(Fig. 31). At the same time the pillars of the proboscis thicken, and the mesogloea is
continued outward along each of these lobes as a midrib. We have then eight oral
arms, each with a longitudinal groove, supported by a midrib, and fringed with digi-
tella,— arms very similar to those characteristic of the genus Aurosa Haeckel ('79). But
it is only the mouth parts of Cassiopea that may be said to pass .through an Aurosa stage,
for the comparison cannot, at this time at least, be carried to the other organs.

Clans has described ('83) some of the principal stages in the metamorphosis of Pilema
(Rhizostoma) and Cotylorhiza. He regards the formation of the eight oral arms as a dif-
ferent process in these forms from what occurs in Aurosa. But it appears to be merely the
same thing expressed differently.

In the next stage we find two oral funnels, or oscula, and a small vesicle developed at
the tip of each oral arm. The other portions of the arm are still open and fringed with
digitella, as before, but the outline is no longer a regular curve, for there are folds in the

226 KOP.KKT PAYNE BIUELOW ON

margin. The deepest folds are the most distal, and they heroine progressively more
shallow toward the base of the arm. The central mouth is still widely open. The subgeni-
tal cavities are well developed at this stage. Figs. 114, (i;~>, and (id show how the oral disc
is formed, and how the subgenital cavities are produced by the great increase in thickness
of the mesogloea at the pillars of the proboscis and the liases of the oral arms. By the
growth of these structures, the siibgenital cavities are necessarily produced. The only
special adaptations are the subsequent growth and folding of the alioral wall and the
narrowing of the orifice.

Tlie marginal lobes of the umbrella now begin to broaden, and thus approach the
adult condition, but (here is only a single •' vellar " lobe between two rhopalial ones.

At a little later stage, when there are three oral funnels at the tips of the arms
(Fig. 33), the re-entrant angles between the pillars of the proboscis have grown inward,
met at the centre, and fused. In this way the lumen of the oesophagus is divided
into four tubes (Fig. •">-), representing the grooves that were present in its angles in the
earlier stages. In the figure the fusion at the centre has gone so far as to involve the
edges of the lips, and the labial grooves of the different pairs of arms are not in open
communication, but a short cross-shaped tube connects them at the centre, and the oral
disc is now completed.

It is interesting to note that Clans has found a stage both in Pilema and in Cotv-
lorhiza that, while showing the characteristic family dilferences, has also a certain
resemblance to this stage in Cassiopea. In all three the walls of the proboscis have fused
so as to divide its lumen into four tubes, and the formation of oscula has begun at the tips
of the arms in such a way that we have on each arm three oscula with a vesicle in the
angles between them. The occurrence of this stage in the ontogeny of three so distinct!}'
separated families must have some morphological significance, and we mav regard these
eight primary vesicles as homologous in the three groups.

The mode of formation of the oral funnels becomes evident at this stage. They are
not formed in Cassiopea simply by a series of fusions of the lips along the line of the
labial groove, as Hamann (81) states to be the casein Cotylorhi'/a. It is more like the
process in Pilema, as described by Clans. Kach of the primary funnels is represented at
first by one of the folds in the margin of the lips referred to above (Fig. 33). The
fold deepens, and its edges are brought together on the ventral side and fuse, leaving an
opening at the apex of the fold, the osciibiin. At the same lime the labial groove in
this region is converted into a canal bv the fusion of the lips on its two sides. After
the fusion all trace of what has occurred quickly disappears.

With the division of the oesophagus into four tubes, and the completion of the oral
disc, our larva comes to be distinctly a rhi/ostomatous medusa. Further development of

CASSIOl'KA XAMACHANA. 227

the mouth parts consists in the continued division of the labial, or brachial, grooves into
oral funnels and brachial canal, together with the development of oral vesicles. ]»y the
time two or three vesicles ha.ve lieen formed on the end of each arm, a vesicle appeal's in
the centre of the oral disc. Except for this interruption, the development of the mouth
parts proceeds regularly in a centripetal direction. The funnels and vesicles are formed
first at the tips of the arms, and then one after another in regular succession toward the
centre. Each of these primary funnels is the rudiment of one of the primary branches of
the arm. When the process of forming funnels has reached about half the length of the
arm, the distal funnels begin to subdivide. l»y this subdivision of the primary funnels
new ones are produced, of which some are the rudiments of secondary branches ; these sub-
divide again, and so on. as long as growth continues. The subdivision is not dichotomous,
but, takes place in such a wa,v as to produce alternate brandies. The formation of a
vesicle takes place at this stage in some way at about the time of the completion of the
adjoining funnel. 1 have not been able to determine whether the vesicle is a funnel with
the orifice dosed, as Hamann claims it to be, or whether it is an evagination from the
pedicle of a funnel, as at first it seemed to me to be, and as Clans thinks it probably is.

According to Haeckel ('79). the genus Archirhi/.a represents a form that was the
ancestor of all the rhizostomatous medusae. Of this genus there are two known species,
A. primordialis Haeckel, and A. iniroxn Haeckd. They agree in having four suhgenital
cavities and eight simple unbranched arms that are provided with a single zig-zag row of
closely set oral funnels, and are devoid of other appendages. Hamann says that a stage
representing this condition is a feature of the ontogeny of rhizostomatous medusae. From
what lias been said it is evident that we have no such stage in the development of (_'<isxi-
opi'K .••(iiiittrhdiHi, for while the labial groove is still open in the proximal half of the oral
arm, in its distal half the vericlrs are formed, and brandies are in the process of
formation.

The outline of the umhrellar margin has not changed essentially since the last stage.
The areas of adhesion have become much wider than the radial canals they separate, and
in them there has appeared a network of anastomosing canals, while the gastric, filaments
have become numerous.

We have now followed the larva of our Cassiopea from its first appearance as a bud
to a point, where, with the exception of the gonads, all the organs of the adult are out-
lined. Here we must take leave of it.

228 ROBERT PAYNE BIGELOW ON

SUMMARY AND CONCLUSIONS.

Cassiopea and Polyclonia are genera of rhizostomatous medusae- peculiarly modified in
adaptation to a sedentary mode of life in shallow water among the mangroves bordering
tropical seas. A comparison of specimens of 1'oIi/cJriiiitt frontloufi Ag. with (_*<ixxioj>i'a
.rii/inic/iin/f/ shows that these two forms are specifically distinct, although in general
appearance they are very similar and they have the same geographical range and
habitat.

C. xamachaiia is remarkable for its variability. This is especially shown in the
appendages to the mouth parts and in the structures at the margin of the umbrella. It
will be noticed that the most frequent number of rhopalia, in the twenty-seven specimens
examined, was sixteen, which is the typical number for the genus. But the variations are
not arranged symmetrically on the two sides of this mode, for specimens having a
greater number of rhopalia are more than twice as many as those having less. The
species shows a strong tendency toward duplication of the rhopalia and associated struc-
tures of the umbrella ; and at the same time the symmetrical relations of the parts tend
to be preserved. The great majority of scyphoinedusae have only eight rhopalia, and in
Cassiopea with its sixteen rhopalia we have a beautiful illustration of Darwin's law that
"Apart developed in any species in an extraordinary degree or manner, in comparison
with the same part in allied species, tends to be highly variable." Study of the color
markings and measurements of the month parts indicates the division of the species into
three varieties; and it was in one of these, var. A, that the duplication of marginal organs
was especially prevalent.

The color of both larvae and adults is due to a great extent to the presence in the
rnesogloea of minute symbiotic algae. That these are plant cells was demonstrated by
micro-chemical tests. Their presence undoubtedly enables the medusae to live in water
that would be too poor in ox}'gen for most marine animals.

The search for developing eggs proved unsuccessful, but scyphistoma larvae were
abundant, and it was found that they were multiplying rapidly by budding.

The bud arises as an evagination of the body wall of the scyphistoma. There is no
evidence of any special gemminal epithelium. The bud, when set free, differs from a
planula chiefly in the possession of a well-defined mesogloea and four septa] muscles. The
septal mnscles are shown to be formed as branches of the two adjacent septal muscles of
the parent. The mouth of the young scyphistoma is formed by a minute perforation at
the former point of attachment, while the distal end of the bud becomes the stem. This
remarkable orientation agrees with what Goette and Clans have found in Cotylorhiza. In

CASSIOPEA XAMACHANA. 229

the formation of the mouth there is no evidence of any in vacillation of ectoderm Ap-
parently the oesophagus, as well as the gastric pouches, is lined with endoderm. On
the other hand, the oesophagus of (lie lower disc of the strobila is formed wholly of
ectoderm.

The four radial tentacles are formed simultaneously, and the four interradial ones
appear at the same time or slightly later. These are followed by eight adradial tentacles
and sixteen more are added a little later, making thirty-two in all. It was found that the
rudiments of the interradial tentacles are not at all constant in position with relation
to the septa. Some were septal, others were interseptal, sometimes on one side of the
septum, sometimes on the other, thus agreeing with the observations of Clans on Aurelia
and Cotylorhiza.

The four gastric pouches are formed at the same time by the ingrowth of the septa.
They are soon brought into communication at the periphery by the perforation of the
septa, which become reduced to columellae surrounding the longitudinal muscles.

Contrary to Goette's opinion, it may be stated positively that the rhopalia are differ-
entiated in the bases of alternate tentacles. After the rhopalium is fully developed, the
distal part of the tentacle undergoes degeneration and is absorbed. The development was
traced for the first, time through all its stages, and the opinions of Agassi/, Chins, and
Lendenfeld are fully confirmed.

The sc\ phomedusae are the only coelenterates, which possess four longitudinal
muscles of ectodermal origin completely imbedded in the mesogloea between the points of
insertion. The buds of Cassiopea have this distinctively inedusoid characteristic long
before they are detached. Moreover, the methods by which the mouth and the gastric
pouches are formed differ entirely from what is said to take plsice in larvae that pass
through an suitho/oan stage. Therefore, whatever may be true of the larvae derived from
eggs, this stage is certainly omitted in larvae produced by budding.

The first step toward the formation of the free-swimming medusa is the perforation
of the gastric septa which takes place in the young scyphistoma before it is fully
developed. Then follows a period of growth and reproduction by budding, and further
metamorphosis begins finally with the process of strobilization. Besides the differentia-
tion of the sense organs and the development of the marginal lobes of the umbrella, the
most important events are the development of the angles of the mouth into quadrate
lobes and the fusion of the two layers of endoderm along certain areas of adhesion so us
to divide (he periphery of the digestive tract into a, series of radial eansvls. Only one
medusa is formed ; but, separated from this by a constriction, is a small basal segment
which, a short time before the medusa becomes free, begins to develop gastric pouches,
tentacles, and a proboscis. This becomes eventually a perfect scyphistoma, and after a
period of growth probably undergoes strobilizationjigain.

230 ROBERT PAYNE BIGELOW ON

There has been considerable discussion in regard to the nature of strobilization ; the
question being whether the medusa is to be regarded as a metamorphosed scyphistoma,
or as derived from the scyphistoma by a process of budding. . In the monodiscous strobila
of Cassiopea we have clearly a metamorphosis. The form of the medusa is the result of a
series of changes which begin very early and involve all the essential organs of the scy-
phistoma. The portion not involved in these changes merely serves as a mechanical sup-
port. That this part is separated off and regenerates the lost parts, instead of being
absorbed, may be regarded as merely an incidental fact.

If there be any question of budding it refers to this basal segment. And this suggests
a striking analogy between the basal polyp and the peculiar planula-like buds. In the
first place they have the same orientation relative to the upper disc. In both, the distal
end forms the stem and the proximal end forms the mouth. In the second place, they
have essentially the same structure. Each one consists of a simple sac with a wall made
up of three layers, ectoderm, mesogloea and endodcrm, and each is provided with four longi-
tudinal muscles imbedded in the mesogloea,. What differences appear in the subsequent
development, may be attributed to the different ways in which the two become sepai'ated
from the disc and to the greater si/-e of the longitudinal muscles in the basal poly}). The
production of supernumerary tentacles, rhopalia, and marginal lobes, is common in this
species. Why may we not regard the buds as supernumerary basal polyps, and their
subsequent development as a process of regeneration preserved and modified by natural
selection for its obvious advantage?

The ephyrula of Cassiopea has the same number of rhopalia as the adult, and differs
in shape from the corresponding stage of ordinary scyphomedusae with eight parameres.

The most important event in the later stages is the metamorphosis of the mouth
parts. The angles of the four quadrate lips become extended to form eight oral arms
somewhat similar to those found in the adult Aurosa. There is no Archirhiza stage, but
there follows a stage with the oesophagus divided into four tubes, and with three oscula
and an oral vesicle on each arm. A similar stage has been found in Pilema and Coty-
lorhiza, and it may have some ph ylogenet.ic significance.

The studies of numerous investigators upon the Ascidians have demonstrated that a
knowledge of the gamogenetic development of an animal will not always enable one to
predict how the organs will lie formed in the agamogenetic process. So it may be
objected that the results set forth in this paper do not afford a valid basis for the criticism
of work done by others on larvae developed from the egg. On the other hand, the eight-
tentacle stage of the bud larva of Cassiopea is so like the same stage of the sexually pro-
duced larva of its near relative L'olyclonia that it would lie impossible to tell them apart,
and, in the absence of any evidence to the contrary, there seems to be no necessary reason

CASSIOPEA XAMACHANA. 2?>1

for supposing that their later history is different. At any rate, we have la-re the lirsl,
fairly complete history of the development of a scyphomedusa from the Imd; :ind when
the sexual development of CiixxiojH'it .mmachana or an allied species is studied, this
memoir will serve as a means of comparison, and will make it possible to determine
whether or not the two modes of ontogeny are alike.

LITERATURE.

Agassiz, A.

'65. North American Aealephae. Illustrated catalogue of the Museum of comp. zool. at Harvard College. No. 2, 1865,

p. 41.

'81. Polyclonia frondosa. Nature, vol. 24, 1881, p. 609.
Agassiz, A.. & Mayer, A. G.

'99. Acalephs from the Fiji Islands. Bull. mus. comp. zool., vol. 32, uo. 9, 1899. (Description of Citssiopea ndroxia.)
Agassiz, L.

'60. Contributions to the natural history of the United States of America. Vols. 3 and 4. Boston, 1800.
Archer, H.

'81. Medusae. Nature, vol. 24, 1881, p. 307.
Berger, E. W.

1900. Physiology and histology of the Cubotnedusae, including Dr. F. S. Conant's notes on physiology. Mem. binl.

lab. Johns Hopkins Univ. vol. 4, no. 4.
Bigelow, R. P.

'90 The marginal sense organs in ilic Pelagidae. Johns Hopkins Univ. circ., vol. 9, 1890, no. 80, p. 65.

'92a. On reproduction by budding in the Disromedusae. J. H. U. circ., vol. 11, 1892, no. 97, p. 71.

'92b. On the development of the marginal sense organs of a rhizostomatous medusa. J. H. U. circ., vol. 11, 1892.

in p. 97, p. 84.

'92c. On a new species of Cassiopea from Jamaica. Zool. Anz., Jahrg. 15, 1892, no. 393, pp. 212-214.
'93. Some observations on Polyclonia frondosa. J. H. U. circ., vol. 12, 1893, no. 106, p. 106.
Brandt, J. F.

'38. Ausfiirliche Beschreibung der von C. H. Martens auf seiner Weltumsegelung beobachteten Sdiiruniualleii. Mem.

Acad. imp. sci. St. Petersbourg, sci. math. phys. et nat., (6) vol. 4 (pt. 2, sci. nat., vol. 2), 1838, pp. 394-3!>'.».
Brooks, W. K.

'86. The life history of Hydromedusae, a discussion of the origin of medusae and the significance of metagenesis,

Mem. Boston soc. nat. hist., vol. 3, no. 12, 1886 ; also Mem. biol. lab. Johns Hopkins Univ., vol. 1.
Claus, C.

"77. Studien liber die Polypen und Quallen der Adria. Denkschr. d. math.-naturw. Classe d. k. Acad. d. Wissen-

schafteu, Wien, Bd. 38, 1877.

'81. Ueber einige bislang noch unbekannte Larvenstadien von Hhiznstoma. Zool. Anz., 1881, No. 7(1. p. 79.
'83. Untersuehungen fiber die Organisation und Entwicklung der Medusen. Prag and Leipzig, 1883, 9(5 pp., 20 pis.
'84. Die Ephyren von Cotylorhiza und Rhizostoma. Arb. zool. Inst. Wien, vol. 5, 1884, p. 169.
'90. Ueber die Entwickluug dps Seyphostoma von Cotylorhiza, Aurelia, und Chrysaora, I. Arb. zool. Inst., Wien,

vol. 9, 1890, p. 29.

'92. Ueber die Entwicklung dfs Sc\ phnstoma von Cotylorhiza, Aurelia und Chrysaora, sowie iiber die Kystrmatisrlie
Stellung der Scyphomedusen, II. Arb. a. d. zool. Inst. d. Univ. Wien, vol. 10, 1H92, pp. 1-70.

232 ROBERT PAYNE BIGELOW ON

du Plessis, G.

'81. Reuiarques sur les metamorphoses de la Cassiopee bourbounieuue. Bull. soc. vaudois, Lausanne, ser. 2, vol. 17,

1881, p. 633.
Fewkes, J. W.

'82. Notes on aealephs from the Tortugas, with description of new genera and species. Bull. mus. comp. Zool., vol. 9,

no. 7, 1882.
Fischer, P.

'88. Note sur des scyphistornes de uieduse acraspede. Bull. soc. zool. de France, vol. 13, 1888, p. 96.
Forskal, P.

1775. Pescriptiones aninialiuni .... <juae in itinere orientali observavit. Hauniae, Moller, 1775, p. 107.

1776. Icones rerum naturalium, quas in itiuere orientali depiugi curavit. Haimiae, 1776, pi. 31.
Frantzius, A. v.

'53. Ueber die Juugeu der Cephea (Cotylorhiza.). Zeitschr. f. wiss. Zool., Bd. 4, 1853, p. 118.
Goette, A.

'87. Entwicklungsgeschichte de.s Aurelia aurita und Cotylorhiza tuberculata. Hamburg and Leipzig, 1887.

'91. Glaus uucl die Entwicklung der Scyphomedusen. Leipzig, 18111.

'93. Vergleichenden Entwicklungsgeschichte von Pelagia noctiluca Per. Zeitschr. f. wiss. Zool. Bd. 55, 1893, pp. 645-

695. [Die Larven von Cotylorhiza tuberculata L. , pp. 645-659.]
Gray, J. E.

'69. Habits of the Medusae. Annals and mag. uat. hist., ser. 4, vol. 4, 1809, p. 295.
Guppy, H. B.

82. Habits of Scyphoinedusae. Nature, vol. 27, 1882, p. 31.
Haacke, W.

'85. Pseudnrliiza Haeckelii spec, nov., der Endspmss des Discomedusenstammes. Biol. Centralb., vol. 4, 1885, p. 291.
Haeckel, E.

'69. Crambessiden. Zeitschr. f. wiss. Zool., vol. 19, 1809, p. 5n!».

'79. Das System der Medusen. Jena, 1879.

'81a. Report on the deep sea medusae. Challenger reports, znnl., vul. 4, 1881.

'81b. Metagenesis and Hypugeiiesis von Aurelia. Jena, 1881.
Hamann, 0.

'81. Die Muiidarme der Hluzostomen und ihre Anhangsurgane. Jenaische Zeitschr. f. Naturw., vol. 15, 1881, p. 243.
Hertwig, 0. & R.

'78. Das Nervensysteni uud die Smnesorgane der Medusen. Leipzig, 1878.

'79. Die Actinien. Jena, 1879.
Hesse, R.

'95. Ueber das Nervensystem und die Siunesorgane von Rhizostoma cuvieri. Zeitschr. f. wiss. Zool., vol. 60, 1895,

p. 411.
Hyde, Ida H.

'94. Entwicklungsgeschichte eiuiger Scyphomedusen. Zeitschr. f. wiss. Zool., vol. 58, 1894, p. 531-565.
Keller, C.

'83. Untersuehungen iiber neue Medusen aus dem Rotlien Meere. Zeitschr. f. wiss. Zool., vol. 38, 1883, p. 621.
Lacaze-Duthiers, H. de.

'93. Scyphistume. Archives de zool. exper., ser. 3, vol. 1, 1893, notes et revues, p. xxx.
Lamarck, J. B. de

'37. Histoire uaturelle des animaux sans vertebres. Ed 3. Paris, 1837, vol. 1, p. 457.
Lang, Albert.

'92. Ueber die Knuspung bei Hydra und einigeu Hydropolypen. Zeitschr. f. wiss. Zool., vol. 54, pp. 363-385.
Lendenfeld, R. v.

'82. Ueber eine Uebergaugsforiii zwischen Semostouieu und Rhizostouien (Pseudorhiza). Zool. Anz., Jahrg. 5, 1882,
p. 380.

< ' ASSTOPEA XAMACIIANA.

2:;;]

'84. Zur Metamorphose der Rhizostomen. Zool. Anz., Jahrg. 7, 1884, p. 429.

'88. Die australischen rliizosl.niucu Meduseii. Zeitsehr. f. wiss. Xool., vnl. 47, Isss, p. aoi.
McMurrich, J. P.

'91. The Development of Cyanea arctica. Amer. nat., vnl. 2"i, 1H91, p. 287.
Metschnikoff, E.

'86. Ernbryologische Studien an Medusen. Wien, 188(5.
Milne-Edwards, H.

'49. Les zoophytes. Cuvier, Ri'gne animal illustre. Vol. 10, 1840, pi. ,r>l
Reid, J.

'48. Observations on the development of the medusae. Annals and mag. nat. hist., ser. 2, vol. 1, 1848, p. 25.
Sars, M.

'41. Ueber die Entwickhmg der Medusa aureta und Cyanea capillata. Arch. f. Naturg., vol. 7, 1841, p. 9.
Schewiakofl, W.

'89. Beitrage zur Kenntnis der Acalephenauges. Morph. Jahrb., vol. 15, 1889, p. 25.
Schneider, A.

'70. Zur Eutwicklungsgeschichte der Aurelia aurita. Arch, f. mik. Anat. vol. fi, 1870, p. 363.
Schultze, L. S.

'98. Rhizostomen von Ternate. Abh. senckenb. naturf. Ges., vol. 24, 1898, p. 153. [Occurrence of C'assiopea, sp.?]
Smith, Frank.

]91. The gastrulat.ion nf Aurelia tiavidula Per. & Les. Bull. mus. comp. zool., vol. 22, no. 2, 1891.
Tilesius, W. G. v.

"29. Zur Naturgeschichte der Medusen. Nova Acta Physico-Medica, N. C., vol. 15. 1829, p. 266.
Vanhoffeu, E.

'88. Untersuchungen iiber semaeostome uud rhizostome Medusen. Leuckart & Chun's Biblioth. Zool., Heft 3, 1888.

EXPLANATION OF PLATES.

All the figures from 1 to 28, and Figs. 34 and 36, are freehand drawings made from the living animals. Figs. 29 to
33 are from well-preserved specimens mounted in balsam, and were outlined with the camera lucida. The remaining figures,
except Fig. 37, are camera-lucida drawings of microtome sections. All the figures are reduced uniformly to .slightly less
than one half the diameter of the original drawings.

LETTERING COMMON TO ALL THE FIGURES.

b. — bud.

c. — columella.

c d. — circum-oral disc, or peristome.

c s. — circular sinus.

ex. — -calyx.

d. — digitellum.

ect. — ectoderm.

e 1. — endodermal lamella, or eathammal plate.

end. — endoderm.

e u. — exmnbrella.

g. — stomach.

g f . — gastric filament.

g p. — gastric pouch.

h. — hood.

i 1. — interrhopalial lobe.

i t. — hit errh opal ia.l tentacle.

1 g. — labial groove.

m. — mouth.

mes. — mesogloea, or supporting membrane.

n f. — nerve fibres.

o a. — oral arm.

oc. — -ocellus.

o d. — oral disc.

oe. — oesophagus.

234

PAYNE BK4ELOW ON

oe. t. — oesopliageal lube, or canal.

os. — osculum.

ot. — rhopalial concretion (otolith).

p. — proboscis.

p p. — pillar of the proboscis.

r c. — radial canal.

rh. — rh opal him.

rh. c. — rhopalial canal.

rh. t. — rhopalial tentacle.

s. — stem.

s e. — sensory epithelium.

sep. — septum.

s f. — septal funnel.

s <;. — subgeiiital cavity.

s ui. — septal muscle.

s u. — subuuibrella.

t. — tentacle.

v. — vesicle.

za. — brownish green cells, symbiotic algae.

PLATE 31.

Fig. 1. Portion of the calyx and stem of a scyphistoma with a fully developed bud (b) attached. A, the outline of the
same bud when contracted.

Fig. 2. Similar to the preceding, except for the formation of a second bud (/i') which bears the older one (ft) upon its
apex,

Fig. 3. The form of the plauula-like bud during the first two days alter becoming free. A, B, and C, are the successive
changes in shape observed in one specimen during a few minutes. Tlie arrows show the direction in which it swims.

Fig. 4. Larva of probably the third day. A, lateral aspect; B, oral aspect.

Fig. 5. A larva of abmit the same age, 48 hours or over.

Fig. li. Another larva of 48 hours or over.

Fig. 7. Larva of the fourth day.

Fig. 8. Free scyphistoma of the fifth day, with rudiments of four tentacles. The arrows show the direction of progres-
sion and rotation.

Fig. 0. A scyphistoma of about the same age, with the rudiments of eight tentacles.

Fig. 10. Oral aspect of ?i similar larva, perhaps somewhat younger.

Fig. 11. A scyphistoma a little more advanced, probably in the fifth day.

Fig. 12. A still more advanced seyphistoma of the fifth day. The four tentacles first formed are much longer than the
other four.

Fig. I-'.. Scyphistoiiri of the sixth day, attached, and with eight tentacles.

Fig. 14. Scyphistoma with rudiments of the second set of eight tentacles. Two tentacles of the first series are bifur-
cated

Fig. l"i. Scyphistoma with sixteen tentacles fully developed, in the attitude of feeding.

Fig. Hi. Oral aspect of a similar specimen.

PLATE ;te.

Fiu. 17. Scyphistoma showing first truces of rhopalial structure.

Fit;. IK. A small portion of the margin more highly magnified.

Fig. l!i. Scyphistoma at a slightly older stage.

l-'iu. i'o. Small part of the margin of a similar larva.

Fig. 21. An early stage in strobilization.

Fig. '2'2. A rhopalial tentacle of the same specimen seen from the side.

Fig. '2'.',. An oliler rhopalial tentacle.

Fit;. '_'l. Strobila in which the rhopalial tentacles have begun to degenerate.

CASSIOPEA XAMACHANA. 235

PLATE 33.

Fig. 2fi. Strobila, in which the degeneration of the rhopalial tentacles is nearly completed, and the inteirhopalial
tentacles have begun to degenerate.

Fig. 20. A complete slrobila. The basal polyp bears a bud which broke off and swam away while the drawing was
being made. The ephyrula was detached during the following night. The rhopalia are visible through the umbrella.
At y, is a pair of twin rhopalia ; compare i/, Fig. 30.

Fig. 27. The basal polyp of the same specimen, a few hours after the separation of the ephyrula.

Fig. 28. Optical section of the same.

Fig. 29. An ephyrula recently set free. Oral aspect. The gastric filaments are visible through the mouth. X 31.

Fig. 30. A specimen of about the same age, showing variations of the margin at u, w, x, y and z. X 31.

PLATE 34.

Fig. 31. Mouth parts of a young medusa in the Aurosa stage. The gastric filaments may be seen through the central
mouth opening. X 33.

Fig. 32. Oral disc of an older specimen. The oesophageal tubes appear as light areas, one at the junction of each pair
of labial grooves.

Fig. 33. One of the oral arms from the same specimen as Fig. 32.

Fig. 34. Floor of the stomach ami the oral arms of an adult viewed from the aboral side. The roof of one subgenital
cavity is removed, and a thread is represented as passing through the external orifice into this cavity, at f.. The ultimate
branches are represented on only one of the oral arms.

Fig. 35. Portion of the aboral surface of an adult. About half natural size.

PLATE 35.

Fig. 36. Section of a young bud. X Zeiss DD -|- oc. 2.'

Fig. 37. Diagram to show the branching of the septal muscles, SHI,, SIM.,, sm3 and s»i4, and the connections of the
septal muscles of the bud, SHI',, sin'.,, and sm'3. Reconstructed from the series of sections of which Fig. 30 is one.

Fig. 38. Section through the distal apex of an older bud, showing the attachment of a septal muscle to the ectodermal
epithelium. X Zeiss H -f- oc. 2.

Fig. .'!!>. Longitudinal section of a planula-like larva. 1> was the distal, and P the proximal, end of the bud while
attached. X Zeiss DD + oc. 2, dt. 100.

Fig. 40. Longitudinal section through the posterior end of a swimming larva, in which changes preparatory to the
formation of the mouth are taking place. X B & L i -f- Zeiss oc. 2.

Fig. 41. Similar section of a slightly older larva, showing the mouth as a small opening not exceeding in width the
thickness of the section. X B & L J + Zeiss oc. 2, dt. 160.

Fig. 42. Adradial section of a scyphistoma a little older than Fig. 9 (5th day). X Zeiss DD + oc. 2.

Fig. 43. Obliquely transverse section of a specimen of the same age, showing the greatest width of one septum, x B
& L J + Zeiss oc. 2, dt. 160.

Fig. 44. A tangential section of an older larva, showing the connection of a septal muscle with the circum-oral disc.
X Zeiss H-f oc. 2.

PLATE 36.

Figs. 45 to 47 are consecutive transverse sections of one individual. Fig. 45 shows the continuity between the endoderm
of adjacent gastric pouches at the base of an interradial tentacle. Fig. 47 is lower, and here the gelatinous septum com-
pletely separates the two pouches. X Zeiss H -f- oc. 2.

1 Unless otherwise noted, the microscope was used with the tlraw tube not drawn out. Length of tube (dt.) = 137 mm.

236 ROBERT PAYNE BIGELOW ON CASSIOPEA XAMACHANA.

Figs. 48 and 49 are from the same series. Fig. 48 is the second section below Fig. 47. It just clears the oesophagus.
Fig. 49 is through the upper part of the stem. X Zeiss DD + oc. 2.

Fig. 50. Longitudinal section of a scyphistonia with sixteen tentacles, probably a little younger than Fig. 15. X Zeiss.
DD + oc. 2, dt. 195.

Fig. 51. An obliquely transverse section of a fully developed scyphistoma, showing the relations of the septal muscles
to the depressions in the circumoral disc. The mesogloea is shaded. X Zeiss B -f- oc. 2, dt. 160.

Fig. 52. Part of an interradial section from a seyphistoma a little older than the last. Owing to a slight obliquity of
the section, the full extent of the circular sinus at the base of the tentacle is not shown. It extends to the point marked x.
X Zeiss C + oc. 2.

PLATE 37.

Figs. 63 to 56 illustrate the development of the rhopalia.

Fig. 53. Median section of the interradial tentacle shown in Fig. 52 ; x marks a corresponding point in the two sections
X Zeiss H + oc. 2.

Fig. 54. A radial section from th* base of a rhopalial tentacle somewhat older than Fig. 23. X /eiss H -f- oc. 2.

Fig. 55. Radial section of a rhopalium in the stage of Fig. 25. x Zeiss H + oo. 2.

Fig. 56. Radial section of a rhopalium in about the stage of Fig. 26. X Zeiss H + oc. 2.

Fig. 57. Radial section showing the course of a septal muscle in a strobila, at the stage of Fig. 24. X Zeiss DD -f- oc. 2.

Fig. 58. A similar section from a specimen a little older than Fig. 25 ; .r, point of separation between the t\vn disrs.
X Zeiss DD + oc. 2.

PLATE 38.

Fig. 59. Median vertical section of a strobila in the stage of Fig. 26;. x, boundary between ephyrula and basal polyp.
X Zeiss B + oc. 2.

Fig. 60. Portion of section from the same specimen, showing the proboscis of the basal polyp ; x marks same point as
in preceding. X B & L f + Zeiss oc. 2, dt. 160.

Fig. 61. Median vertical section of a basal polyp, stage of Fig. 27. X Zeiss B -f- oc. 2.

Fig. 62. Part of a section from the same specimen, showing the proboscis, p, and the vestige of the former connection
with the ephyrula at x. X B & L f .+ Zeiss oc. 2, dt. 160.

Fig. 63. Median vertical section of an ephyrula that has recently become free ; .c is opposite the opening that formerly
led into the cavity of the basal polyp. Cf. Fig. 29. X Zeiss DD + oc. 2.

Fig. 64. Median vertical section from a young medusa intermediate in age between Figs. 31 and 32. The section is
nearly interradial in position. X Zeiss AA + oc. 2, dt. 160.

Figs, (if) and (ill. Tangential sections of the same specimen, parallel to the last, nearly at right angles to an interradins.
Fig. 60 is the one nearer the periphery, x Zeiss AA + oc. 2, dt. 16O.

Printed, August, 1900.

VITA.

The author of this thesis, Robert Payne Bigelow, is the son of Otis and Margaret
Payne Bigelow, and was born in Baldwins ville, New York, on the 10th of July, 1863.
Since 1868 he has resided in the City of Washington., where he attended the public schools
and afterward studied for two years, until 1882, in the Preparatory School of the Colum-
bian University. A year was spent as bookkeeper for the firm of Otis Bigelow & Co.,
after which he entered the Lawrence Scientific School at Harvard University, whence he
was graduated in 1887 with the degree of Bachelor of Science, magna cum laude.

Another year was spent in business, and then he entered the Johns Hopkins Univer-
sity as a candidate for the degree of Doctor of Philosophy, taking Animal Morphology as
his principal study, with Physiology and Botany as subordinate subjects. While in this
University he has held the positions of University Scholar, Fellow, and Adam T. Bruce
Fellow.

MEMOIRS HOST. Soc. NAT. HIST. VOL. 5

PLATE 31.

RPB, del.

BlGELOW, C'ASSKH'KA XAMACIIAXA.

MEMOIRS BOST. Soc. NAT. HIST. VOL. 5.

rht

PLATE 32.

RPB M

23

BlGELOW, ( '.VSSliil'KA XAMAi IIAXA.

MEMOIRS BOST. Soc. NAT HIST. VOL. 5.

PLATE 33.

27

FtPB del.

BIGELOW, CASSIUI-KA XAMATIIANA.

<:N

33.

•V« •» ,&

M5 •Jfe«;W

*f ;•' ••.'ix

••1 ->'•'' - -"^

? /^ #"•

* <^^w^

V.-

^v;' MF'

B^lJ

m:

32

RPB del

l!li.l;l.ii\\ . (' \^-liPl'KA XAMACHAXA.

MEMOIRS BOST. Soc. NAT. HIST. VOL. 5.

PLATE 35.

end

40

41

Ft.f?B. d,l.

D

39

43

sin

e—

BIGELOW, CASSIOPEA XAMACHANA.

MEMOIRS BOST. Soc NAT HIST VOL. 5.

PLATE 36.

45

-' 6 * p'

,

•

,

_ - SHI

-sin

'

—ft

--

*••* !

'

•:—i

1

\

'>

46

47

RPB del

cd

l'.h.i:i,<i\\ , ( ' V

XAMAOIIAJfA.

MEMOIRS BOST. Soc. NAT HIST. VOL. 5.

PLATE 37.

P ' fe-"

- • . >

5<S

BIGELOW, CASSIOPEA XAMACIIAXA.

MEMOIRS BOST. Soc, NAT. HIST. VOL. 5

PLATE 08.

end-

15iii];i.(i\\ , CASSIOPBA XAV M'IIA